Lone Pair and Multiple Bond Effects

Ammonia's H–N–H angle is 106.7°. This is not the tetrahedral 109.5° you'd guess from counting four electron domains. Water bends further still, to 104.5°. Sulfur tetrafluoride puts its lone pair squarely in the equatorial belt of a trigonal bipyramid. It does not sit at an axial vertex. Phosphorus pentachloride and its relatives distort in a completely predictable way. This happens whenever a double bond or a lone pair joins ordinary single bonds around a central atom. None of this is arbitrary. It is what happens when you take the electron-domain geometries of VSEPR Molecular Geometries seriously. Then you ask a sharper question — do all electron domains repel equally? The answer is no. The reason matters far beyond a set of memorized bond angles. Lone pairs are held by one nucleus alone. Their electron density piles up closer to the central atom. It spreads into a fatter, more diffuse cloud. This cloud is broader than a bond pair squeezed between two nuclei. Multiple bonds pack two or three electron pairs into a single bonding region. This region occupies one "direction" in space. Still, its multiply-occupied electron cloud is broader. It is also more repulsive than a single bond's. Both effects push neighboring domains around. They close some angles and open others. Both effects also favor specific coordination sites. This applies in structures with inequivalent sites to offer. Examples include the axial and equatorial vertices of a trigonal bipyramid. This article turns a qualitative VSEPR ranking into a working toolkit. That ranking is "lone pair–lone pair > lone pair–bond pair > bond pair–bond pair." It explains why the ranking exists. The explanation rests on electron density and orbital hybridization. It shows how it explains the bond-angle progression from methane through ammonia to water. It explains why sulfur and phosphorus fluorides adopt the shapes they do. It also shows how multiple-bond domains distort angles in molecules like formaldehyde, phosgene, and xenon oxytetrafluoride. We build the reasoning from electron-domain repulsion energetics. We apply it to a ladder of real molecules with measured angles. We show precisely where the simple picture breaks down. By the end you will be able to predict something important. You will know, before consulting any table, whether a lone pair or a multiply-bonded substituent will sit axial or equatorial in a five-coordinate species. You will understand why the O–S–O angle in SO₂F₂ differs from the F–S–F angle. You will see why bond-angle deviations are a genuine measurement of relative electron-pair "size." They are not a fudge factor. This groundwork feeds directly into Bent's Rule and Orbital Character. There, the same asymmetry is reinterpreted through hybridization. It also feeds into Walsh Diagrams. There, it is recovered from molecular orbital correlation diagrams.

Key takeaways

Electron domains repel unequally: lone pair–lone pair > lone pair–bond pair > bond pair–bond pair, because lone pairs are held by one nucleus and spread into fatter, closer clouds.
This ranking explains the angle sequence CH₄ (109.5°) → NH₃ (106.7°) → H₂O (104.5°) as lone pairs replace bond pairs one at a time.
Multiple bonds behave like "fat" single-bond domains: extra electron density in the bonding region increases their effective repulsive size, widening flanking angles.
In five-coordinate (trigonal bipyramidal) geometries, lone pairs and double-bonded groups preferentially occupy equatorial sites, minimizing 90° repulsive contacts.
The same hierarchy governs distortions in six-coordinate octahedral lone-pair-bearing species and rationalizes measured angles in SF₄, ClF₃, XeOF₄, and carbonyl-containing molecules.
Quantitative bond-angle changes correlate with electronegativity and π-bond order, setting up the more detailed hybridization argument of Bent's rule.
Key idea: electron domains are not point charges of equal size — lone pairs and multiple bonds are effectively "fatter" than single bond pairs, and geometry bends to keep the fattest domains as far apart as possible.

The core insight

Not all electron domains are created equal: lone pairs and multiply-bonded groups occupy more angular "space" than ordinary single bond pairs, so real molecular geometries relax away from idealized VSEPR angles to relieve the resulting excess repulsion, with the biggest domains claiming the roomiest (equatorial) sites when inequivalent positions exist.

