Is B2 a compound with no sigma bond but only a pi bond?

[Indranil]

Could suggest a compound with no sigma bond only with the pi bond?

 

[DoItForYourself]

Hi,

This kind of compound does not exist, as long as simple covalent bond is sigma bond and a double/triple covalent bond have one sigma bond and 1 or 2 pi bonds (respectively).

 

[Jason2]

When you think about it, any atom with, say, p orbitals available to form pi bonds would also be able to form sigma bonds with them; and that sigma would be more energetically favoured. So we would have to rely on steric constraints preventing sigma bond formation. It is difficult to imagine a scenario where that would be the case. Can you think of one?

 

[TeethWhitener]

This depends on the precise definition of a sigma bond. I'm not sure there is one. If you require cylindrical symmetry of the bond (as in the Wikipedia article), then you have to contend with the fact that C₂H₂ (or any alkyne) has a cylindrically symmetric pi electron density. On the other hand, it's unclear whether you can call the interaction between Fe and Cp in ferrocene a sigma interaction, even though the bonding is clearly covalent and quite strong.

I suppose you could make the argument that a sigma bond is one where the maximum electron density lies on the line connecting the two atom centers (this would exclude the pi electron sheath in acetylene, whose wavefunction has a node along this line). I can't think of any attractive atom-atom interaction where this criterion would not be present. Maybe in side-bonded H₂ or ethylene η² transition metal complexes? The bond here is formed between a filled metal d-orbital and the empty antibonding orbital (σ* for H₂ and π* for C₂H₄), along with a "backbond" between the filled H-H or C=C bonding orbitals and an empty metal d-orbital. (I use quotes because I'm not sure that the metalheads consider this a true backbond.) It would require an eelctronic structure calculation to figure out the maximum electron density of this bonding configuration. Maybe I'll give it a shot if I have time.

 

[Jason2]

Yes, the back-bonding in complexes of ethylene with metal ions would certainly be candidates for pi bonds - though the forward donation part of the bonding could be construed as 2 sigma bonds. This is territory where things are not so clean-cut!

 

[DrDu]

Could suggest a compound with no sigma bond only with the pi bond?

B2

 

[Jason2]

B2

B2 is a good call! It does indeed seem to be a case where 2 "half pi" bonds are stronger that a single sigma bond. However B2 is paramagnetic so high spin could play a part in that.

 

[DrDu]

B2 is a good call! It does indeed seem to be a case where 2 "half pi" bonds are stronger that a single sigma bond. However B2 is paramagnetic so high spin could play a part in that.

Another effect which contributes here is that the 2pz orbitals and the resulting sigma bond get pushed up by the 2s orbitals on the neighbouring atom. On the other hand, the 2s orbitals get lowered in energy. Probably a careful calculation of the bond orders would reveal that there is a net sigma bonding contribution though much smaller than in typical sigma bound compounds.
This mechanism becomes unimportant in later compound of the series, as the s-p energy difference increases when going from B to C, N, O, F. E.g. in N2, the sigma bond formed from the pz orbitals is lower than the pi bond orbitals.