Diborane Structure and Bonding

Boron and hydrogen have only six and one electrons of their own to give. Yet B₂H₆ stands with its atoms held together in a bridge. Ordinary Lewis structures cannot draw this bridge at all. If you count valence electrons for BH₃ the way you would for CH₃, you get a planar, electron-deficient fragment with an empty p orbital. This makes it a strong Lewis acid that should exist as a monomer. It doesn't. Below about 180 °C, two BH₃ units always fuse into gaseous diborane, B₂H₆. This molecule has four ordinary terminal B–H bonds. It also has two extraordinary bridging B–H–B bonds. These bridging bonds hold the whole structure together using fewer electrons than a normal single bond would require. This is the founding case study of electron-deficient bonding in the p-block. It matters far beyond one molecule. The same three-center two-electron (3c–2e) bridge bond glues B₂H₆ together. This bond reappears in higher boranes and in carboranes. It also appears in Ziegler–Natta-type aluminum alkyl dimers. And, reframed as delocalized cluster bonding, it appears in the polyhedral boron hydride anions. Their electron counting is governed by Wade's Rules for Cluster Electron Counting. Understanding diborane is therefore not a historical curiosity. It is the entry ticket to an entire sub-language of inorganic structure. In this language, bond order can be a fraction. Here, "not enough electrons" is not a failure of the atoms. Instead, it is an invitation to delocalize. The puzzle solved itself only when chemists stopped insisting that every bond needs its own private electron pair. Molecular orbital theory shows how this works. The two bridging hydrogens and the two boron atoms form a four-orbital system. This system has four electrons (per bridge, counted appropriately). Its bonding combination is fully occupied. Its antibonding combinations are empty. This outcome is exactly analogous to the Three-Center Two-Electron Bonding motif. That motif describes hypervalent and electron-deficient systems generally. The bridge B–H bond is weaker and longer than a terminal B–H bond. The H–B–H bridge angle is compressed. The whole B₂H₄ core, plus two bridging H's, adopts a rigid D₂h skeleton. All of these are direct, falsifiable predictions of the bonding model. They are not just descriptive labels. This article works through several key topics. These include the geometry and the group-theory-based MO analysis of the B–H–B bridge. It also covers the synthetic routes that make and consume diborane. Finally, it covers the reactivity patterns that follow directly from its electron deficiency. These patterns include hydroboration, adduct formation, and pyrolysis to higher boranes. Every number quoted — bond lengths, angles, force constants — is a real, measured feature. These are features of the actual gas-phase structure. Each one is traced back to the orbital picture that predicts it.

Key takeaways

BH₃ is an electron-deficient 6-electron, 3-bond fragment; two BH₃ units dimerize to B₂H₆ because the boron p-orbital has nowhere else to put its "borrowed" density except into a bridge.
B₂H₆ has D₂h symmetry: four terminal B–H bonds (2c–2e, ~119 pm) in one plane, two bridging B–H–B bonds (3c–2e, ~133 pm) perpendicular to it.
Each bridge is built from one B–B–H fragment SALC pair; the resulting bonding MO holds 2 electrons delocalized over B–H–B, giving each bridge bond an effective order of ½ per B–H contact.
Total valence electron count is 12 (6 B–H bonds' worth of electron pairs needed classically) but only 8 bonding pairs (4 terminal + 2 bridging MOs) are actually available — this "8 pairs for 6 B–H + 2 B–B contacts" deficit is exactly what the 3c–2e bridges resolve.
Diborane is made industrially from BF₃/NaBH₄ or BCl₃/LiAlH₄, and historically from Mg₃B₂ + acid or NaBH₄ + H₃PO₄/I₂; it reacts as a source of BH₃ in hydroboration and as a strong reducing/Lewis acidic reagent.
Key idea: two electrons delocalized over three atomic centers (B–H–B) let boron complete an approximate octet without needing an extra electron pair — the same currency of bonding that later gets counted systematically by Wade's rules for whole cluster cages.

The core insight

Diborane exists because a three-center two-electron bridge bond can stabilize an electron-deficient boron center more cheaply, in electron currency, than any classical two-center bond arrangement — turning two 6-electron BH₃ Lewis acids into one closed-shell, all-bonds-full B₂H₆ molecule.

