Carboranes and Heteroboranes
A cage made of twelve boron atoms and two carbon atoms survives being heated past 600 °C. All fourteen atoms lock into a perfect icosahedron. It does not fall apart. This is ortho-carborane, C₂B₁₀H₁₂. Its sheer indestructibility — chemical and thermal — is not an accident of a strong single bond. Instead, it is a property of the whole cluster acting as one delocalized electronic unit. Replacing a B–H vertex with a C–H vertex, or with a transition metal fragment, barely perturbs the framework. The cage does not care very much which atom sits where. It only cares how many skeletal electrons the whole assembly commands. That single fact — electron count, not atom identity, dictates shape — is the organizing principle of the entire subject. Carboranes and heteroboranes matter because they sit at the crossroads of three big ideas in inorganic chemistry. These are Wade's Rules for Cluster Electron Counting, the Isolobal Analogy and Cluster Extensions, and the practical chemistry of boron hydrides. The Isolobal Analogy connects clusters to organometallic fragments and even to naked metal atoms. Boron hydrides, meanwhile, are among the most electron-deficient, three-dimensionally bonded species known. Where a hydrocarbon uses one two-electron bond per edge, a borane cluster typically has too few electrons to do that. It is forced instead into multicenter bonding spread over triangular faces of a deltahedron. Carboranes are what you get when you swap CH for BH⁻ vertices (isoelectronic substitution) inside these deltahedra. Heteroboranes generalize the game to almost any main-group or transition-metal vertex. The payoff is enormous structural richness from a small number of rules. A single empirical formula, C₂B₁₀H₁₂, corresponds to three distinct, isolable, thermally interconvertible isomers. These are ortho, meta, and para. They are distinguished only by how far apart the two carbons sit on the icosahedron. Removing a vertex from ortho-carborane with a strong base gives the nido anion [C₂B₉H₁₂]⁻. This ligand is so good at binding metals in a pentahapto fashion that it is often called "carbon's answer to cyclopentadienide." But it is bigger, more electron-rich, and far more robust. Metallacarboranes built this way have been used in boron neutron capture therapy. They also serve as building blocks for redox-active materials. They act too as ligands supporting catalytically active metal centers that ordinary Cp ligands cannot stabilize. This article develops the electron-counting logic that predicts carborane and heteroborane shapes. It works through the closo/nido/arachno structural hierarchy with real bond-length and angle data. It derives the famous ortho → meta → para isomerization pathway. It also derives the vertex-removal rule that generates nido-carboranes. Finally, it finishes with worked examples spanning a textbook count, a real synthetic transformation, and a competition-level multi-step electron-counting problem.
- Carboranes replace one or more BH vertices in a borane cluster with isoelectronic CH vertices; the skeletal electron count is unchanged.
- Wade's Rules for Cluster Electron Counting (n+1 pairs = closo, n+2 = nido, n+3 = arachno) predict carborane shapes exactly as they do for boranes.
- o-, m-, and p-C₂B₁₀H₁₂ are all closo 13-pair (n = 12) icosahedral clusters; they differ only in carbon placement and interconvert thermally at high temperature.
- Removing the electron-poorest vertex of o-carborane with base gives nido-[C₂B₉H₁₂]⁻, the "dicarbollide" ligand, isolobal with cyclopentadienide.
- Heteroboranes extend the same electron bookkeeping to main-group heteroatoms (S, P, As) and to transition-metal vertices, unifying cluster and organometallic chemistry.
- Key idea: in cluster chemistry, shape is a function of total skeletal electron pairs, not of which particular atoms supply them.
