Screening the fabrication conditions of ultrafiltration membranes by using the uniform design and regression analysis methods

Suei Yee Lam Leung, Wing Hong Chan, Chi Hung Leung, Chi Hang Luk

Research output: Contribution to journalJournal articlepeer-review

23 Citations (Scopus)

Abstract

In this work, the concerned casting conditions for the fabrication of the ultrafiltration (UF) membrane from polysulfone were the gelation medium used, the temperature of the gelation bath, the gelation period, and the evaporation time of the cast sheet. The uniform design method was applied to substantially reduce the number of experiments for studying the influence of each of four casting conditions introduced in the preparation of UF- membranes. The pure water flux and the molecular weight cutoff (MWCO) of each of the four macromolecule-standards, namely egg white lysozyme, egg white albumin, bovine serum albumin and bovine γ-globulius, of each prepared UF- membrane were determined. These experimental responses were then analyzed by using a non-linear model with the backward regression method. The results of this work showed that the uniform design-non linear model (UD-NLM) method can substantially reduce the number of experimental runs for predicting the influences of the studied casting conditions on the ultrafiltration performance of the prepared asymmetric polysulfone membranes.

Original languageEnglish
Pages (from-to)203-213
Number of pages11
JournalChemometrics and Intelligent Laboratory Systems
Volume40
Issue number2
DOIs
Publication statusPublished - 24 Jun 1998

Scopus Subject Areas

  • Analytical Chemistry
  • Software
  • Process Chemistry and Technology
  • Spectroscopy
  • Computer Science Applications

User-Defined Keywords

  • Lithium bromide
  • Non-linear regression analysis method
  • Polysulfone membrane
  • Ultrafiltration membrane
  • Uniform design method

Fingerprint

Dive into the research topics of 'Screening the fabrication conditions of ultrafiltration membranes by using the uniform design and regression analysis methods'. Together they form a unique fingerprint.

Cite this