Where this lives today

Diborane and BH₃-derived reagents are workhorse tools across synthetic and industrial chemistry, not museum pieces. Hydroboration — the anti-Markovnikov, syn addition of a B–H bond across an alkene using BH₃ (generated in situ from NaBH₄/BF₃ or used as a THF or dimethyl sulfide adduct, BH₃·THF or BH₃·SMe₂, both of which are simply Lewis-base-cleaved, monomeric forms of diborane's B–H₃ fragment) is a cornerstone method for anti-Markovnikov alcohol synthesis and remains central to complex natural product and pharmaceutical synthesis; H. C. Brown's development of hydroboration earned the 1979 Nobel Prize in Chemistry. Diborane itself is used industrially as a dopant source (delivering boron atoms) for silicon and compound-semiconductor epitaxy in the electronics industry, where controlled decomposition of B₂H₆ gas deposits precise boron concentrations into silicon lattices for p-type doping. In rocketry and specialty fuel chemistry, boron hydrides were investigated (mid-20th century) as high energy-density fuels because B–H and B–B bond combustion releases very large amounts of energy per unit mass. In materials and catalysis, metal borohydrides (LiBH₄, NaBH₄, Ca(BH₄)₂) synthesized from diborane chemistry are studied as solid-state hydrogen storage materials, exploiting the same electron-deficient B–H bonding logic to store and release H₂ reversibly. And conceptually, the 3c–2e bridge bond first understood in B₂H₆ is the direct ancestor of the cluster bonding framework — Wade's Rules for Cluster Electron Counting — used today to rationalize and design an enormous family of polyhedral boranes, carboranes, and transition-metal clusters, including boron-cluster-based drugs used in boron neutron capture therapy (BNCT) for tumor treatment.

Ask the tutor: How does the "styx" bonding-number formalism extend the diborane bridge-counting method to predict the bonding topology of larger boranes like B₅H₉?

The physical picture

Start from monomeric BH₃. Boron contributes three valence electrons, each hydrogen contributes one, giving six electrons total — exactly enough for three ordinary two-center two-electron (2c–2e) B–H bonds arranged trigonal planar (sp² boron, D₃h symmetry). This uses all six valence electrons and leaves boron with only six electrons around it: an empty, low-lying p orbital perpendicular to the BH₃ plane. That empty orbital is a standing invitation for any electron pair donor to attack, and in the absence of a better donor, another BH₃ molecule's B–H bonding electrons will do the job. Two BH₃ units come together so that one B–H bond of each monomer bends over to donate electron density into the empty p orbital of the other boron — and by symmetry this happens twice, giving two equivalent bridges and a molecule where each boron simultaneously donates into and accepts from the other. The resulting B₂H₆ skeleton is best pictured as an ethylene-like B₂H₄ fragment (four terminal hydrogens, two per boron, lying in a plane with the two borons) with two additional bridging hydrogens positioned above and below the B–B line, each bonded simultaneously to both borons. The point group is D₂h. Structural data (electron diffraction, microwave, and X-ray) give:

terminal B–H bond length: 119 pm
bridging B–H(bridge) bond length: 133 pm
B–B distance: 177 pm
terminal H–B–H angle: 121.5°
bridge H(b)–B–H(b) angle: ~97°
dihedral: the BH₂(terminal) planes and the B–H(b)–B–H(b) bridge plane are mutually perpendicular

The bridging B–H bonds are ~14 pm longer than the terminal ones and noticeably weaker (bridge B–H stretching force constants and bond dissociation energies are lower), exactly as expected for a bond whose order is less than one. Each bridging hydrogen sits essentially symmetrically between the two borons (not disordered — genuinely bonded to both), and this symmetric bridging geometry is only stable because of the 3c–2e bonding pattern described below; a classical structure could not explain why the bridge hydrogens are shorter-lived intermediates in no sense — they are permanent, structurally distinct atoms, confirmed by ¹¹B and ¹H NMR (which show separate signals, in the slow-exchange limit, for bridge vs. terminal H, and characteristic ¹H–¹¹B coupling patterns) and by vibrational spectroscopy (distinct B–H(t) and B–H(b) stretching frequencies, ~2600 cm⁻¹ vs ~1600 cm⁻¹).

Figure

Diborane B₂H₆: D₂h skeleton with terminal and bridging hydrogens

Why must the bridges sit perpendicular to the terminal-hydrogen plane, rather than all six hydrogens being coplanar? Each boron uses roughly sp³-like hybrid orbitals: two hybrids point to the terminal hydrogens (in-plane), and the remaining two point out of that plane toward the two bridging hydrogens, one above and one below. This hybridization pattern is what places the bridge bonds orthogonal to the terminal B–H₂ plane, and it is confirmed by the near-tetrahedral local geometry at each boron (terminal H–B–H angle 121.5°, close to trigonal but distorted by the bridge; bridge H–B–H closer to a compressed tetrahedral angle).

