The ionization of a C-X bond (X=halogen) to form what we call a carbocation and which is known as the SN-1 reaction goes way back in the history of chemistry. Julius Steglitz was probably the first person to suggest such an ionization, back in 1899 (Steglitz, J.; Am. Chem. J., 1899, 21, 101). The reaction became very famous during the 1920s onwards, and could be regarded as one of the cornerstones of organic chemistry. A question I like to ask whenever talking about a reaction is “what is the transition state like?”. Although answering such a question can get tricky, one might imagine that it should be relatively simple for such a fundamental reaction as the SN-1.
Well, it turns out it is not! Looking at the energy of the system as a function of extending the C-Cl distance usually produces a curve that rises sharply, way past the actual free energy barrier for e.g. the solvolysis of tert-butyl chloride (which is around 21 kcal/mol), and which produces no actual transition state. Such a gas-phase model is simply not realistic, and to make it so, we have to include solvent. Such a model does yield a transition state, about which the following aspects can be noted:
- At least 13 water molecules are needed to model this reaction; more would be better, but it gets increasingly difficult to fully optimise their positions as you add more
- The reason so many are needed is that they cross-polarise each other. One water molecule initiates an SN2 like attack from the back side of the t-butyl chloride. As the O…C bond develops, that water molecule becomes positively charged. This positive charged water forms an unusually strong hydrogen bond to a second molecule, which transfers part of this positive charge to the second water. This forms a H-bond to a third water, again stronger than usual because of the charge of the second. At the other end of the molecule, the C…Cl bond is gradually leaving and becomes a chloride anion. This too stabilises by forming a hydrogen bond to an adjacent water. Again this is stronger than normal due to the partial transfer of charge. Gradually, a chain of waters bridging the first water and the chloride forms.
- In fact, its better to form at least two such bridges rather than just one. Perhaps even three chains might form (but I have not yet succeeded in locating the transition state for such!).
- The bridging waters also form little water trimers as they go, a particularly stable arrangement for water. These too gain from the cross-polarisation.
- This arrangement allows a proper close ion-pair to form, and a properly locatable transition state on the way to this species to be characterized.
- Once the ion pair is formed, one of two things can now happen
- The original water is still fully protonated, with the associated positive charge. It is easily the best leaving group in the system. So a neutral water can come in from the side of the chloride, and undergo a second SN2 like displacement of that charged water. Net effect? Production of t-butanol with retention of configuration at the carbon.
- The polarisation induced by the first charge water, and the last charged chloride amounts to partial proton transfer via a chain like mechanism. In fact, it takes very little further energy to fully transfer one proton from the water end of things to the chloride. This produces t-butanol with inversion of configuration at the carbon.
- If each of the above two steps were to be equally likely. the outcome would be that the production of t-butanol occurs with apparent racemisation! In reality, the dynamics of the system will probably also play an important role in determining the outcome.
- Take a look at the transition state shown above. You will notice a remarkable degree of rotation of two of the methyl groups. This makes the transition state highly dependent on the mass of the hydrogens on these methyls. Now, it is known that replacing the Hs with deuterium induces a very large isotope effect on the reaction. Now we know why! It is more difficult to move a heavy atom in the TS than a light one. In fact, the isotope effect can be calculated from the transition state, and it agrees almost exactly with experiment.