Where this lives today

The lone-pair and multiple-bond distortion rules are not museum pieces — they are load-bearing tools in modern structural and mechanistic chemistry. In catalysis and reaction mechanism, predicting whether an incoming nucleophile or a leaving group occupies an axial or equatorial site in a five-coordinate phosphorane or sulfurane transition state (relevant to phosphate ester hydrolysis and to the Berry pseudorotation pathways central to nucleophilic substitution at phosphorus in nucleic acid chemistry and organophosphate pesticide/nerve-agent hydrolysis) rests directly on the same axial/equatorial energetics developed here. Main-group and noble-gas chemistry — XeF₄, XeOF₄, IF₅, BrF₅ — is essentially a running exam in lone-pair-driven distortions of octahedral and trigonal-bipyramidal frameworks, still used to benchmark computational methods against precise microwave and electron-diffraction geometries. In materials chemistry, the same reasoning explains "stereochemically active" lone pairs in solid-state structures — the distorted coordination environments of Pb(II), Sn(II), Sb(III), and Bi(III) in perovskite-type semiconductors and ferroelectrics (PbTiO₃-type distortions, lone-pair-driven polar structures in BiFeO₃) trace back to exactly the same lone-pair "size" and directional-repulsion arguments, now applied to lattice coordination polyhedra rather than isolated small molecules. In organic and organometallic chemistry, understanding why a carbonyl oxygen widens the H–C–H or R–C–R angle in ketones and aldehydes relative to a saturated analogue, or why a metal-bound oxo or imido ligand distorts a coordination sphere, is a direct multiple-bond-domain application. Finally, this entire qualitative framework is the launching point for the more quantitative hybridization-based reasoning of Bent's Rule and Orbital Character and for the molecular-orbital-based explanation of angle-dependent stability captured in Walsh Diagrams, both of which recover and sharpen the same bond-angle trends from a completely different, orbital-energetic starting point — a valuable cross-check that the electron-domain picture, despite its simplicity, gets the right physics.

Ask the tutor: Why does the lone pair in ClF₃ produce a smaller absolute angle compression than the single lone pair in SF₄, even though ClF₃ has two lone pairs?

The physical picture

Basic VSEPR treats every pair of bonding or nonbonding electrons around a central atom as an equivalent point of negative charge, and asks that these points spread out to minimize pairwise repulsion — exactly like charges on a sphere. This "hard-sphere" idealization gets you the coordination geometry right (linear, trigonal planar, tetrahedral, trigonal bipyramidal, octahedral) but it cannot, by construction, explain why the H–N–H angle in ammonia is smaller than the H–C–H angle in methane even though both are AX₃E-type... no, wait: methane is AX₄ (no lone pair) and ammonia is AX₃E (one lone pair), and it is precisely this substitution of a bond pair by a lone pair that the simple point-charge model cannot resolve without an extra rule. That extra rule is the electron-domain size hierarchy. A lone pair belongs to a single nucleus. It is not stretched between two positively charged centers the way a bonding pair is, so nothing pulls its electron density away from the central atom. The result is a lone pair cloud that sits closer to the central atom, on average, and that consequently occupies a wider angular cone as seen from that nucleus — think of it as a squat, broad balloon rather than a narrow cigar-shaped one. A bond pair, by contrast, is shared between the central atom and a ligand; the ligand's nucleus pulls density away from the central atom, thinning and elongating the domain along the bond axis. Because repulsion falls off steeply with the angular overlap of two electron clouds, the broader lone-pair domain pushes harder on its neighbors than a bond pair does. This is the microscopic origin of the ranking

\text{lone pair–lone pair} \;>\; \text{lone pair–bond pair} \;>\; \text{bond pair–bond pair}

Multiple bonds add a second, related effect. A double or triple bond crams two or three electron pairs into the same bonding region rather than one, and even though the σ and π components occupy the same gross spatial direction, the combined electron density is larger and closer to the central atom in the perpendicular (angular) sense than a single bond pair's density is. VSEPR treats a double bond as occupying "one domain" in terms of counting steric number, but that domain behaves as an oversized, fat single-bond domain for repulsion purposes — pushing neighboring bonds away and widening the angles flanking it, exactly as an extra lone pair would.