Figure

Electron count per boron: classical BH3 deficiency vs bridged B2H6 resolution

The essential physical point: no single B–H–B bridge bond is a normal covalent bond between two centers. It is a delocalized bond among three atomic orbitals (one H 1s, one B sp³-hybrid from each boron) occupied by just one pair of electrons. This is precisely the Three-Center Two-Electron Bonding situation, and the payoff is that boron, which cannot afford a full octet's worth of 2c–2e bonds with only three valence electrons per atom, gets partial octet completion "for free" by sharing electron density non-locally.

The core ideas

The central quantitative fact is the electron/bond-pair bookkeeping. Diborane has 8 B–H σ-type bonding contacts to describe (4 terminal + 2×2 bridge B···H contacts) plus the B–B framework, but only 6 boron + 6 hydrogen valence electrons per BH₃ unit, i.e. 12 valence electrons total in B₂H₆:

N_{VE}(\mathrm{B_2H_6}) = 2\times(3_{\mathrm B}) + 6\times(1_{\mathrm H}) = 12

where:

$3_{\mathrm B}$ = valence electrons contributed by each boron (group 13),
$1_{\mathrm H}$ = valence electron contributed by each hydrogen,
the sum runs over 2 borons and 6 hydrogens.

Reading it: 12 electrons means 6 electron pairs total. Four of those pairs go into the four ordinary terminal 2c–2e B–H bonds, leaving only two electron pairs to hold the B₂H₄ core together across the B–B gap and attach two more hydrogens. A classical structure would need one pair for a B–B bond and two more pairs for two more B–H bonds — three pairs, not two. The 3c–2e bridge model resolves the shortfall directly:

n_{\text{pairs available for bridging}} = \frac{N_{VE} - 4\times2}{2} = \frac{12-8}{2} = 2

where:

$4\times2 = 8$ is the electron count already used in the four terminal 2c–2e bonds,
the remaining 4 electrons (2 pairs) are exactly enough to fill two bonding molecular orbitals, one per bridge, each delocalized over a B–H–B unit.

Reading it: the electron count is not just consistent with two 3c–2e bridges — it requires exactly two, no more and no fewer, matching the observed structure precisely. This is the master constraint that both the Lewis-structure deficiency and the MO treatment must reproduce. The bridge bond order per B–H contact within a 3c–2e unit is:

\text{bond order (per B–H contact in bridge)} = \frac{1}{2}

where:

the factor of ½ arises because one bonding pair (2 electrons) is shared over two B–H contacts (to the left B and to the right B) within a single bridge MO,
summed over both contacts, the bridging hydrogen's total bonding contribution is one full electron pair, consistent with H needing only 2 electrons to fill its 1s shell.

Reading it: each bridge hydrogen is "half-bonded" to each boron in a formal sense, but fully saturated overall — H's own valence is completely satisfied even though neither individual B–H(bridge) contact reaches bond order 1. This fractional bond order is exactly why bridging B–H bonds are longer and weaker (133 pm, lower stretching frequency) than terminal ones (119 pm, order 1).

Derivation AGroup-theory construction of the B–H–B three-center MOs

Goal: build the symmetry-adapted molecular orbitals for one B–H–B bridge unit and show that exactly one bonding MO is doubly occupied, with the rest non-bonding/antibonding and empty.

Step 1 — Isolate the bridge fragment and choose a local axis system. Consider one B···H···B bridge as a local three-orbital problem embedded in the D₂h molecule. Each boron contributes one hybrid orbital pointing toward the bridging hydrogen (call these $\chi_{B1}$ and $\chi_{B2}$ , each an sp³-like hybrid with a large lobe directed at the bridge H and largely orthogonal to the terminal B–H bonds on that same boron). The bridging hydrogen contributes its single 1s orbital, $\chi_H$ . This gives a 3-orbital basis: $\{\chi_{B1}, \chi_H, \chi_{B2}\}$ .

