- (a) Leach AR (2001) Molecular modelling, 2nd edn. Prentice Hall, Essex, England, chapter
12; (b) H€oltje H-D, Folkers G (1996) Molecular modelling, applications in medicinal
chemistry. VCH, Weinheim, Germany; (c) van de Waterbeemd H, Testa B, Folkers G
(eds) (1997) Computer-assisted lead finding and optimization. VCH, Weinheim, Germany - Dewar MJS, Storch DM (1985) J Am Chem Soc 107:3898
- For a series of small, mostly nonbiological molecules AM1 seemed better than PM3, except
for O-H/O hydrogen bonds: Dannenberg JJ (1997) J Mol Struct (Theochem) 410:279 - In model systems of biological relevance, mostly involving water, PM3 was superior to
AM1: Zheng Y-J, Merz KM (1992) J Comp Chem 13:1151
Easier Questions.............................................................
- Outline the similarities and differences between the extended H€uckel method
on the one hand and methods like AM1 and PM3 on the other. What advantages
does the EHM have over more accurate semiempirical methods? - Outline the similarities and differences between molecular mechanics, ab
initio, and semiempirical methods. - Both the simple H€uckel and the PPP methods arepelectron methods, but PPP
is more complex. Itemize the added features of PPP. - What is the main advantage of an all-valence-electron method like, say, CNDO
over a purelypelectron method like PPP? - Explain the terms ZDO, CNDO, INDO, and NDDO, showing why the latter
three represent a progressive conceptual improvement. - How does an AM1 or PM3 “total electron wavefunction”Cdiffer from theC
of an ab initio calculation? - Ab initio energies are “total dissociation” energies (dissociation to electrons
and atomic nuclei) and AM1 and PM3 energies are standard heats of formation.
Is one of these kinds of energy more useful? Why or why not? - For certain kinds of molecules molecular mechanics can give better geometries
and relative energies than can even sophisticated semiempirical methods. What
kinds of properties can the latter calculate that MM cannot? - Why do transition metal compounds present special difficulties for AM1 and
PM3? - Although both AM1 and PM3 normally give good molecular geometries, they
are not too successful in dealing with geometries involving hydrogen bonds
(PM3 seems to be the better one here). Suggest reasons for this deficiency.
Harder Questions.............................................................
- Why are even very carefully-parameterized semiempirical methods like AM1
and PM3 not as accurate and reliable as high-level (e.g. MP2, CI, coupled-
cluster) ab initio calculations?
Harder Questions 443