Where this lives today
Boron neutron capture therapy (BNCT) is the most cited biomedical application: carborane cages (and related closomers) are attached to biomolecules or antibodies that accumulate in tumor tissue, delivering a high local density of ¹⁰B nuclei; thermal neutron irradiation then triggers the ¹⁰B(n,α)⁷Li nuclear capture reaction, depositing lethal, highly localized energy only in boron-loaded cells while sparing surrounding tissue, because the high-linear-energy-transfer alpha and lithium recoil particles travel only about one cell diameter. Carboranes are attractive delivery vehicles precisely because the icosahedral cage is chemically and thermally so robust that it survives the biological delivery and cellular uptake process intact. Metallacarboranes built on the dicarbollide ligand — direct products of the vertex-removal chemistry derived above — function as ligands in homogeneous catalysis where their steric bulk, negative charge, and redox-tunable metal centers offer advantages over ordinary cyclopentadienyl chemistry: cobaltabisdicarbollide anions are used industrially in nuclear waste reprocessing (selective extraction of Cs⁺ and Sr²⁺ from radioactive raffinates) because the anion's extreme chemical and radiolytic stability lets it survive the harsh acidic, radioactive environment for years. Carborane-based extractants and ion-exchange materials exploit exactly the thermal/chemical inertness first noted in the earliest icosahedral carborane syntheses. Materials chemistry uses closo-carborane cages as three-dimensional, electron-withdrawing, extremely robust substituents ("super-adamantane-like" spacers) in liquid crystals, luminescent materials, and polymer backbones, since the cage's high symmetry, thermal stability, and tunable dipole (depending on which isomer, o-, m-, or p-, is incorporated, since the C–C separation changes the local dipole moment substantially) allow fine control over macroscopic material properties without sacrificing chemical robustness. Heteroborane clusters more broadly — with S, P, As, Sb, or Se vertices replacing BH units — extend the same design logic to main-group cluster catalysis and to ligands for unusual metal oxidation states, while metal-vertex heteroboranes (metallaboranes) form a direct conceptual bridge to the Isolobal Analogy and Cluster Extensions that connects all of cluster chemistry, from bare boranes to giant transition-metal carbonyl clusters, under one electron-counting framework.
:::quiz[The nido dicarbollide ligand [C2B9H11]2- is isolobal with which classic ligand?]
- CO
- Cyclopentadienide, C5H5- ✓
- Benzene, C6H6
- Chloride, Cl-
Ask the tutor: How does the isolobal analogy extend beyond CH/BH⁻ to transition-metal fragments like Fe(CO)₃ or Cp*Co, and what electron-counting convention keeps mixed organometallic-carborane clusters consistent?
The physical picture
Start from the borane hydrides themselves. Boron has three valence electrons and one empty 2p orbital more than it needs for four two-electron bonds, so simple boron hydrides cannot be built from the usual two-center two-electron (2c-2e) bonds alone. Instead, boron clusters adopt three-dimensional cage or cage-fragment geometries in which some bonding electron pairs are smeared out over three or more atoms — three-center two-electron (3c-2e) B–B–B or B–H–B bonds. The deltahedra that result (triangulated polyhedra: octahedron, pentagonal bipyramid, icosahedron, and so on) are the closo (closed) shapes; removing one, two, or three vertices from the next-larger closo deltahedron gives the open nido, arachno, and hypho shapes respectively. Carbon has one more valence electron than boron. A neutral CH vertex therefore contributes one more skeletal electron than a neutral BH vertex, but it can only occupy a cluster position if the overall electron count is still consistent with a Wade closo/nido/arachno classification. Because the isoelectronic anion BH⁻ has exactly the same electron count as neutral CH, chemists say that CH and BH⁻ are isolobal: they have the same number, symmetry, and approximate energy of frontier orbitals available for cluster bonding, so one can be substituted for the other with only modest changes to the overall geometry. This is the seed of the broader Isolobal Analogy and Cluster Extensions used throughout cluster and organometallic chemistry. Icosahedral C₂B₁₀H₁₂ is the flagship example. The parent closo borane dianion [B₁₂H�12]²⁻ is a perfect Ih icosahedron with n = 12 vertices and n+1 = 13 skeletal electron pairs. Replacing two BH⁻ units with two isoelectronic CH units removes the 2− charge (two fewer electrons contributed net, since each BH⁻→CH swap removes one electron) and gives neutral C₂B₁₀H₁₂, still a 13-pair closo icosahedron. Three isomers exist because the two carbons can occupy adjacent, meta, or antipodal vertices of the icosahedron, and interconversion between them requires vertex-to-vertex atom migration over the cage surface, not bond cleavage of the whole cluster.