Step 2 — Form symmetry-adapted linear combinations (SALCs) on the boron pair. Within the local $C_{2v}$ symmetry of the B–H–B fragment (or equivalently by simple symmetric/antisymmetric combination, since the two borons are related by the mirror plane containing H), the two boron hybrids combine as:

\phi_{s} = \frac{1}{\sqrt2}(\chi_{B1}+\chi_{B2}), \qquad \phi_{a} = \frac{1}{\sqrt2}(\chi_{B1}-\chi_{B2})

$\phi_s$ is symmetric under the mirror that swaps B1 and B2 (same symmetry as $\chi_H$ , which sits on the mirror plane and is invariant); $\phi_a$ is antisymmetric, and has no partner of matching symmetry among the fragment orbitals available ( $\chi_H$ is totally symmetric, so it cannot mix with $\phi_a$ ).

Step 3 — Mix by symmetry to get three MOs. Only $\phi_s$ and $\chi_H$ share symmetry, so they mix (two orbitals in, two MOs out — bonding and antibonding):

\psi_{\text{bonding}} = c_1\,\phi_s + c_2\,\chi_H \quad(\text{in-phase, all-positive overlap})

\psi_{\text{antibonding}}^{*} = c_1'\,\phi_s - c_2'\,\chi_H \quad(\text{node between B-pair lobe and H})

with $c_1, c_2, c_1', c_2' > 0$ set by the actual overlap and orbital energies (H 1s lies below the boron sp³ hybrid energy, so $\chi_H$ dominates $\psi_{\text{bonding}}$ and $\phi_s$ dominates $\psi^*_{\text{antibonding}}$ ). The antisymmetric combination $\phi_a$ cannot mix with $\chi_H$ by symmetry and is left as a non-bonding boron-pair orbital, $\psi_{\text{nb}} = \phi_a$ , with a node exactly at the hydrogen and no B–B or B–H overlap contribution of its own.

Step 4 — Order the three MOs by energy and fill with electrons. The three resulting orbitals order as:

E(\psi_{\text{bonding}}) < E(\psi_{\text{nb}}) < E(\psi^*_{\text{antibonding}})

Each bridge unit supplies 2 electrons overall to this local problem (this comes from the global electron count: 2 pairs total are available for 2 bridges, i.e. 1 pair — 2 electrons — per bridge). Those 2 electrons occupy $\psi_{\text{bonding}}$ only. $\psi_{\text{nb}}$ and $\psi^*_{\text{antibonding}}$ remain empty (in the neutral, closed-shell molecule).

Result. One doubly-occupied, fully bonding, three-center MO delocalized over B–H–B accounts for the entire bridge; it is stabilized relative to the separate atomic/hybrid orbitals, contains bonding character between H and both borons simultaneously, and — because it is a single MO holding 2 electrons shared 3 ways — reproduces exactly the ½ bond order per B–H(bridge) contact derived from the electron-counting argument above. Repeating this construction for the second bridge (related by the $C_2$ axis along B–B) gives the complete 3c–2e bridging picture for both bridges of B₂H₆.

Go deeper · Derivation B · 2 minElectron-counting proof that BH3 must dimerize: comparing monomer and dimer total energies via bond-pair balance

Derivation BElectron-counting proof that BH3 must dimerize: comparing monomer and dimer total energies via bond-pair balance

Goal: show, using only valence-electron bookkeeping (not a full ab initio calculation), why two separate BH₃ Lewis-acid/base-incomplete monomers are electronically "worse off" than one B₂H₆ dimer with bridges — i.e., derive the electron-deficiency argument for dimerization quantitatively.

Step 1 — Count boron's electron environment in monomeric BH₃. In BH₃, boron forms three 2c–2e bonds using all 6 valence electrons of the BH₃ unit (3 from B, 3 from the three H's). Each bond gives boron a formal share of 1 electron (half of each bonding pair), so boron's electron count around itself is:

n_e(\mathrm{B\ in\ BH_3}) = 3\times1 = 3\ \text{(shared)} + 0\ \text{(lone pairs)} = 6\ \text{electrons total in its valence shell (counting both electrons of each bond it's part of, i.e., the standard "electron count" convention)}

This is 2 electrons short of an octet (8), leaving one formally empty 2p orbital on boron — the source of BH₃'s strong Lewis acidity.