Figure

Electron-domain "fatness" hierarchy: lone pair > multiple bond > single bond

The consequences ripple through every part of a structure that has more than one kind of electron domain. In tetrahedral AX₃E and AX₂E₂ species, successive replacement of bond pairs by lone pairs compresses the remaining bond angles because the growing lone-pair population pushes harder on its neighbors than the bond pairs it replaced. In trigonal bipyramidal AX₅ frameworks, the two coordination sites are not equivalent: an axial position has three 90° neighbors and one 180° neighbor, while an equatorial position has only two 90° neighbors and two 120° neighbors. Because 90° close contacts dominate repulsion energy, the least crowded site — equatorial — is where you want to park your fattest domain, whether that domain is a lone pair or a double-bonded oxygen. This single geometric fact, that equatorial sites see fewer 90° contacts than axial sites, is the master key to essentially every distorted trigonal-bipyramidal and octahedral structure you will meet.

Figure

Axial vs equatorial contact counting in a trigonal bipyramid

The core ideas

The qualitative electron-domain ranking can be written as a repulsion energy that falls steeply with angle and scales with a domain-dependent "size" factor:

E_{ij} \;\propto\; \frac{s_i \, s_j}{\theta_{ij}^{\,n}}

where:

$E_{ij}$ is the pairwise repulsion energy between domains $i$ and $j$
$s_i, s_j$ are effective "size" (steric bulk) factors for each domain, with $s_{\text{lone pair}} > s_{\text{multiple bond}} > s_{\text{single bond, less electronegative substituent}} > s_{\text{single bond, more electronegative substituent}}$
$\theta_{ij}$ is the angle subtended between the two domains at the central atom
$n$ is an empirical exponent (typically taken $\gtrsim 2$ – $6$ in point-charge and later ab initio–calibrated VSEPR models), reflecting the steep, short-range character of exchange repulsion between electron clouds

Reading: because the size factors multiply rather than add, a lone pair interacting with a bond pair contributes more repulsion than two ordinary bond pairs at the same angle, and the molecule responds by increasing $\theta_{ij}$ for the lone-pair-containing interactions (pushing the lone pair away from its neighbors) while decreasing the angle between the remaining bond pairs (since they are pushed together by the retreating lone pair, and their own mutual repulsion is comparatively weak). A second, complementary relationship tracks how the total repulsion energy changes as one bond angle opens at the expense of another when the central atom's total coordination geometry (steric number) is fixed:

\delta\theta_{\text{small}} \;\approx\; -\left(\frac{s_{\text{LP}} - s_{\text{BP}}}{s_{\text{BP}}}\right)\delta\theta_{\text{large}}

where:

$\delta\theta_{\text{small}}$ is the compression of the bond–bond angle not involving the lone pair
$\delta\theta_{\text{large}}$ is the opening of the lone-pair–bond angle relative to the ideal (unperturbed) polyhedral angle
$s_{\text{LP}}, s_{\text{BP}}$ are the lone-pair and bond-pair size factors respectively

Reading: this is a statement of angular "conservation" — a fixed total amount of angular relaxation is redistributed among the domains, so if the lone pair muscles open its own angles to neighboring bonds, the bonds pushed together must close ranks proportionally more the "fatter" the lone pair is relative to a bond pair.

Derivation AFrom point-charge repulsion to the NH₃/H₂O bond-angle sequence

Goal. Show, using only the domain-size hierarchy and a minimal energy argument, why the H–X–H angle shrinks steadily as CH₄ (0 lone pairs) → NH₃ (1 lone pair) → H₂O (2 lone pairs), and estimate the direction and rough magnitude of each step.

Step 1 — Set up the idealized tetrahedral reference. All three molecules are AX $_n$ E $_{4-n}$ species built on a tetrahedral (steric number 4) electron-domain arrangement, so in the point-charge limit — where every domain is an identical, structureless negative charge — all six pairwise domain–domain angles equal the tetrahedral angle,

\theta_{\text{tet}} = \arccos\!\left(-\tfrac{1}{3}\right) = 109.47^{\circ}

This is the geometry-only prediction, blind to what occupies each vertex.