Because the cage is only mildly perturbed by carbon substitution, most C–C, C–B, and B–B distances in o-carborane are close to the values in B₁₂H₁₂²⁻ (B–B ≈ 1.77 Å), but the C(1)–C(2) edge in ortho-carborane contracts to about 1.62–1.64 Å, noticeably shorter than a normal C–C single bond distance found in ethane-like fragments, reflecting the compact multicenter bonding around the electron-poor carbon pair. Vertex removal converts a closo carborane into a nido carborane. Treating the strongest Lewis base attack at the boron atom adjacent to both carbons (the most electrophilic, least electron-rich vertex — the one flanked by two carbons, which donate less electron density to the cage than borons would), degradation of o-C₂B₁₀H₁₂ with ethoxide/ethanol removes the B(3) vertex to give the eleven-vertex, eleven-skeletal-pair open cage [C₂B₉H₁₂]⁻, whose open pentagonal face can then deprotonate to the dianion [C₂B₉H₁₁]²⁻ — the "dicarbollide" ligand, isoelectronic and isolobal with cyclopentadienide, C₅H₅⁻.
:::svg[Vertex removal: closo-C2B10H12 to nido-[C2B9H11]2- dicarbollide]
The core ideas
Wade's rules count skeletal electron pairs (SEP) and classify shape by comparing the vertex count n to the number of pairs.
where: = number of valence electrons of vertex atom (3 for B, 4 for C, 5 for P/As, etc.); = number of electrons donated by exo-cluster substituents on vertex (1 for a terminal H); the per vertex removes the electrons used in the exo B–H (or C–H) 2c-2e bond, leaving the electrons available for skeletal bonding; = overall ionic charge (electrons added for anions, subtracted for cations).
where: n = number of vertex atoms actually present in the cluster (not the parent deltahedron size). Reading this relationship: as the electron count rises relative to the vertex count, the cluster opens up — the extra electron pairs occupy orbitals that would otherwise have to be strongly antibonding in a closed deltahedron, so the cage relieves the strain by losing a vertex and becoming more open, exposing a face of the parent polyhedron.
where: identifies which closed deltahedron (from the closo series: 4 = tetrahedron ... 12 = icosahedron ... ) the observed open cluster is derived from, by adding back the missing vertices until SEP = .
Goal. Show from first principles that C₂B₁₀H₁₂ is a 13-pair, 12-vertex closo cluster, i.e. a closed icosahedron, and confirm this is consistent with the isoelectronic parent [B₁₂H₁₂]²⁻.
Step 1 — Assign vertex electron contributions. There are 10 BH vertices and 2 CH vertices. Each vertex atom uses two of its own valence electrons plus the one electron from its terminal H substituent for the exocluster 2c-2e B–H (or C–H) bond, leaving electrons for skeletal bonding. For boron, , (one H), so each BH vertex donates skeletal electrons. For carbon, , , so each CH vertex donates skeletal electrons.
Step 2 — Sum over all vertices. Ten BH vertices contribute
Two CH vertices contribute
The molecule is neutral, so . Total skeletal electrons:
Step 3 — Classify against vertex count. There are n = 12 vertices in the cage (10 B + 2 C). Compare SEP to n+1:
Since SEP = n+1 exactly, the cluster is closo: a closed 12-vertex deltahedron, i.e. the icosahedron, matching the experimentally determined Ih (idealized) cage geometry of all three C₂B₁₀H₁₂ isomers.