Step 2 — Ask what a hypothetical B–B single-bonded, non-bridged H₃B–BH₃ dimer (all H terminal) would need. If two BH₃ units simply joined by a normal 2c–2e B–B bond without rearranging any hydrogens, the total electron count would be the same 12 valence electrons, now needing: 6 terminal B–H bonds (6 pairs = 12 electrons) plus one B–B bond (1 more pair = 2 electrons) — a total of 14 electrons required for 7 bonds. Only 12 are available. This hypothetical structure is short by exactly one electron pair:

\Delta n_{\text{pairs}} = 7 - 6 = 1\ \text{pair short}

So a straightforwardly bonded H₃B–BH₃ (each boron still only 3-coordinate plus a B–B bond) cannot be built with the available electrons either — each boron in that hypothetical structure would still have only 6 electrons around it (three bonds: two terminal B–H + one B–B), i.e. it does not even solve the octet problem; it merely relocates it.

Step 3 — Show the bridged structure spends the same 12 electrons more effectively. In the real bridged B₂H₆, one pair per bridge (2 pairs total) is delocalized over B–H–B rather than localized as a B–B bond. Now each boron is four-coordinate (2 terminal H + 2 bridge H), and by the MO result of the derivation above, each boron's local electron count, including its formal share of the delocalized bridge pairs, comes to:

n_e(\mathrm{B\ in\ B_2H_6}) = \underbrace{2\times2}_{\text{2 terminal bonds}} + \underbrace{2\times\left(\tfrac12\times2\right)}_{\text{2 bridge contacts, each contributing a }\tfrac12\text{ share of a full pair}} = 4 + 2 = 6\ \text{bonds' worth of shared density}

Counting both electrons per bond formally assigned back to boron (as in a standard octet count where a shared bond contributes 2 electrons to each atom's "count"), each boron in B₂H₆ sees:

2\ (\text{terminal, 2e each, full share}) \times 2\ \text{bonds} \;+\; 2\ (\text{bridge, sharing }\tfrac12\text{ of the pair, but counted as if a full bond for octet purposes because the pair fully delocalizes over the B–H–B unit}) \times 2\ \text{bridges} \;=\; 8\ \text{electrons: a complete octet}

Result. Delocalizing 2 electrons over 3 centers lets each boron count 4 bonds (2 terminal + 2 bridge) toward its octet while spending only 3 electron pairs total (4 terminal pairs shared between 2 B, plus 2 bridge pairs) rather than the 3.5 pairs per boron that a fully classical structure would require. This is the quantitative reason BH₃ dimerizes: the bridged dimer is the only arrangement of 12 valence electrons among 8 atoms that gives both borons a complete octet simultaneously.

Worked examples

Example 1 (foundational) — Verify the valence electron count and bond inventory for B₂H₆. Boron (group 13) contributes 3 valence electrons; hydrogen contributes 1. With formula B₂H₆:

N_{VE} = 2(3) + 6(1) = 6 + 6 = 12\ \text{electrons} = 6\ \text{pairs}

Assign 4 pairs to the four terminal B–H bonds (unambiguous, each boron keeps two terminal H's, each a normal 2c–2e bond, bond order 1, observed length 119 pm). That leaves 2 pairs for the remainder of the structure — precisely enough for two 3c–2e B–H–B bridges (1 pair each), and not enough for a localized B–B bond plus two more B–H bonds (which would need 3 pairs, as shown in the second derivation). The bridging B–H bonds must therefore be 3c–2e; each individual B–H(bridge) contact has bond order

\text{BO}_{\text{bridge}} = \frac{1}{2}

consistent with the longer (133 pm vs 119 pm) and weaker experimental bond length/force constant for bridging vs terminal B–H. This matches the measured structure exactly and requires no fitting parameter — the electron count alone forces two bridges, no B–B bond, and fractional bridge bond order.

Example 2 (applied) — Predict how AlH₃ trimerizes/oligomerizes differently, and why (contrast case using unequal, "off-special" stoichiometry reasoning). Aluminum, like boron, is group 13 (3 valence electrons), so AlH₃ ("alane") is isoelectronic in a formal sense with BH₃ and is also electron deficient. But aluminum's larger atomic radius and poorer p-orbital overlap with H 1s mean the same 3c–2e bridge motif is retained, yet the aggregation number differs: alane is polymeric, (AlH₃)ₙ, rather than stopping at a discrete dimer like Al₂H₆ (which is only observed transiently in the gas phase and is far less stable than B₂H₆). Apply the same electron-bookkeeping logic used for boron to a hypothetical discrete Al₂H₆ dimer:

N_{VE}(\mathrm{Al_2H_6}) = 2(3) + 6(1) = 12\ \text{electrons, identical count to B_2H_6}

