Applications of topological indices to structure-activity relationship modelling and selection of mineral collectors
Ramanathan Natarajan, Palanichamy Kamalakanan, I. Nirdosh · INDIAN JOURNAL OF CHEMISTRY- SECTION A · 2003
Topological indices (TI) are ex tensively used as molecul ar descriptors in building Quantitative Structure-Activity Relationships (QSAR), Quantitat ive Structure-Property Re lati onships (QSPR) and Quantitative Structure-Toxicity Relationships (QSTR). We have appli ed th e QSAR approach for the first time to froth flotation and :nodelled the separation effic iencies (Es) of four seri es of coll ectors namely, cupferrons, o-aminothiophenols, merc aptobenzothiozole s and ary lhydroxamic acids used for the beneficiation of uranium, lead, zinc and copper ores, respectively. Principal components extrac ted from topological indices are found to form linear regression equati ons th at can predict the separati on efficienci es of the first three cla sses of coll ectors. Inclusion of substituent constant or quantum mechanical parameters did not improve th e fit and th is may be due to the absence of large variation in electronic effects among th e molecules in each class. A quadratic fit has been obtain ed instead of linear regression in the case of arylhydroxamic acids. The separati on efficienci es -of the members have been predicted withi n experimental error (±5%) by the best regression equations in each series. From the statistica l models built, it may be generalized that Es = f(TI). To ex tend the applicati ons of TI in the field of mineral processing, a simi larity space of 587 ary lh ydroxamic acids has been constru cted usi ng principal components extracted from Tis. Ten compounds have been selected for synthesis and testin g. The approach appears to provide an efficient sc ientifi c tool for the selection of coll ectors from a vi rtual database and may lead to an inexpensive and quick method for th e identification of new collectors.