Step 4 — Cross-check against the parent borane. The isoelectronic all-boron dianion is [B₁₂H₁₂]²⁻: twelve BH vertices each donating 2 electrons gives electrons from vertices, plus the 2− charge adds 2 more electrons, for electrons, SEP = 13, n+1 = 13 — identical count and identical closo icosahedral shape. Substituting two BH⁻ units (each an "effective" 3-electron donor once its own extra negative charge is included) with two neutral CH units (3-electron donors intrinsically) leaves the total skeletal electron count and hence the shape completely unchanged, which is exactly the isoelectronic/isolobal replacement rule stated in the overview.
:::derivation[From closo-C₂B₁₀H₁₂ to nido-[C₂B₉H₁₁]²⁻: vertex removal and the dicarbollide ligand] Goal. Derive the electron count and shape of the "dicarbollide" dianion produced by base degradation of o-carborane, and show why it is isolobal with cyclopentadienide.
Step 1 — The degradation reaction. Treatment of o-C₂B₁₀H₁₂ with sodium ethoxide in ethanol (or with piperidine/other strong bases) removes the boron vertex adjacent to both carbons — the most electrophilic, most positively polarized vertex, because it is flanked by two carbons that withdraw cage electron density less effectively than boron neighbors would, leaving that particular boron most deficient and thus most attacked by the incoming nucleophilic base. The base attacks this BH vertex, is incorporated as a B–OEt leaving unit (formally removed as "BH(OEt)⁻" chemistry, with H₂ evolved), and the cage opens to a nido, eleven-vertex fragment:
Step 2 — Deprotonation to the dianion. The monoanion [C₂B₉H₁₂]⁻ carries an extra "endo" bridging hydrogen on the open pentagonal face. Treatment with a strong base (e.g. NaH) removes this proton to give the dianion:
Step 3 — Electron count of [C₂B₉H₁₁]²⁻. Now n = 11 vertices (9 B + 2 C), each with one terminal H (x=1). Boron vertices contribute electrons; carbon vertices contribute electrons; the charge contributes electrons. Total:
Step 4 — Classify. Compare SEP to n+2 (the nido criterion):
SEP = n+2, so the cluster is nido: derived from a 12-vertex closo parent (the icosahedron again — consistent, since we removed exactly one vertex from C₂B₁₀H₁₂'s parent icosahedron) with one vertex missing, leaving an open pentagonal face of five boron/carbon atoms bearing the cage's frontier orbitals.
Step 5 — Isolobal comparison with Cp⁻. The open C₂B₉ face presents five roughly coplanar atoms with a set of frontier orbitals — one totally symmetric orbital plus two doubly degenerate pairs of matching symmetry — pointing outward, exactly analogous to the π-system frontier orbitals of cyclopentadienide, C₅H₅⁻, which is a 6 π-electron (aromatic) monoanion. Both ligands are formally dianion-equivalent donors of six electrons to a metal in the η⁵ sense once charge and hapticity are matched consistently (Cp⁻ donates 6 electrons as an anionic η⁵ ligand under the ionic counting convention; [C₂B₉H₁₁]²⁻ similarly donates 6 electrons as a dianionic η⁵ face), so the two ligands are isolobal and dicarbollide complexes such as [Fe(C₂B�9H₁₁)₂]²⁻ ("ferrabis-dicarbollide") are structurally and electronically the direct analogues of ferrocene.
Worked examples
Example 1 — Foundational: classify nido-2-CB₉H₁₂⁻ ("monocarbaborane"). This anion has one carbon and nine boron vertices, all bearing a single terminal H, plus one bridging H somewhere on an open face (bridging hydrogens are counted separately as one skeletal electron each, distinct from terminal exo-H's which are already folded into the term for their own vertex; here we instead use the more careful atom-by-atom count). Ten vertex atoms total: nine BH units, one CH unit. Each BH vertex donates 2 skeletal electrons (); the CH vertex donates 3 skeletal electrons (). That is
The formula CB₉H₁₂⁻ has ten vertex atoms carrying ten terminal H's (one each), plus two extra H atoms beyond that (12 total H, 10 used as terminal substituents), which must be bridging hydrogens; each bridging H contributes one electron to the skeletal count (it is not tied up in an exo 2c-2e bond exclusively — it bridges two boron atoms using cluster-type bonding). Two bridging H atoms add
The single negative charge adds
Total skeletal electrons:
There are n = 10 vertices. Compare to n+2 = 12: SEP = 12 = n+2, so this is nido, derived from an 11-vertex closo parent (octadecahedron/edge-contracted icosahedron-type cage), consistent with the known structure of monocarbaundecaborate anions used as ligand precursors. As a limiting-case sanity check, setting the bridging-H count to zero and adjusting charge instead would need to keep SEP the same — showing the electron count, not the particular hydrogen distribution, is what fixes the shape classification.