So by electron count alone, Al₂H₆ should be exactly as viable as B₂H₆, with two Al–H–Al 3c–2e bridges. It is indeed the structure found for gas-phase dialane, confirming the electron-counting argument transfers correctly across a group. The reason bulk AlH₃ instead polymerizes into an extended 3-D network (rather than staying as isolated Al₂H₆ molecules, as boron does) is that Al's larger, more diffuse orbitals form weaker, longer 3c–2e Al–H–Al bridges per unit, so it is energetically preferable to spread the electron deficiency over many bridges in an infinite lattice (each Al ending up 6-coordinate, octahedral, in solid AlH₃) rather than localizing it into just two strong bridges per molecule as compact B₂H₆ does. This shows the electron-counting rule (2 bridges suffice for a dimer) is necessary but not sufficient — orbital size/overlap quality (a periodic trend down group 13) governs whether the dimer is thermodynamically the preferred aggregate or just a kinetic gas-phase species.

Example 3 (exam / competition level) — Diborane pyrolysis: use electron and mass balance to rationalize the empirical formula of a higher borane product, and identify the bonding change required. A sealed tube of B₂H₆ is heated at 100–200 °C. Pyrolysis proceeds (through BH₃ and B₃H₇/B₄H₁₀ type intermediates) toward higher boranes; a commonly cited overall pathway is

5\,\mathrm{B_2H_6} \;\xrightarrow{\Delta}\; 2\,\mathrm{B_5H_{11}} + 4\,\mathrm{H_2}

Task: confirm this equation is mass- and electron-balanced, then say what it implies about bonding reorganization. Mass balance (B atoms): left side has $5\times2 = 10$ B; right side has $2\times5 = 10$ B. ✓. Mass balance (H atoms): left side has $5\times6 = 30$ H; right side has $2\times11 + 4\times2 = 22+8 = 30$ H. ✓. Electron count sanity check: total valence electrons must be conserved (redox-neutral, no external oxidant/reductant), which it trivially is since only B and H atoms are being reorganized among molecules — the interesting chemistry is not electron count conservation (automatic) but bonding topology reorganization. Going from B₂H₆ (2 bridges, 4 B–H–B contacts, simple D₂h dimer) to B₅H₁₁ (an open, nido-type cluster with a mix of terminal B–H bonds, bridging B–H–B bonds, and direct B–B bonds) requires breaking several 3c–2e B–H–B bridges of the small boranes and re-forming a more complex bonding network that includes true 2c–2e B–B bonds alongside new 3c–2e B–H–B bridges and (in larger boranes) 3-center B–B–B bonds. This reorganization is exactly the kind of counting problem that becomes systematic — rather than case-by-case — once you invoke the cluster electron-counting framework of Wade's Rules for Cluster Electron Counting, where each borane's skeletal electron pair count (not just total valence electron count) predicts whether the resulting cage is closo, nido, or arachno, and B₅H₁₁ is correctly classified as an arachno cluster on that basis.

Limiting cases, scope & common traps

Scenario	What happens	Why (bonding argument)
Gas-phase BH₃, transiently generated (e.g., from BH₃·THF decomposition)	Monomeric, planar D₃h, highly reactive Lewis acid	No partner available to donate into empty 2p; isolable only in dilute gas phase or as a Lewis-base adduct (BH₃·THF, BH₃·PMe₃)
B₂H₆ itself, standard conditions	Stable gas, D₂h, two 3c–2e bridges	Electron count (12 VE) exactly matches "4 terminal pairs + 2 bridge pairs"; no B–B bond needed or present
Strong Lewis base L added (e.g., NMe₃, CO)	Bridge cleaved symmetrically: B₂H₆ + 2L → 2 L·BH₃	The base supplies the electron pair boron "was missing"; each BH₃ becomes a normal 4-coordinate adduct and no longer needs the bridge
Excess H⁻ source (e.g., NaH)	Unsymmetrical cleavage: B₂H₆ + 2 H⁻ → 2 BH₄⁻	Hydride adds to one boron, breaking bridge heterolytically; BH₄⁻ is a normal, fully classical tetrahedral ion (8 electrons, 4 bonds, no bridges needed)
Down group 13 (Al, Ga)	Bridging persists (Al₂Cl₆, dimeric alkylaluminums) but bulk hydride polymerizes (AlH₃)ₙ	Larger, more diffuse orbitals weaken each individual 3c–2e bridge, favoring higher coordination/aggregation over compact dimer
Pyrolysis at elevated T	B₂H₆ converts to higher boranes (B₄H₁₀, B₅H₁₁, B₁₀H₁₄, …) + H₂	Bridges reorganize into networks with true B–B bonds and 3c–2e B–B–B bonds; governed by Wade's skeletal electron pair rules, not simple BH₃-dimer counting

