This preview shows page 1. Sign up to view the full content.
Unformatted text preview: air to its mole fraction xa in water. The kH is a key parameter for assessing the distribution of trace organic compounds among the various environmental media (e.g. air, water, soil, sediment). However, determination of k H is difficult particularly for organic compounds with very low vapor pressures or low aqueo us solubilities. In this work, we generated quantitative structure -property relationship (QSPR) models for predicting Henry’s law constant (kH) of hydrocarbons and halogenated hydrocarbons using connectivity indices (CI) as predictors. A total of 177 organic compounds were used (152 in training set and 25 in test set) for QSPR modeling by stepwise multiple-linear regression (MLR) and artificial neural network (ANN). The MLR 2 model generated has 9 constitutional descriptors that have significant influence to kH (R = 0.924). The 2 MLR model showed high prediction performance (Q = 0.928) when applied to the test set (N = 25). However, more accurate results were obtained when ANN -based QSPR models were used to predict the 2 kH (Q > 0.950, N = 25). 113 23rd-25th June 2010 3rd ICYC 2010 Universiti Sains Malaysia ENV-P23 QSPR PREDICTION OF SOLUBILITY OF ORGANIC COMPOUNDS IN WATER USING CONSTITUTIONAL AND TOPOLOGICAL DESCRIPTORS Ronald A. Simporios, Joanna Marie A. Dimailig, Jay-Ar Carlo S. Aurelio, Marie Joy Rellona, Ma. Jocel L. Hernandez, Elisa A. Molinyawe, Jennifer C. Sy, and *Alex A. Tardaguila Department of Physical Science, Pamantasan ng Lungsod ng Maynila, Gen. Luna St., Intramuros 1002, Manila, Philippines E-mail: email@example.com The solubility of a substance is defined as the concentration of dissolved solu te in equilibrium with undissolved solute at a specified temperature and pressure. The study of solubility is important for a wide variety of processes, including oxygen transport in the blood, drug interactions, environmental pollution, and industrial processes. In this work, quantitative structure-property relationship (QSPR) method was employed to model the solubility values (expressed as LN Sol) of 103 organic compounds. Stepwise multiple-linear regression (MLR) was applied to the data...
View Full Document
- Spring '11