Example 2 — Applied: is [(η⁵-C₂B₉H₁₁)₂Fe]²⁻ electronically analogous to ferrocene? This is the classic "ferrabis(dicarbollide)" complex, made by reacting FeCl₂ with two equivalents of Na₂[C₂B₉H₁₁]. Each dicarbollide ligand was shown above to be a nido, 13-pair, 11-vertex cluster whose open C₂B₉ face is isolobal with Cp⁻ and donates 6 electrons in the ionic counting convention (as a formal 2− ligand). Count electrons at iron the same way one counts ferrocene: Fe(II) center: , contributes 6 electrons. Two dicarbollide ligands, each donating 6 electrons η⁵-fashion: electrons. Total electron count at the metal:
This is an 18-electron, coordinatively saturated sandwich complex, exactly like ferrocene (). The complex is overall dianionic because each dicarbollide ligand is 2− and Fe is 2+: , matching the observed [FeC₂B₉H₁₁)₂]²⁻ charge. As a limiting check, replacing one dicarbollide with ordinary Cp⁻ (as has been done synthetically to give mixed "half-sandwich" metallacarboranes) leaves the count at again — confirming that the isolobal replacement of Cp⁻ by [C₂B₉H₁₁]²⁻ (with appropriate charge bookkeeping) truly preserves the electron-counting logic, not merely the geometry.
Example 3 — Competition level: track the electron count through an anti-Markovnikov cage-opening cascade. A sample of o-carborane, C₂B₁₀H₁₂ (SEP = 13, closo, n=12, as derived above) is subjected to two sequential reactions: (i) degradation with KOH/EtOH to remove one boron vertex and give a nido anion, then (ii) that nido anion is further treated with excess strong base and heated, causing loss of a second boron vertex (an unusual but documented further degradation under forcing conditions) to give an arachno dianion of formula [C₂B₈H₁₁]³⁻ after appropriate protonation state adjustments. Verify the electron count is internally consistent at each stage and identify the parent deltahedron size implied at the arachno stage. Stage 0: C₂B₁₀H₁₂, SEP = 13, n = 12, SEP = n+1: closo (icosahedron), as derived earlier. Stage 1: remove one BH vertex, add appropriate charge from base attack, as derived above: [C₂B₉H₁₁]²⁻ (after full deprotonation), SEP = 13, n = 11, SEP = n+2: nido, derived from the same 12-vertex icosahedral parent with one vertex vacant. Stage 2: remove a second BH vertex. Now n = 10 vertices (8 B + 2 C). Suppose the product, after adjusting protons and charge under forcing base conditions, is [C₂B₈H₁₀]²⁻ complexed with one bridging hydrogen retained plus the stated 3− charge in a differently protonated tautomer — for the purposes of a clean count, take the well-defined arachno species [C₂B₈H₁₁]⁻ (a real characterized arachno-carborane skeleton). Count: eight BH vertices contribute electrons; two CH vertices contribute electrons; one extra bridging H contributes 1 electron; the single negative charge contributes 1 electron. Total:
With n = 10 vertices, compare SEP to n+3 (the arachno criterion):
This does not satisfy arachno counting for n=10; instead compare to n+2:
So [C₂B₈H₁₁]⁻ is actually nido (not arachno) relative to an 11-vertex parent deltahedron — a useful reminder that successive vertex removal from a closo parent does not automatically step the classification from closo → nido → arachno in lockstep with each individual boron lost; the actual class depends on recomputing SEP versus the current n at every stage, because bridging hydrogens and protonation state change along with vertex count and these must be retallied explicitly rather than assumed. (The genuinely arachno eight-vertex carboranes reported experimentally, such as [C₂B₇H₁₃]²⁻-type species, carry a different bridging-hydrogen count that pushes SEP up to n+3 = 11, i.e. 11 pairs for 8 vertices — a distinct, correctly-tallied case.) This example is included specifically to show that Wade's-rule classification is a per-molecule recomputation, never a mechanical decrement.