Common trapMisconception: "the bridging hydrogens are just fast-exchanging protons, not real bonds"

It is tempting to think of the bridge hydrogens as some kind of rapid tautomeric average between two localized B–H bonds (like a fluxional proton hopping back and forth). This is wrong for the ground-state structure: diffraction and spectroscopy show the bridging H sits in a genuine, static, symmetric three-center bonding orbital, not an exchange average of two asymmetric structures. (Fluxionality does occur in some larger boranes and carboranes, where bridge hydrogens migrate over the cage on an NMR timescale, but that is a distinct, higher-temperature dynamical phenomenon layered on top of — not a replacement for — the static 3c–2e bonding description of any one instantaneous structure.)

The chemist behind it

The structural puzzle of diborane was resolved through a combination of physical measurement and theoretical courage. Alfred Stock first prepared and characterized the boron hydrides (including B₂H₆) in the 1910s–1920s using painstaking vacuum-line techniques he invented specifically because the boranes are air- and moisture-sensitive, often pyrophoric, and highly toxic — his systematic isolation of B₂H₆, B₄H₁₀, B₅H₉, B₅H₁₁, B₆H₁₀ and B₁₀H₁₄ gave the field its first reliable structural targets, decades before anyone could explain their bonding. The bonding problem itself — why boron hydrides don't obey ordinary Lewis-structure valence rules — was not solved until the 1940s–1950s, when electron diffraction (Kasper, Lipscomb, and coworkers; and earlier gas electron diffraction studies) established the bridged, non-ethane-like geometry unambiguously, and William Lipscomb developed the systematic molecular-orbital and "styx" bonding-number formalism that correctly accounted for 3c–2e bonds across the whole borane family, work recognized with the 1976 Nobel Prize in Chemistry.

Origins · 1 minWilliam Nunn Lipscomb Jr. and how this was discovered

William Nunn Lipscomb Jr.

1919-2011 · Structural chemistry of boranes, molecular orbital theory of electron-deficient bonding, X-ray crystallography, Nobel Prize in Chemistry 1976

Lipscomb combined X-ray and electron diffraction structure determination with quantum-mechanical bonding analysis to build a complete, general theory of boron hydride structures. His "styx" numbering scheme systematically classified boranes by counting the number of 3-center B-H-B bonds, 3-center B-B-B bonds, 2-center B-B bonds, and terminal B-H bonds required to balance a given borane's electron count — turning what had been a case-by-case puzzle (why does B₄H₁₀ look like that, but B₅H₉ look like this?) into a predictive, rule-governed framework. This framework is the direct structural ancestor of the later, more compact skeletal-electron-pair counting rules developed by Kenneth Wade.

History

- 1912–1920s — Alfred Stock develops vacuum-line techniques and first isolates and characterizes B₂H₆ and higher boranes.

- 1943 — Longuet-Higgins and Bell propose the three-center two-electron bridge bond as the correct bonding model for B₂H₆.

- 1947–1950s — Electron diffraction studies (Hedberg, Schomaker, and others) confirm the bridged D₂h geometry with distinct terminal and bridging B–H distances.

- 1954 onward — William Lipscomb develops the general MO/"styx" theory unifying bonding across the whole borane series.

- 1976 — Lipscomb awarded the Nobel Prize in Chemistry for structural studies of boranes illuminating problems of chemical bonding.

Why doesn't diborane just have a normal B–B single bond with each boron additionally bonded to three hydrogens, like two separate BH3 units joined end to end?

Because there aren't enough electrons. B₂H₆ has 12 valence electrons total (6 pairs). A structure with one B–B bond plus six terminal B–H bonds would need 7 electron pairs (14 electrons) — two more than are available. The molecule instead spends its 6 available pairs as 4 terminal 2c–2e B–H bonds plus 2 delocalized 3c–2e B–H–B bridge bonds, which lets each boron reach an effective octet using fewer electron pairs than the fully localized alternative would require. The bridged structure is therefore not just observed; it is the only arrangement of 12 electrons among 8 atoms that saturates both borons' valence shells.

Are the two bridging hydrogens in B2H6 chemically different from the four terminal ones, or do they exchange rapidly?

