Skip to main navigation Skip to search Skip to main content

Separation of dichloromethane-nitrogen mixtures by adsorption: Experimental and molecular simulation studies

  • Carolyn A. Koh
  • , Robin E. Westacott
  • , Robert I. Nooney
  • , Viviane Boissel
  • , Saad F. Tahir
  • , Valeria Tricarico

Research output: Contribution to journalArticlepeer-review

Abstract

Experimental and grand canonical Monte Carlo simulation results for the separation of a CH2Cl2 (1.5 mol%)-N2 binary gas mixture in molecular sieve materials are presented. AlPO4-5 and MCM-41 molecular sieves have been used as the selective adsorbents because they consist of uniform arrays of uni-dimensional channels of micro and meso length scales, respectively. Adsorption isotherms were measured at 318 K and at pressures between 50 kPa and 130 kPa. Two MCM-41 materials have been used, one with a 33 Å pore diameter and the other with a 42 Å pore diameter. For AlPO4-5 at 110kPa the total amount adsorbed from experiment was found to be independent of equilibration time at 0.0542, 0.0538 and 0.0547 mmol per g AlPO4-5 for 2, 24 and 48 hours, respectively. However, the selectivity for CH2Cl2 was found to increase with time from 1.29, to 4.59, to 10.74. For MCM-41 at 110 kPa the selectivity for CH2Cl2 was found to be dependent on pore size. On increasing the pore size from 33 Å to 42 Å the selectivity for CH2Cl2 increased considerably. Grand canonical Monte Carlo simulations agreed qualitatively with the experimental results, showing a greater selectivity for CH2Cl2 than for N2. The simulations indicate that MCM-41 has a lower selectivity for CH2Cl2 than AlPO4-5, which contradicts the experimental results. Reasons for these discrepancies are presented and discussed.

Original languageEnglish
Pages (from-to)2087-2095
Number of pages9
JournalMolecular Physics
Volume100
Issue number13
DOIs
Publication statusPublished - 10 Jul 2002

Fingerprint

Dive into the research topics of 'Separation of dichloromethane-nitrogen mixtures by adsorption: Experimental and molecular simulation studies'. Together they form a unique fingerprint.

Cite this