Limiting cases, scope & common traps
| Situation | What Wade's rules predict | Caveat |
|---|---|---|
| n=12, SEP=13 (C₂B₁₀H₁₂, all isomers) | closo icosahedron | true for o-, m-, p- identically; isomers differ only in C positions, not electron count |
| n=11, SEP=13 ([C₂B₉H₁₁]²⁻) | nido, from 12-vertex parent | open pentagonal face is the reactive/coordinating face |
| n=10, SEP=12 ([C₂B₈H₁₀]²⁻-type) | nido from 11-vertex parent, not arachno | must recompute against current n at each degradation step |
| n=8, SEP=11 (genuine arachno C₂B₆ cages) | arachno from 11-vertex parent | bridging-H count changes with each deprotonation/protonation |
| Transition-metal vertex (e.g. Fe in metallacarboranes) | treat as isolobal donor fragment, count via ionic or neutral convention consistently | mixing ionic and covalent/neutral electron-counting conventions in the same molecule gives wrong totals |
| High-temperature carborane isomerization (o→*m→p*) | no change in SEP, no change in closo classification | this is a geometric rearrangement (vertex migration on a fixed polyhedron), not an electron-count change |
The chemist behind it
Origins · 1 minWilliam Nunn Lipscomb Jr. and how this was discovered
Lipscomb solved the crystal structures of boranes and carboranes using early X-ray diffraction techniques of his own devising (low-temperature data collection, careful correction methods for light-atom, electron-deficient systems where conventional bonding pictures gave nonsensical bond orders), and developed the styx numbering and topological/molecular-orbital treatments of three-center bonding that made sense of these otherwise bewildering cage shapes. His structural and bonding work on boron hydrides, several of which are direct parents or close relatives of the icosahedral carboranes discussed here, won him the Nobel Prize in Chemistry in 1976. Kenneth Wade, building on this structural foundation together with the molecular orbital calculations of Rudolph Pitzer, Kenneth Wade himself, and Michael Mingos, then distilled the observed shape patterns into the electron-counting rules that carry Wade's name and that this article applies throughout.
Why does carbon substitution barely change the icosahedral cage shape?
Carbon and the boron-hydride anion fragment BH⁻ are isolobal: both present the same number and symmetry of frontier orbitals to the cluster, and swapping one for the other leaves the total skeletal electron count unchanged. Since Wade's Rules for Cluster Electron Counting show that shape is fixed by the total electron pair count relative to vertex number, not by which specific atoms occupy which vertices, the icosahedron survives the substitution essentially intact. Small local distortions do occur — the C–C edge in ortho-carborane contracts to about 1.62–1.64 Å versus a typical B–B edge of about 1.77 Å — because carbon's extra nuclear charge and different orbital energies still perturb local bonding, but the global deltahedral topology is preserved because the electron count, the fundamental variable controlling closo/nido/arachno classification, is unchanged.
What is the difference between ortho-, meta-, and para-carborane, and how do they interconvert?
All three are C₂B₁₀H₁₂ icosahedra differing only in the separation of the two carbon vertices: adjacent (1,2- or ortho), meta (1,7-), or antipodal (1,12- or para). They interconvert thermally, with ortho converting to meta around 400-500 °C and meta to para requiring still higher temperatures (600 °C+), via vertex-migration mechanisms on the closed polyhedral surface (diamond-square-diamond-type rearrangements) that never break the cage open. No skeletal electron pairs are gained or lost during this rearrangement; it is a purely geometric isomerization driven by the thermodynamic preference to separate the two electronegative, less electron-donating carbon atoms as far apart as possible, which lowers electron-electron repulsion and cage strain.