Step 2 — Introduce the size hierarchy for one lone pair (NH₃). Replacing one X (bond pair, size $s_{\text{BP}}$ ) by a lone pair (size $s_{\text{LP}} > s_{\text{BP}}$ ) raises the repulsion energy of every angle that includes the lone pair, for fixed angle. The molecule minimizes total repulsion by increasing the three lone-pair–bond-pair angles above 109.47° (pushing the lone pair's neighbors away from it) while decreasing the three bond-pair–bond-pair angles below 109.47° to compensate, since geometric closure (the three N–H bonds plus the lone pair must still fill all of space around nitrogen) forces some angles to shrink if others grow. Because there are three lone-pair–bond angles and only three bond–bond angles in a trigonal-pyramidal AX₃E system, and because the bond-pair–bond-pair repulsion is the weakest term in the hierarchy, essentially all of the "give" is absorbed by compressing the H–N–H angles.

Step 3 — Quantify using the measured NH₃ angle. Experimentally, the H–N–H angle in ammonia is 106.7°, a compression of

\Delta\theta_{\text{NH}_3} = 109.47^{\circ} - 106.7^{\circ} = 2.8^{\circ}

This 2.8° closure is the direct fingerprint of the extra repulsive push from a single lone pair replacing a single bond pair — small in absolute terms because the lone pair only has to out-compete one bond pair's worth of repulsion at each of the three H–N–H angles, redistributed over three angles simultaneously.

Step 4 — Add a second lone pair (H₂O). In water, two of the four tetrahedral vertices are lone pairs. Now there are two lone-pair-containing domains, and by the ranking lone pair–lone pair repulsion is the single strongest interaction in the molecule — stronger than either lone pair–bond pair or bond pair–bond pair term. This dominant term forces the lone-pair–lone-pair angle open substantially, and because water has only one remaining bond–bond angle (H–O–H) to absorb the geometric compensation, that single angle must close much more than any one angle did in ammonia's three-way distribution.

Step 5 — Compare the measured compressions. The measured H–O–H angle is 104.5°, giving

\Delta\theta_{\text{H}_2\text{O}} = 109.47^{\circmm} - 104.5^{\circ} = 4.97^{\circ} \approx 5.0^{\circ}

almost double the ammonia compression, consistent with Step 4's reasoning: water has two lone pairs mutually repelling each other at the strongest level of the hierarchy, and only one H–O–H angle available to release that strain, compared to ammonia's three H–N–H angles sharing a single, weaker lone-pair perturbation. The trend

109.47^{\circ} \;(\text{CH}_4) \;>\; 106.7^{\circ}\;(\text{NH}_3) \;>\; 104.5^{\circ}\;(\text{H}_2\text{O})

is exactly the monotonic sequence the lone-pair-size argument predicts, and the accelerating (nonlinear) size of the compression — 2.8° then a further 2.2° — mirrors the disproportionate strength of lone-pair–lone-pair repulsion relative to lone-pair–bond-pair repulsion.

Limiting check. If lone pairs had exactly the same effective size as bond pairs ( $s_{\text{LP}} = s_{\text{BP}}$ ), the point-charge model would predict no deviation at all from 109.47° in either NH₃ or H�2O — contradicting experiment. The nonzero, monotonically increasing compressions observed are precisely the experimental proof that $s_{\text{LP}} > s_{\text{BP}}$ , not an assumption smuggled in after the fact.

Go deeper · Derivation B · 2 minWhy lone pairs and double bonds prefer equatorial sites in trigonal bipyramids

Derivation BWhy lone pairs and double bonds prefer equatorial sites in trigonal bipyramids

Goal. Derive, from 90°-contact counting, why a lone pair (or a doubly-bonded oxygen) in a five-coordinate AX₄E (or AX₄Y, Y = O) species occupies an equatorial rather than an axial vertex, and use this to predict the shape of SF₄.

