Prediction of Reduced Ion Mobility Constants from Molecular Structure

Abstract

The specific goal of this project was to develop mathematical models to predict the reduced ion mobility constants, Ko values, for organic compounds directly from their molecular structures. These models are generated by a three-step procedure that involves the representation of the compounds by calculated molecular structure descriptors, selection of the most important descriptors, and the subsequent development of the models using computational neural networks. We have completed and published a high quality model for the prediction of Ko values for monomer ions of 168 compounds using a 6-4-1 (6 input, 4 hidden, and 1 output neuron) computational neural network model. A subset of 93 compounds which exhibited good dimer ion peaks was used to develop a successful 4-2-1 CNN model. A study of phosphorus-containing compounds was also successfully completed. The significance of this work is that it provides fundamental information for ion mobility spectrometry, a sensitive analytical technique used to detect chemical warfare agents in the field.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 1998
Accession Number
ADA365190

Entities

People

  • Peter C. Jurs

Organizations

  • Pennsylvania State University

Tags

Communities of Interest

  • Ground and Sea Platforms

DTIC Thesaurus Topics

  • Chemical Compounds
  • Chemistry
  • Computational Science
  • Computer Programs
  • Computers
  • Data Analysis
  • Data Sets
  • Mathematical Models
  • Mobility
  • Models
  • Molecular Structure
  • Neural Networks
  • Organic Compounds
  • Phosphorus
  • Spectrometry
  • Students
  • Three Dimensional

Fields of Study

  • Chemistry

Readers

  • Computational Modeling and Simulation
  • Organic Chemistry
  • Quantum Chemistry

Technology Areas

  • AI & ML
  • AI & ML - Neural Networks