One of the more unusual publications recently involved a theoretical computation of a hypothetical carbonium ion (or at least a very short-lived molecule-ion) C-(CH₃)₅ . (Angew. Chem. Int. Ed 53: 7875 – 7878.) Computations concluded that the structure was that of a trigonal bipyramid, which effectively had three methyl groups around the central carbon atom, which was in sp² configuration, and two other methyl groups bonded to the p orbital of the central carbon atom. All methyl groups were in sp³ configuration. The important point about such a computation is that the ion is argued to be sufficiently stable that it exists, albeit short-lived, as it has two computed decay modes. The question now is, is it right? The issue is important because it proposes a type of bonding that so far has not been recognized, or if it has, the recognition passed by me.
 
There is one important point to note. Computations indicate that the CH₅ ion does not follow the same structure. This ion can be considered as a distorted CH₃ system that bonds to a H₂ molecule. This gives three equivalent hydrogen atoms and two further equivalent, but different atoms. This is supported by the infrared spectrum (Science 309: 12219 – 1222) which shows a fluxional molecule consistent with that structure and with full hydrogen scrambling. Why does the replacement of hydrogen atoms with methyl groups make such a difference? Then again, does it?
 
The CH₅ ion is conceptually simple, in that it is really a carbenium ion making an electrophilic attack on a two-electron bond. Now, if it will do that to the hydrogen molecule bond, why does the same thing not happen with, say the (CH₃)₃ – C ion which could make an electrophilic attach on the C – C bond of ethane?
 
The next question is, does it matter? I think it does, because it calls into question a number of bond issues. The first is, where is the formal positive charge? In my view, it starts on the central carbon atom. I argued that the gas phase stabilities of the usual carbenium ions is given quite satisfactorily by assuming the positive charge is first located at the formal ion centre, and it then polarizes the substituents (Aust J. Chem. 26 : 301-310.) That makes the (CH₃)₃ – C ion considerably more stable than the CH₃ ion, and that ion would more readily polarize the bond in ethane. The issue then resolves itself to whether the formation of a two-electron C – C – C bond, plus the polarization energy is lower than the energy of the C – C bond in ethane, plus its polarization energy. A further question then is, is a two-electron C – C – C bond even possible? What we are asking, at least in conventional chemical thinking, is for the two methyl electrons that are separated over that distance to pair, and get the appropriate phase relationship. What disturbs me about this is that there are no other examples that I can think of where a vacant p orbital can bind two electrons in that way. My immediate thinking then makes me ask, is there any equivalent in boron chemistry? I am not sufficiently familiar to say there is not, but I am certainly unaware of any. Therefore the question is, does B-(CH₃)₅ exist? If the trigonal bipyramid structure for C-(CH₃)₅ is correct, one would think it should because the troublesome ionic character that leads to rearrangement is missing. If on the other hand, such an ion represents the (CH₃)₃ – C ion polarizing ethane, then there should be no B-(CH₃)₅.