I. Motoc - Böcker

Although the importance of steric fit for receptor-effector 1 interactions was recognized since Emil Fischer proposed his "lock and key" theory, the whole area of steric properties is still in a very 2-4 early stage of development. We have a fairly good idea about el- tronic and hydrophobic parameters, but it is not easy to describe ste- ric shapes of molecules without a large number of data. There are se- veral cases of good QSAR's developed for rather large series of mole- 5 cules without steric parameters - for example see papers by Hansch , 6 or Franke , but the state of steric parameters is nevertheless one of the most important drawbacks, especially concerning the ability of en- compassing, within a single QSAR, molecules of different shapes and stereoisomers. From today's steric parameters, one may mention the 7 Taft parameters E ' which gave good results in organic chemistry, the S 8 10 ra th er cum b ersome way 0 f measurIng * s h ape d'ff I ere h ces 0 f Amoore - and , 11 12 AllInger ,and the L, B -B parameters of Verloop 1 4 The work described here consists of two types of approaches to the steric fit problem.The first approach consists of developing new parameters to describe different characteristics of the molecular shape (i. e. , branching, bulkiness); this is done by means of topological in- dices.

The aim of this chapter is to discuss in detail the Monte Carlo algorithms developed to compute the sequence distributions in polymers. Because stereoregular polymers constitute a unique form of copolymer, the stereosequence distributions in vinyl homopolymers and the sequence distributions in copolymers can be computed using the same algorithms. Also included is a brief review of probabilistic models (i. e. , Bernoulli trials and Markov chains) frequently used to compute the sequence distribtuion. The determination of sequence distributions is important for the under- standing of polymer physical properties, to compute the monomer reactivity para- meters and to discriminate among polymerization mechanisms. 2. 2. Short review of analytical models, Monte Carlo algorithms and computer programs. l A Bernoullian model was developed by Price. Within this model the probability of a given state of the system is independent of the previous state and does not condition the next state.The Bernoullian behaviour has been shown 24 to describe cls-trans distributions among 1, 4 additions in polybutadienes - , 5 the comonomer distribution in ethylene-vinyl acetate copolymer , and configura- 6 tional distributions in polystyrene , poly (vinyl chloride)7, poly (vinyl alcohol)7 Consider the binary copolymerization:;1,J=1,2 (1) where - MI* , I = 1,2, is an ionic or radical polymeric chain end, and M, J = 1,2, J is a monomer. Because the final state (i. e.