in particular modelling is often done not on an isolated molecule but on a molecule
and its environment of solvent and ions) may be simulated with quantum mechanical
methods [ 23 , 24 ]. In Monte Carlo methods random numbers decide how atoms or
molecules are moved to generate new conformations or geometric arrangements
(“states”) which are then accepted or rejected according to some filter. Tens of
thousands (or more) of states are generated, and the energy of each is calculated by
MM, generating a Boltzmann distribution.
3.3.5 As a (Usually Quick) Guide to the Feasibility of, or Likely
Outcome of, Reactions in Organic Synthesis
In the past 15 years or so MM has become widely used by synthetic chemists,
thanks to the availability of inexpensive computers (even very modest personal
computers will easily run MM programs) and user-friendly and relatively inexpen-
sive programs [ 5 ]. Since MM can calculate the energies and geometries of ground
state molecules and (within the limitations alluded to above) transition states, it
can clearly be of great help in planning syntheses. To see which of two or more
putative reaction paths should be favored, one might (1) use MM like a hand-held
model: examine the molecule for factors like steric hindrance or proximity of
reacting groups, or (2) approximate the transition states for alternative reactions
using an intermediate or some other plausible proxy (cf. the treatment of solvolysis
in the discussion of transition states above), or (3) attempt to calculate the
energies of competing transition states (cf. the above discussion of transition state
calculations).
The examples given here of the use of MM in synthesis are taken from the
review by Lipkowitz and Peterson [ 28 ]. In attempts to simulate the metal-binding
ability of biological acyclic polyethers, the tricyclic 1 (Fig.3.12) and a tetracyclic
analogue were synthesized, using as a guide the indication from MM that these
molecules resemble the cyclic polyether 18-crown-6, which binds the potassium
ion; the acyclic compounds were found to be indeed comparable to the crown ether
in metal-binding ability.
Enediynes like 2 (Fig.3.12) are able to undergo cyclization to a phenyl-type
diradical 3 , whichin vivocan attack DNA; in molecules with an appropriate
triggering mechanism this forms the basis of promising anticancer activity. The
effect of the length of the constraining chain (i.e. ofnin 2 ) on the activation energy
was studied by MM, aiding the design of compounds (potential drugs) that were
found to be more active against tumors than are naturally-occurring enediyne
antibiotics.
To synthesize the very strained tricyclic system of 4 (Fig.3.12), a photochemical
Wolff rearrangement was chosen when MM predicted that the skeleton of 4 should
be about 109 kJ mol$^1 less stable than that of the available 5. Photolysis of the
diazoketone 6 gave a high-energy carbene which lay above the carbon skeleton of
66 3 Molecular Mechanics