Structures of main-group compounds

Result

Structures of main-group compounds

T-100Home CU-306Threads bonding · structure

Statement

Applying VSEPR across the elements.

Why it matters

Basic VSEPR handles simple cases such as methane or ammonia cleanly, but many important main-group compounds — especially those of the heavier p-block elements, which p-block-trends shows behave increasingly differently from their lighter congeners — involve expanded coordination numbers, multiple lone pairs, or genuinely irregular geometries that require the fuller version of the theory developed here to predict correctly.

Because molecular shape directly controls where a molecule's electron density, and specifically any exposed lone pair, is sterically accessible, the geometries predicted here feed directly into hsab-principle's classification of Lewis acid/base behaviour, making this page a structural prerequisite for much of the reactivity chemistry covered later in the unit.

Hypotheses

Lone pairs occupy more angular space than bonding pairs and so preferentially occupy equatorial rather than axial positions in a trigonal bipyramidal arrangement.A "fatter," less directional lone pair compresses adjacent bond angles more than a bonding pair would; treating lone pairs and bonding pairs as occupying exactly the same angular space would fail to predict the correct, experimentally observed geometry for many five- and six-domain species. Central atoms beyond period 2 can accommodate more than four electron domains ("expanded octets"); period-2 elements cannot.Period-2 elements (C, N, O, F) lack accessible orbitals of comparable energy beyond the standard \(2s\)/\(2p\) set and are limited to a maximum of four electron domains, while heavier main-group elements routinely form five- and six-domain species such as \(\text{PCl}_5\) or \(\text{SF}_6\). Whether expanded-octet species are more accurately described by genuine \(d\)-orbital participation, or, in the modern view, by three-centre-four-electron bonding and increased ionic character instead, remains a real point of refinement beyond simple electron counting, though the VSEPR geometric predictions themselves remain reliable regardless of which bonding description is preferred.

Proof

\text{Count total electron domains (bonding domains + lone pairs) to assign the base electron-domain geometry.}

A bonding domain is counted once per attached atom regardless of bond order; two through six domains give linear, trigonal planar, tetrahedral, trigonal bipyramidal, or octahedral base geometries respectively. A

\text{Molecular geometry (atoms only) vs electron-domain geometry (atoms + lone pairs).}

Removing lone pairs from the domain count while keeping their angular influence gives the characteristic "bent," "trigonal pyramidal," "seesaw," "T-shaped," and "square pyramidal" molecular shapes that differ from the parent electron-domain geometry. A

\text{In 5-domain systems, lone pairs occupy equatorial positions preferentially.}

An equatorial position in a trigonal bipyramid has only two adjacent \(90^\circ\) neighbours, while an axial position has three; since \(90^\circ\) neighbours are more repulsively costly than \(120^\circ\) ones, and lone pairs are the most repulsive domain type (Hypotheses), lone pairs preferentially occupy equatorial sites, determining shapes such as \(\text{SF}_4\)'s seesaw geometry. B

\text{Repulsion order: lone pair-lone pair} > \text{lone pair-bonding pair} > \text{bonding pair-bonding pair}.

This ranking predicts systematic bond-angle compression below the idealised domain-geometry angle wherever one or more domains is a lone pair, in a specific, testable order (e.g. water's H-O-H angle compressed further below tetrahedral than ammonia's H-N-H angle, since water has two lone pairs to ammonia's one). A

\text{Apply Steps 1-4 uniformly across the main group, including expanded-octet centres.}

\(\text{ClF}_3\) (5 domains, 2 lone pairs, both equatorial by Step 3) gives a T-shaped molecule; \(\text{XeF}_4\) (6 domains, 2 lone pairs, positioned trans to each other to maximise mutual separation) gives a square planar molecule — the same electron-counting and repulsion-ranking logic applied consistently, including to species with more than four domains. A

Result

\text{Electron-domain count} \to \text{base geometry}; \quad \text{lone pairs (equatorial-preferring, most repulsive) distort it to the observed molecular shape}

Reading. A single, consistent electron-counting and repulsion-ranking procedure predicts molecular shape across the whole main group, including expanded-octet heavier p-block centres.

Scope. Requires the correct total domain count, including any lone pairs, to be identified first (Step 1); expanded-octet predictions apply specifically beyond period 2 (Hypotheses).

Corollaries & converses

p-block-trends' observation that heavier p-block elements behave increasingly differently from period-2 elements is explained in part directly by the expanded-octet rule here — period-2 elements simply cannot form the five- and six-domain species common among their heavier congeners.
The same lone-pair repulsion ranking used here (Step 4) to predict bond-angle compression is the identical logic used generally whenever comparing related hydrides or compounds down a main group.
hsab-principle's classification of a species as a hard or soft Lewis acid/base often correlates with its VSEPR-predicted geometry, since an exposed, sterically accessible lone pair, readily identified from the electron-domain count developed here, is a prerequisite for that species acting as a Lewis base at all.

