Characterization of Carbon Molecular Sieve Membranes Supported on Ceramic Tubes
K. Briceño, J. Silvestre-Albero, A. Silvestre-Albero, J.I. Calvo, D. Montane, R. García-Valls, A. Hernández, F. Rodríguez-Reinoso
2013
Adsorption Science & Technology
Carbon molecular sieve membranes have been analyzed in supported and unsupported configurations in this experimental study. The membranes were used to adsorb CO 2 , N 2 and CH 4 , and their adsorption data were analyzed to establish differences in rate and capacity of adsorption between the two types of samples (supported and unsupported). Experimental results show an important effect of the support, which can be considered as an additional parameter to tailor pore size on these carbon
more »
... . Immersion calorimetry values were measured by immersing the membranes into liquids of different molecular dimensions (dichloromethane, benzene, n-hexane, 2,2dimethylbutane). Similarities were found between adsorption and calorimetric analysis. The pore volume of the samples analyzed ranged from 0.016 to 0.263 cm 3 /g. The effect of the pyrolysis temperature, either 550 or 700 °C, under N 2 atmosphere was also analyzed. Quantification of the pore-size distribution of the support was done by liquid-liquid displacement porosimetry. The composite membrane was used for CO 2 /CH 4 separation before and after pore plugging was done. The ideal selectivity factors value (4.47) was over the Knudsen theoretical factor (0.60) for membrane pyrolyzed at 600 °C, which indicates the potential application of these membranes for the separation of low-molecular weight gases. (K. Briceño). †Published in the Festschrift of the journal dedicated to Professor K.S.W. Sing to celebrate his 65 years of research in the field of adsorption. those involving temperature and elevated partial pressure of water. Furthermore, synthesis of molecular sieves based on carbon is simpler and cheaper than those based on other materials such as zeolites, whose methods are intricate and require additional post-treatments steps (Foley 1995) . Beyond their use as adsorbents, CMS have been explored in the past for membrane applications (CMSM), building on the pioneering works of Rao (Rao and Sircar 1996) and Koresh (Koresh and Soffer 1987) . CMSM can be produced in different geometries-planar, hollow fibre and supported filmsby pyrolysis of polymers at temperatures between 550 and 1000 °C (Saufi and Ismail 2004). The selection of membrane presentation depends on the final application of the membrane. Supported membranes are asymmetric structures that are preferred over the unsupported ones owing to their superior mechanical properties. However, even when using the same precursor, membrane properties can largely vary depending on the preparation procedure used. For example, in their attempts to reduce the number of coatings, Fuertes and Centeno (1999) reported differences between supported membranes obtained after three carbon-coating cycles and those obtained after deposition of three layers of the same precursor polyimide with a subsequent carbonization step. The latter sample exhibited higher permeation rates but lower selectivity compared with the former one. It is obvious that the way in which the carbon structure is formed depends on the coating procedure, which is responsible for differences in the properties of the carbon layer. Singh-Goshal and Koros (Singh-Ghosal and Koros 2000) outlined that it is easier to compute entropic selectivity on dense symmetric films than on asymmetric membranes. Identification of the carbon structure becomes more complex if variables that are often considered to affect the carbon structure of flat membranes have to be considered on asymmetric configurations. Vu et al. (2002) confirmed an unexpected decrease in the CO 2 permeance values for CMS fibres (asymmetric membranes) after pyrolysis. On the contrary, permeance increased in planar films made from the same polymer (unsupported) following the same pyrolysis conditions. Apparently, when the polymer precursor is pyrolyzed in an asymmetric structure, as is the case of hollow fibre or supported samples, the carbonaceous structure obtained after pyrolysis becomes different from those samples obtained in a planar shape and dense film. It could be hypothesized that the presence of the support could determine differences in carbon structure because of the differences in the properties of the carbon layer. The reasons explaining this difference have been scarcely explored in the past because it is not easy to characterize the changes in pore structure of these materials. In fact, characterization of unsupported carbon films has been used in the past to explain the morphological changes of its supported counterparts (Fuertes and Centeno 1999; Sedigh et al. 2000) . However, little evidence has been shown in the past to understand the extent of this extrapolation. In addition, excluding permeance analysis, there are few studies that match the characterization of the overall supported membrane and its isolated carbon supported layer. The aim of this work is to identify structural differences and similarities between carbon samples obtained in supported tubes and those obtained in unsupported planar configuration. The effect of coating layer on modification of a ceramic support has been explored using liquid-liquid displacement porosimetry (LLDP), a characterization technique that can distinguish active pores, thus contributing to actual flux, from those that do not actually inter-connect both sides of the membrane. Although it is not an objective of this work to propose a mechanism of carbon formation, a correlation has been found between the characteristics of the composite membrane and the structure of the carbon layer obtained. 234 K. Briceño et al. Figure 3 . CH 4 and CO 2 adsorption-desorption isotherms at 0 °C for the different carbon non-supported membranes obtained at 550 and 700 °C pyrolysis temperatures.
doi:10.1260/0263-6174.31.2-3.233
fatcat:v442c36jtbg23khyejyjlahrui