What exactly is a "dicarbollide" ligand and why is it important?
Dicarbollide, [C₂B₉H₁₁]²⁻, is the nido, eleven-vertex, open-faced anion obtained by removing one boron vertex from ortho-carborane with strong base and then deprotonating the resulting bridging hydrogen. Its open pentagonal C₂B₃ face presents a set of frontier orbitals isolobal with the π-system of cyclopentadienide, so it binds metals in an η⁵ fashion exactly as Cp⁻ does, but it is bulkier, more electron-rich, dianionic rather than monoanionic, and vastly more thermally and chemically robust. This makes dicarbollide complexes attractive both as structural analogues of metallocenes for fundamental study and as practical, radiation- and acid-resistant extractants for separating radioactive cesium and strontium from nuclear waste.
Can heteroatoms other than carbon replace boron vertices in these clusters?
Yes — this is exactly what defines the broader class of heteroboranes. Any main-group atom that can be tallied consistently in the Wade electron-counting scheme (its valence electrons plus any exo-substituent electrons, minus the 2 electrons tied up in its own external bond) can, in principle, occupy a cluster vertex: sulfur, phosphorus, arsenic, antimony, and selenium heteroboranes are all known, along with metallaboranes in which an entire transition-metal-ligand fragment occupies a vertex and donates a calculated number of skeletal electrons determined by the same isolobal bookkeeping used for organometallic fragments. The unifying feature is always that shape follows from total skeletal electron pairs versus vertex count, regardless of the specific chemical identity of the vertex atoms.
Why is o-carborane so much more chemically and thermally stable than typical boranes?
Ordinary electron-poor boron hydrides are often air- and moisture-sensitive, reactive species because their exposed B–H–B bridging bonds and low-coordinate boron vertices are kinetically and thermodynamically reactive toward nucleophiles, oxidants, and hydrolysis. In the closed 12-vertex icosahedral carborane cage, every vertex is fully coordinated within the deltahedral framework, each boron or carbon atom bears only a single, unreactive terminal C–H or B–H bond pointing outward, and the delocalized multicenter bonding of the closo cage is a genuinely closed-shell, thermodynamically very stable electronic configuration (SEP = n+1 exactly, the electronically optimal closo count). This combination of steric encapsulation and electronic saturation is why o-carborane survives strong acids, strong oxidants, and temperatures well above 400 °C without decomposing.
How does one decide which boron vertex is removed during nido-cage formation from o-carborane?
The vertex removed by nucleophilic base attack is the boron atom bonded to both carbon vertices simultaneously (conventionally labeled B(3) in ortho-carborane's numbering). Carbon atoms are more electronegative than boron and pull electron density toward themselves and away from their immediate boron neighbors, so the boron flanked by two carbons is left the most electron-poor, most electrophilic vertex in the whole cage, and is preferentially attacked by an incoming nucleophile such as ethoxide. Once that vertex is excised (as a boron ester leaving group, with concurrent hydrogen loss), the cage relaxes into the eleven-vertex nido dicarbollide precursor whose open pentagonal face contains both original carbon atoms.
Does isomerizing o-carborane to p-carborane change the molecular formula or electron count?
No. All three isomers share the identical molecular formula C₂B₁₀H₁₂ and the identical skeletal electron pair count of 13 for 12 vertices (SEP = n+1, closo). What changes between isomers is purely the relative position of the two carbon atoms on the fixed icosahedral framework — adjacent in ortho, separated by one vertex in meta, and diametrically opposite in para. The interconversion proceeds through vertex-migration (rearrangement) mechanisms on the intact cage surface at elevated temperature, driven by the reduced electronic/steric repulsion when the two carbons are farther apart, but no bonds to the outside world are broken and no electrons are gained or lost in the process.