Electronegativity (EN) is one of the most-used concepts in chemistry: it predicts bond polarity, the direction of electron transfer, ionic vs covalent character. But a rarely asked question: how do you measure it? There isn't one single electronegativity. There are at least three scales, computed differently, giving slightly different numbers — all valid depending on context.
Conceptual definition (Pauling, 1932)
Linus Pauling, in 1932, proposed the still-cited definition:
> "The electronegativity of an atom is the measure of its tendency to attract toward itself the electrons of a covalent bond."
Conceptually simple, operationally vague. How do you measure that "tendency"?
Scale 1: Pauling — from bond energies
Pauling started from a thermochemical observation. If an A-B bond were purely covalent (energy-wise like identical atoms), its energy should be the geometric mean of A-A and B-B bonds:
E(A-B)_ideal = √(E(A-A) × E(B-B))
In practice, E(A-B) > E(A-B)_ideal. The excess Δ comes from the bond's partial ionic character — when A and B have different ENs, partial charges form and stabilize the system. Pauling postulated:
χ_A − χ_B = √(Δ / 96) (in kJ/mol, scale 0.7-4.0)
with a calibration factor chosen so F = 3.98 (highest in the table). By fixing a reference point (initially χ_H = 2.1), all other ENs follow from experimental bond energies.
Pros: experimentally anchored (calorimetry), historical, still the dominant scale in textbooks.
Con: requires reliable bond energies, which don't exist for every pair (noble gases, rare transition metals, transactinides).
Scale 2: Mulliken — from ionization and electron affinity
Robert Mulliken, in 1934, proposed a different definition based on two measurable atomic properties:
χ_M = (IE + EA) / 2
where IE is first ionization energy and EA electron affinity. Logic: an atom that resists losing an electron (high IE) AND eagerly accepts another (high EA) strongly attracts bonding electrons.
The natural unit is eV. To compare with Pauling, Mulliken proposed:
χ_M(Pauling) ≈ (IE + EA) / 5.4
Pros: fully based on individual atomic properties, no molecules needed, computable even for synthetic elements (transactinides).
Con: EA is notoriously hard to measure for many elements (noble gases: negative EA, alkaline earths: EA near zero). Mulliken values are scarce for less-studied elements.
Scale 3: Allred-Rochow — from effective charge and radius
Allred and Rochow, in 1958, proposed a third formulation based on pure electrostatics:
χ_AR = 0.359 × (Z* / r²) + 0.744
where Z* is the effective nuclear charge felt by the valence electron (computed via Slater's rules) and r the covalent radius in pm. Factor 0.359 and constant 0.744 are calibrated so Allred-Rochow values approximately match Pauling.
Pros: no experimental measurement needed, just structural data (covalent radii) and Slater's formula for Z*. Self-consistent, no circularity.
Con: covalent radii aren't always well-defined (noble gases, lanthanides); strong dependence on the radius definition.
Some values compared
| Element | Pauling (1932) | Mulliken (1934) | Allred-Rochow (1958) |
|---|---|---|---|
| H | 2.20 | 2.25 | 2.20 |
| C | 2.55 | 2.67 | 2.50 |
| N | 3.04 | 3.08 | 3.07 |
| O | 3.44 | 3.21 | 3.50 |
| F | 3.98 | 3.90 | 4.10 |
| Cl | 3.16 | 3.15 | 2.83 |
| Na | 0.93 | 0.95 | 1.01 |
| Cs | 0.79 | 0.73 | 0.86 |
| Au | 2.54 | — | 1.42 |
Values are very close between Pauling and Mulliken (within 0.1-0.2) but can diverge for Allred-Rochow, especially for heavy elements (cf. Au).
Which scale to use?
Case 1: teaching and general organic chemistry. Use Pauling. It's the textbook standard, the basis of classification rules (difference > 1.7 = ionic, 0.4-1.7 = polar, < 0.4 = covalent), and 90 years of empirical intuition.
Case 2: heavy-element chemistry, transactinides, ab initio calculations. Use Mulliken or modern derivatives (Allen, Sanderson). Pauling becomes unreliable above Z ≈ 80 because experimental bond energies are scarce.
Case 3: mineral chemistry, predictions on unknown compounds. Use Allred-Rochow. Its purely structural formulation is useful when working with a known crystalline solid (radii available in structural tables).
Recent evolutions
The Allen scale (1989), based on valence orbital energies from spectroscopy, has become the reference for modern quantum chemistry. It gives consistent values for every element in the table, including transactinides. For fluorine: χ_Allen = 4.193, slightly higher than Pauling, reflecting improved spectroscopic precision.
The Sanderson scale (1955) is still used in mineral chemistry to predict partially ionic bonds in complex oxides.
Why the differences matter to you
If you read a paper that says "EN(F) = 3.98", that's Pauling. "EN(F) = 4.10" is Allred-Rochow. "EN(F) = 4.193" is Allen. None is "more correct" than the others — they measure the same intuition with different tools.
The classic student trap: never compare ENs from different scales. If you compare χ(F) = 3.98 (Pauling) to χ(O) = 3.21 (Mulliken), you're comparing apples to oranges. Stay within one scale.
Electronegativity, despite its imperfections, remains one of the most efficient descriptors in chemistry: a single number per atom to predict bonds, polarities, chemical behaviors. It's the perfect illustration of 20th-century scientific philosophy — reducing a complex phenomenon (chemical bonding) to an operational scalar.