Step 1 — Enumerate near-neighbor contacts for each site type. In an ideal trigonal bipyramid, an axial vertex has three neighbors at 90° (the equatorial belt) and one neighbor at 180° (the opposite axial vertex). An equatorial vertex has two neighbors at 90° (the two axial vertices) and two neighbors at 120° (the other two equatorial positions). Repulsion energy is dominated by the closest contacts because $E \propto \theta^{-n}$ with $n$ large, so the number of 90° contacts is the single most important geometric variable.

Step 2 — Compare 90°-contact counts. An axial site suffers three 90° contacts; an equatorial site suffers only two. Placing the largest, most repulsive domain — whether a lone pair or a double-bond oxygen — at the site with fewer strong (90°) contacts always lowers total repulsion energy, all else equal.

Step 3 — Apply to SF₄ (AX₄E, steric number 5, one lone pair). Sulfur tetrafluoride has four bonding pairs (to F) and one lone pair around a steric-number-5 sulfur center. By Step 2, the lone pair occupies an equatorial site, leaving three fluorines that would occupy the remaining two equatorial and two axial slots... but with only three F left for four remaining slots, one equatorial slot is taken by the lone pair and the fourth fluorine (the other equatorial position) is occupied by F, while the two axial slots are occupied by the remaining two F atoms. The overall shape is a "seesaw" (disphenoidal): two axial F, one equatorial F, one equatorial lone pair.

Step 4 — Predict and check the resulting distortions. The lone pair, sitting equatorial, pushes hardest on its two axial neighbors (each at a nominal 90°) and more gently on its one equatorial neighbor (nominal 120°). Both angles should therefore compress from their idealized values:

\angle F_{\text{ax}}\text{–S–}F_{\text{ax}} < 180^{\circ}, \qquad \angle F_{\text{eq}}\text{–S–}F_{\text{eq(LP)}} < 120^{\circ}

Experimentally, the axial F–S–F angle in SF₄ is 173°, compressed by 7° from the idealized 180°, and the equatorial F–S–F(lone pair) angle is 101.6°, compressed by roughly 18° from the idealized 120°. Both compressions have the correct sign predicted by an equatorial lone pair, and the much larger compression of the equatorial angle (101.6° vs. 120° ideal) versus the axial angle (173° vs. 180° ideal) makes sense because the equatorial F–S–F angle directly competes with the lone pair for one of sulfur's two "roomy" 120° equatorial slots, while the axial fluorines are merely bent slightly toward each other by the lone pair's push from the side.

Limiting check. If the lone pair were forced axial instead (the wrong assignment), it would create three 90° repulsive contacts rather than two, raising the total repulsion energy — this configuration is never observed for AX₄E species and serves as the counterfactual that confirms the equatorial preference is an energy minimum, not an arbitrary convention.

Worked examples

Example 1 (foundational): Ranking bond angles in PF₃ vs. PF₅'s equatorial belt. Phosphorus trifluoride, PF₃, is AX₃E on phosphorus: three P–F bond pairs and one lone pair, steric number 4, tetrahedral electron-domain geometry, trigonal pyramidal molecular shape (isostructural in domain-count terms with NH₃). Apply the same logic as the ammonia derivation: the ideal tetrahedral angle is 109.47°, and the lone pair — being fatter than a bond pair — compresses the F–P–F angle below this value. The measured F–P–F angle in PF₃ is 96.3°, a striking compression of

\Delta\theta_{\text{PF}_3} = 109.47^{\circ} - 96.3^{\circ} = 13.2^{\circ}

far larger than ammonia's 2.8°. Why so much bigger? Fluorine is highly electronegative, so it pulls bonding electron density strongly away from phosphorus, thinning the P–F bond-pair domains considerably (a preview of the Bent's Rule and Orbital Character argument: electronegative substituents make bond pairs "skinnier," effectively widening the size gap between them and the lone pair). A skinnier bond-pair domain offers even less resistance to the lone pair's push than the N–H bond pairs in ammonia did, so the compression is much larger. This example shows that the lone-pair/bond-pair size ranking is necessary but not sufficient on its own — substituent electronegativity modulates the magnitude of the effect even while the direction (compression) remains fixed by the lone-pair-size argument alone.

