%0 Journal Article
%T Impact of Measuring Devices and Data Analysis on the Determination of Gas Membrane Properties
%J Journal of Membrane Science and Research
%I FIMTEC & MPRL
%Z 2476-5406
%A Wu, Haoyu
%A Kruczek, Boguslaw
%A Thibault, Jules
%D 2018
%\ 01/01/2018
%V 4
%N 1
%P 4-14
%! Impact of Measuring Devices and Data Analysis on the Determination of Gas Membrane Properties
%K Gas Membrane Characterization
%K Time-lag
%K Pressure Transducer
%K Data Variability
%K Nonlinear Least Squares
%R 10.22079/jmsr.2017.63433.1136
%X The time-lag method, using a gas permeation experiment, is currently the most popular method for determining the membrane properties: diﬀusivity coefcient and permeability coefcient, and from which the solubility coefcient can be calculated. In this investigation, the impact of systematic, random (noise), resolution and extrapolation errors associated with gas permeation experiments on the determination of the membrane properties using the time-lag method is investigated. A comprehensive error analysis for each type of errors and their combination is presented. Random and resolution errors have a greater impact on the determination of the time lag for low rates of downstream pressure accumulation which can be alleviated by increasing the capacity parameter. Increasing the feed pressure lowers the resolution errors, but has no eﬀect on random errors. Extrapolation errors associated with the time-lag method, which increase with time, can be reduced by increasing the number of evaluation points and the length of the evaluation window. Because of their strong correlation, it is difcult to decouple solubility and diﬀusivity coefcients accurately without using the time-lag. A judicious balance between data precision, the drop in the driving force and the duration of an experiment must be considered in the design of a constant-volume membrane system and in the selection of experimental operating conditions to minimize the impact of pressure variability. The necessity of a small capacity parameter for the accurate determination of membrane properties needs to be reconsidered in the presence of experimental noise.
%U http://www.msrjournal.com/article_26198_dd87a0903c1adbfbc92a9605a71377fe.pdf