Fails without

Drop the equatorial lone-pair preference (Hypotheses): placing a lone pair axially instead in a 5-domain system predicts the wrong molecular shape entirely — \(\text{SF}_4\) would be incorrectly predicted to have a different geometry from its observed seesaw shape, since the more repulsively costly \(90^\circ\) axial-equatorial interactions would be left unminimised.
Ignore the period-2 orbital limitation (Hypotheses): proposing an expanded-octet species for a period-2 element, such as a hypothetical five-coordinate nitrogen or fluorine compound analogous to \(\text{PF}_5\), predicts a species that simply does not exist, since period-2 elements lack the accessible orbitals needed to accommodate more than four electron domains.

Common errors

Forgetting to count a lone pair as an electron domain when assigning the base geometry, giving an incorrect predicted domain count before any distortion is even considered.
Placing a lone pair axially rather than equatorially in a trigonal-bipyramidal (5-domain) system, the opposite of the preferred, lower-energy arrangement (Step 3).
Assuming every second-period p-block element can form expanded-octet species just like its heavier congeners, without recognising the accessible-orbital limitation specific to period 2 (Hypotheses).
Treating a multiple bond (double or triple) as contributing more than one electron domain to the repulsion count, when VSEPR counts a multiple bond as a single domain regardless of its bond order.

Discussion

VSEPR theory in its modern form was largely developed and systematised by Ronald Gillespie and Ronald Nyholm through the late 1950s, building on earlier, more limited ideas from Nevil Sidgwick and Herbert Powell in the 1940s. The theory's continued usefulness, despite being a comparatively simple electrostatic/steric picture rather than a full quantum-mechanical bonding treatment, is precisely that it predicts molecular shape correctly across almost the entire main group using only a straightforward electron count.

The modern preferred bonding description of expanded-octet species has shifted away from literal \(d\)-orbital hybridisation, which more detailed quantum-chemical calculations show contributes only a small amount to the true bonding picture, toward descriptions emphasising three-centre-four-electron bonding and substantial ionic character instead; VSEPR's geometric predictions remain valid regardless of which underlying bonding picture is preferred, since the theory was never intended as a bonding theory in the first place, only a geometry-prediction tool.

Common misconception: that VSEPR explains, in a deep quantum-mechanical bonding sense, why molecules adopt their observed shapes. VSEPR is a successful predictive, largely empirical electron-counting procedure, not a fundamental bonding theory — a separate, more involved question addressed by molecular-orbital-based treatments.

Worked examples

\text{SF}_4: \text{S has 5 domains (4 bonding + 1 lone pair)} \to \text{trigonal bipyramidal base} \to \text{seesaw shape}

With the single lone pair placed equatorially (Step 3, minimising the more costly \(90^\circ\) interactions), the four fluorine atoms occupy two axial and two equatorial positions, giving the characteristic seesaw shape with somewhat compressed F-S-F angles (Step 4). A

\text{XeF}_4: \text{Xe has 6 domains (4 bonding + 2 lone pairs)} \to \text{octahedral base} \to \text{square planar shape}

The two lone pairs are positioned trans to each other (maximising their mutual separation, the lowest-repulsion arrangement available among six octahedral domains), leaving the four fluorine atoms in a perfectly square planar arrangement. A

\text{SF}_4:\text{ seesaw (5 domains, 1 lone pair)} \qquad \text{XeF}_4:\text{ square planar (6 domains, 2 lone pairs)}

Reading. Both classic expanded-octet examples follow directly from the same systematic domain-counting and lone-pair-placement procedure used for ordinary, non-expanded-octet molecules.

Scope. The identical procedure predicts the shape of any main-group species once its total electron-domain count and lone-pair number are correctly identified.

Problems

Predict the molecular shape of \(\text{ClF}_3\), showing the domain count and lone-pair placement used.
Solution
Chlorine has 5 electron domains (3 bonding + 2 lone pairs), a trigonal bipyramidal base geometry. Both lone pairs occupy equatorial positions (Step 3, minimising lone pair-lone pair and lone pair-bonding pair \(90^\circ\) interactions), leaving the three fluorine atoms in a T-shaped arrangement.
\(\text{ICl}_4^-\) has 6 electron domains (4 bonding + 2 lone pairs). Predict its shape, and identify a species discussed on this page that is isoelectronic (in domain terms) with it.
Solution
With 6 domains and 2 lone pairs positioned trans to each other (as in Step 5/Worked example 2), \(\text{ICl}_4^-\) is square planar, directly isoelectronic in domain-counting terms with \(\text{XeF}_4\).
Explain why \(\text{PF}_5\) is a well-known, stable compound but an analogous \(\text{NF}_5\) does not exist.
Solution
Phosphorus, in period 3, has accessible orbitals beyond its standard valence set that allow it to accommodate 5 electron domains (Hypotheses), forming a stable trigonal bipyramidal \(\text{PF}_5\). Nitrogen, in period 2, lacks any such accessible orbitals and is limited to a maximum of 4 electron domains, so a 5-domain \(\text{NF}_5\) analogous to \(\text{PF}_5\) cannot form.