Example 2 (applied): Predicting the O–S–F vs. F–S–F angles in thionyl fluoride, SOF₂. Thionyl fluoride, SOF₂, has sulfur as the central atom bonded to one oxygen (by a formal S=O double bond), two fluorines (single bonds), and one lone pair — steric number 4 (treating the S=O as one domain), giving a tetrahedral-based, trigonal-pyramidal-like AX₃E shape (with one of the three "X" substituents being the double-bonded O). We predict the ranking of all three angle types using the domain-size hierarchy, now including the double bond:

s_{\text{lone pair}} > s_{\text{S=O double bond}} > s_{\text{S–F single bond}}

Since the double-bonded oxygen is the second-fattest domain in the molecule (after the lone pair), it should compress the F–S–F angle (between the two "skinny" single bonds) more than it compresses the O–S–F angles, and the lone pair compresses everything relative to ideal tetrahedral values. Measured structural data for SOF₂ show F–S–F ≈ 92.8° and O–S–F ≈ 106.8°, confirming both predicted trends simultaneously: F–S–F is the smallest angle in the molecule (squeezed between two "skinny" single-bond domains flanked by both the lone pair and the fat S=O domain), while O–S–F, involving the bulky double bond on one side, remains noticeably larger, closer to (though still below) the ideal tetrahedral value. This is a compound application of both the lone-pair effect and the multiple-bond effect acting in the same molecule, each pushing the F–S–F angle down from two independent directions.

Example 3 (exam/competition level): Predicting the shape and rationalizing angles of ClF₃. Chlorine trifluoride, ClF₃, is AX₃E₂ — three bond pairs, two lone pairs, steric number 5, trigonal-bipyramidal electron-domain geometry. First, decide where the two lone pairs go among the five vertices (2 axial + 3 equatorial). By the equatorial-preference rule (Step 2 of the second derivation above), each lone pair independently prefers an equatorial slot to avoid a third 90° contact. Since there are three equatorial slots and only two lone pairs, both lone pairs occupy equatorial positions, leaving one equatorial and both axial slots for the three fluorines. This gives the well-known T-shaped molecular geometry: two axial F atoms and one equatorial F atom, with the two lone pairs occupying the remaining two equatorial sites. Now predict the angle distortions. The two equatorial lone pairs are mutually at 120° in the idealized geometry — the strongest repulsive interaction in the molecule (lone pair–lone pair, the top of the hierarchy) — so this angle should open considerably beyond 120°, at the expense of compressing the axial F–Cl–F angle (nominally 180°, flanked on both sides at 90° by the two lone pairs) inward, and also compressing the equatorial-F–Cl-to-axial-F angles (nominally 90°) inward as the lone pairs swell outward from their 120° positions. Quantitatively, the two mutually repelling equatorial lone pairs push the axial fluorines away from the equatorial plane and toward each other on the opposite side, so the measured axial F–Cl–F angle is 175°, compressed by

\Delta\theta_{\text{ax}} = 180^{\circ} - 175^{\circ} = 5^{\circ}

and the F $_{\text{eq}}$ –Cl–F $_{\text{ax}}$ angle is compressed to about 87.5°, i.e.

\Delta\theta_{\text{eq-ax}} = 90^{\circ} - 87.5^{\circ} = 2.5^{\circ}

Both compressions have the sign predicted by two mutually repelling equatorial lone pairs pushing all bonding domains away from themselves — a two-lone-pair intensification of exactly the seesaw distortion seen in one-lone-pair SF₄, now doubled in the sense that two fat domains (rather than one) are doing the pushing, even though the individual angle compressions here happen to be smaller in magnitude than SF₄'s single-lone-pair case because the T-shape spreads the strain differently among fewer independent bond angles.

