Just enter your email
A peer-reviewed article of this Preprint also exists.
I found out that the basis set used is insufficient for proper optimizations. I have now redone the whole computations with (aug)-pcseg-1 and computed single points with DSD-BLYP. The corrected values show that the proposed pathway seems to be too costly. The conclusions of this preprint can therefore not be relied upon presently.
I am exploring other pathways now.