Methanol vapour permeation

Kunioka K, Shugen Z, Okita K, Kakitani T, and Yazawa T. Permeation of methanol vapour through siUca membranes prepared by the eVD method with aid of evacuation. J. Membr. Sci. 1999 160 31-39. 

Figure 4.8 Permeation curves of dichloromethane and methanol vapour in Hyflon AD80X membranes at 25 °C. Time lag calculation by the tangent method (top) and by a direct least squares fit of the entire permeation curve according to Eq. (4.6) (bottom). The thick dark line represents the experimental data the thin brighter line gives the tangent (top) or the least squares fit. The experimental and fitted lines superimpose completely in the case of DCM, and only one curve can be distinguished, whereas MeOH gives a poor fit. See text for further explanation...

In this work preliminary vapour permeation measurements were carried out with two different species, the rather bulky dichloromethane (DCM) molecules and the much smaller methanol molecules. Two typical permeation curves are displayed in Figure 4.8. The transport parameters, determined on the basis of the tangent method and Equations (4.9)-(4.11), are listed in Table 4.3. It contains the parameters dehned above as well as solubility C in the membrane in equilibrium with the feed pressure of penetrants. 

Figure 4.9 Left side Experimental permeation curve of methanol vapour in the Hyflon AD80X membrane of Figure 8 (thick dark line). The shaded areas and the thin brighter lines represent the different steps of the fitting procedure according to Eq. (4.6). Right side, B, C, D experimental data or residual experimental data (noisy dark lines), with the corresponding fit of the individual steps in the three different time intervals (thin brighter lines). The sum of the three fits coincides perfectly with the experimental data. In most cases the fit is nearly perfect and only one curve can be distinguished. See further explanation in the text...

Silica layered Nation membranes prepared by olasma enhanced chemical vapour deposition [45] also exhibit a moderate reduction of permeability (F, =0.33) by those membranes fabricated by dip-coating [38] show extremely high barrier properties to methanol permeation, with 0.002 < P < 0.005. However, it should be noted that the permeability reported for the Nation 117 without silica layers is more than one order of magnitude higher than the values reported by several authors, including those shown in Fig. 6.18. 

More careful analysis of the permeation curves show that the two vapours actually display completely different behaviour. The tangent method can be applied without problems to the DCM permeation curve, resulting in a time lag of ca. 400 s. hi contrast, this method shows that methanol has an unusually wide transient period. While the extrapo-... 

A commercial nickel catalyst was used for methane steam reforming performed at a 500 °C reaction temperature, a S/C ratio of 3.0 and atmospheric pressure, while the permeate side was evacuated. The performance of the vapour deposited platinum membrane was similar to the plated dense palladium membrane. In the permeate of the deposited ruthenium and palladium membranes, small amounts of carbon oxides and also methane were observed. While it was expected that all these species had passed through the membranes by diffusion, in addition some methane was converted into carbon dioxide over the noble metals of the membranes. Kikuchi et al. demonstrated by simulations that conversion and hydrogen permeation in a membrane reactor is higher, where the first portion of the catalyst bed is not coupled to the membrane. Such an arrangement as shown in Figure 7.16 would clearly save expensive membrane surface area. Experimental work by Itoh et al. performed for methanol steam reforming [521] confirmed the assumptions of Kikuchi et al. 

---