Limiting cases, scope & common traps

Situation	What the model predicts	What actually happens / caveat
Equal-size domains (hypothetical)	No deviation from idealized polyhedral angles	Never physically realized — all lone pairs are measurably "fatter" than bond pairs; this is the null/limiting check baked into every real molecule
Very electronegative substituents (e.g., F, O) on the central atom	Bond pairs thin out, amplifying lone-pair-driven compression	Confirmed (PF₃ vs. NH₃); this is the seed of Bent's Rule and Orbital Character
Steric number 5, one lone pair (AX₄E)	Lone pair goes equatorial (fewer 90° contacts)	Universally observed (SF₄, ClF₃ core logic, SeF₄, TeCl₄)
Steric number 5, two lone pairs (AX₃E₂)	Both lone pairs equatorial, T-shaped molecule	Observed in ClF₃, BrF₃; small deviations from ideal T-shape angles as derived above
Steric number 6, one lone pair (AX₅E)	Lone pair occupies any vertex (octahedral vertices are all equivalent by symmetry) but pushes the four cis bonds toward the lone pair's opposite face	Square pyramidal shape with basal atoms bent slightly toward the unique axial position opposite the lone pair (e.g., IF₅, BrF₅); no axial/equatorial choice needed since all six octahedral vertices are equivalent before the lone pair is placed
Multiple bonds treated as single point domains	Same coordination number, but flanking angles should widen relative to an all-single-bond analogue	Confirmed in carbonyl and sulfonyl compounds (formaldehyde H–C–H = 116.5° vs. tetrahedral 109.5°, opened by the fat C=O domain)
Very large exponent $n$ in the repulsion law	Only nearest (90°) contacts matter; 120° and 180° contacts negligible	Reasonable first approximation, but real molecules show measurable, if smaller, effects even from 120°/180° neighbors — the point-charge model is qualitative, not quantitative

The chemist behind it

The systematic ranking of electron-domain repulsion strengths is usually credited to Ronald Gillespie, working with Ronald Sydney Nyholm in the 1950s to formalize what had been Sidgwick and Powell's earlier qualitative electron-pair repulsion ideas into the full VSEPR (Valence Shell Electron Pair Repulsion) model, including the crucial lone-pair-versus-bond-pair size hierarchy that explains bond-angle deviations rather than just idealized geometries.

Origins · 1 minRonald James Gillespie and how this was discovered

Ronald James Gillespie

1924–2021 · McMaster University, Ontario, Canada, mid-to-late 20th century

Gillespie trained as a physical chemist in London under Nyholm and Ingold, working initially on the physical chemistry of superacid solutions before turning to systematize why molecular shapes deviated from the simple electrostatic point-charge predictions of early valence-shell repulsion theories. Together with Nyholm, he published the 1957 review that crystallized VSEPR as a predictive, teachable framework, explicitly ranking lone-pair and multiple-bond repulsions above ordinary bond-pair repulsions and using this ranking to rationalize measured bond angles across the periodic table — from ammonia and water through the interhalogens and noble-gas compounds. Gillespie spent most of his career at McMaster University, where he continued refining and defending VSEPR against competing bonding models, later co-developing the complementary Ligand Close-Packing model to address some of VSEPR's quantitative limitations while preserving its qualitative, pedagogically powerful core.

History

- 1940: Sidgwick and Powell publish the earliest systematic electron-pair repulsion rules for predicting molecular shape from electron count.

- 1957: Gillespie and Nyholm publish their foundational review formalizing VSEPR, including the lone-pair/bond-pair/multiple-bond size hierarchy.

- 1960s–1970s: Gillespie extends and defends VSEPR systematically against ligand-field and early molecular-orbital critiques, applying it to interhalogens and noble-gas compounds such as XeF₄ and ClF₃.

- 1970s–1980s: microwave and electron-diffraction structural data on SF₄, ClF₃, and related molecules provide the precise bond angles that quantitatively confirm the lone-pair-size predictions.

- 1990s: Gillespie introduces the Ligand Close-Packing model as a complementary, more quantitative refinement addressing systematic deviations in VSEPR's angle predictions for larger central atoms.

Does the lone-pair-size argument mean lone pairs are physically bigger atoms/particles than bond pairs?