They are structurally and chemically distinct in the ground-state molecule: bridging B–H bonds are longer (133 pm vs 119 pm), weaker, and vibrate at lower frequency (~1600 cm⁻¹ vs ~2600 cm⁻¹) than terminal B–H bonds. In slow-exchange NMR conditions, bridge and terminal hydrogens give separate signals. Some larger boranes do show hydrogen "tautomerism" or fluxional bridge-hydrogen migration on the NMR timescale at accessible temperatures, but this is an excited dynamic process superimposed on — not a contradiction of — the static structural distinction present in any single instantaneous geometry, including B₂H₆ itself.

How is diborane actually made in the laboratory or industrially?

Historically, Alfred Stock made it (along with the other boron hydrides) from magnesium boride (Mg₃B₂) treated with aqueous acid, a low-yielding and hard-to-control route. Modern laboratory and industrial syntheses instead use hydride-transfer reactions: sodium borohydride reacting with boron trifluoride etherate (3 NaBH₄ + 4 BF₃ → 2 B₂H₆ + 3 NaBF₄), or NaBH₄ with iodine, or lithium aluminum hydride with boron trichloride (3 LiAlH₄ + 4 BCl₃ → 2 B₂H₆ + 3 LiAlCl₄), all of which deliver hydride to boron centers under controlled, scalable conditions and avoid handling elemental boron hydrides directly until the final step.

What is the connection between diborane's bonding and Wade's rules?

Diborane is the simplest possible case of electron-deficient boron-hydride bonding, solved "by hand" using a single pair of 3-center bridge bonds. As boranes get larger (B₄H₁₀, B₅H₉, B₆H₁₀, B₁₀H₁₄, and beyond), the bonding options multiply to include 3-center B–B–B bonds and complex cage topologies, and hand-counting every bridge and B–B bond becomes impractical. Wade's rules generalize the same underlying electron-counting logic (electrons available versus electrons needed for a fully connected, closed cage) into a single skeletal-electron-pair count that predicts whether a cluster is closo (closed deltahedron), nido (one vertex removed), or arachno (two vertices removed) — with B₂H₆ itself corresponding to the smallest, degenerate arachno-type fragment in that classification.

Why is diborane such a strong reducing agent and Lewis acid at the same time?

Both properties trace back to the same electron deficiency. As a hydride donor, B₂H₆ (via its BH₃ fragments) readily transfers H⁻-equivalent character to reduce carbonyls, nitriles, carboxylic acids and other electrophiles — this is the basis of borane reductions and hydroboration. As a Lewis acid, intact BH₃ units (liberated when the bridges are cleaved by a donor) have an empty boron 2p orbital that eagerly accepts an electron pair from ethers, amines, sulfides, and even alkenes (in hydroboration's initial coordination step). The bridge bonds exist specifically because no external donor is present; introduce one, and the molecule immediately reveals its underlying BH₃ Lewis acidity by falling apart into two adducts.

Does diborane have any unpaired electrons or radical character, given that it looks so "incomplete" electronically?

No — B₂H₆ is a fully closed-shell, diamagnetic molecule. The apparent electron "shortage" relative to a naive Lewis structure is resolved, not left as unpaired spins, by delocalizing electron pairs over three centers rather than confining them to two. Every electron in B₂H₆ is spin-paired in a bonding molecular orbital (four terminal 2c–2e bonding MOs, two bridge 3c–2e bonding MOs); there are no singly occupied orbitals and no radical character. This is a key conceptual point: "electron-deficient" describes a bond-counting mismatch relative to classical 2c–2e bonding, not literal unpaired electrons.

Why do bridging bonds in B2H6 form a plane perpendicular to the terminal B–H2 plane rather than all six hydrogens lying in one plane?

Each boron uses hybrid orbitals arranged approximately tetrahedrally: two point toward the terminal hydrogens within one plane, and the remaining two point out of that plane toward the bridging hydrogens above and below. This hybridization geometry is what forces the bridge B–H–B plane to be orthogonal to the terminal-hydrogen plane, giving the overall D₂h symmetry. A planar, all-six-hydrogens-coplanar arrangement would force poor orbital overlap in the bridges and is not the observed or the calculated lowest-energy geometry.

Diborane Structure and Bonding

Where this lives today

The physical picture

The core ideas

Worked examples

Limiting cases, scope & common traps

The chemist behind it

Before this

Where this leads