No — a lone pair is not a different kind of particle, just a differently distributed cloud of the same two electrons. Because it is held by only one nucleus rather than shared between two, its electron density is not pulled outward toward a second atom, so it sits closer to the central atom and spreads over a wider angular range as seen from that center. "Bigger" here means angularly broader and more repulsive at a given angle, not more massive or containing more charge. The distinction matters because it grounds the whole ranking in real electron-density distribution rather than treating lone pairs as mysterious oversized balloons by fiat.

Why do multiple bonds count as only "one domain" in steric number counting, yet still act like a fat domain for angle purposes?

Steric number counts the number of directions in space occupied by electron density around a central atom — a double bond, however many electron pairs it contains, occupies exactly one direction, so it counts once. But the repulsive strength of that single direction depends on how much electron density is packed into it, and a double bond has twice the bonding electron pairs of a single bond squeezed into the same angular slot. So the "one domain, but a fat one" description is entirely consistent: one direction (relevant to counting), fat/repulsive occupation of that direction (relevant to angle distortion).

If I forgot which is bigger, lone pair or multiple bond, is there a quick way to remember the order?

Yes: think about how tightly each type of electron density is held to the central nucleus. A lone pair belongs entirely to the central atom — nothing pulls it away — so it is the most diffuse and repulsive. A multiple bond is shared with a neighboring atom, so some density is pulled toward that neighbor, but it still packs extra pairs into one direction, making it fatter than an ordinary single bond. A single bond to an electronegative atom is thinnest of all, since the substituent pulls density furthest away from the center. The order lone pair > multiple bond > single bond (electronegative substituent thinnest) follows directly from "how much of this electron density stays close to the central atom."

Does this framework work equally well for third-row and heavier central atoms like sulfur, phosphorus, and xenon?

Qualitatively yes — the same equatorial-preference and lone-pair-compression rules correctly predict the shapes of SF₄, ClF�3, XeF₄, XeOF₄, and IF₅. Quantitatively, heavier central atoms show some systematic deviations that plain VSEPR undercounts, which is part of why Gillespie later developed the Ligand Close-Packing model — treating ligand–ligand (rather than purely electron-domain) contacts as the dominant term for larger central atoms with more crowded coordination spheres. The qualitative direction of every distortion discussed here, however, remains correct across the periodic table.

Is there a simple experimental way to "see" that lone pairs are bigger than bond pairs, beyond just measuring bond angles?

Bond-angle deviations are themselves the primary experimental evidence, obtained from microwave spectroscopy, electron diffraction, and X-ray/neutron crystallography, all of which independently confirm the same compressions discussed here. A second, independent line of evidence comes from photoelectron spectroscopy and computed electron-density maps (e.g., from ab initio or DFT calculations), which show lone-pair electron density genuinely occupying a broader solid angle around the central nucleus than bond-pair density does, visually confirming the "fat balloon versus thin cigar" picture described qualitatively above.

How does this topic connect to Bent's rule, which I'll meet next?

Bent's rule reinterprets the same lone-pair/bond-pair asymmetry through hybridization: it states that central atoms concentrate more p character in bonds to more electronegative substituents (and more s character in bonds to less electronegative ones or in orbitals holding lone pairs), which is really a more mechanistic, orbital-based restatement of exactly the electronegativity-amplification effect seen in the PF₃ versus NH₃ comparison here. Where this article treats lone-pair size as an empirical ranking, Bent's rule explains why that ranking has the magnitude it does, in terms of s/p orbital mixing.

Why does a lone pair in an AX₅E octahedral-based molecule (like IF₅) not need an axial/equatorial choice the way SF₄'s lone pair does?

In a perfect octahedron, all six vertices are symmetry-equivalent before anything is placed on them — there is no separate "axial" and "equatorial" class the way there is in a trigonal bipyramid, where axial and equatorial vertices have genuinely different numbers of 90° neighbors. Once the lone pair occupies any one octahedral vertex, the remaining five ligand positions become inequivalent (four "basal" plus one "apical," trans to the lone pair), producing the square-pyramidal shape with basal atoms bent toward the lone pair's opposite face — but this asymmetry arises only after placement, not as a pre-existing choice between site classes.