(573b) Characterization of Deteriorated Zeolite Membrane Used in the Dehydration of Commercial Bio-Ethanol Plants | AIChE

(573b) Characterization of Deteriorated Zeolite Membrane Used in the Dehydration of Commercial Bio-Ethanol Plants

Authors 

Nakane, T. - Presenter, Mitsubishi Chemical Corporation
Sato, K. - Presenter, Mitsubishi Chemical Corporation
Saito, J. - Presenter, Mitsubishi Chemical Corporation
Shimotsuma, N. - Presenter, Mitsubishi Chemical Corporation
Sugimoto, K. - Presenter, Mitsubishi Chemical Corporation
Kurata, T. - Presenter, Mitsubishi Chemical Corporation


Tubular zeolite NaA (LTA) membrane has been a topic of much research in recent years, as a possible means of performing low-cost and energy saving dehydration processes of bio-ethanol [1-4]. Several studies have been conducted on LTA membranes hydrothermally synthesized on an alumina porous tube for use in vapor permeation (VP) dehydration process of bio-ethanol [5, 6]. Recently, these membranes have been used in large commercial-scale and pilot-scale bio-ethanol dehydration plants [7-10]. The biggest subject in commercial plant is to clarify deterioration mechanism and lifetime of the membrane. Once the deterioration mechanism of the LTA membranes in the VP process is clarified, then it may be possible to improve the process to avoid such deterioration. This also would result in a reduction in the running cost of a dehydration plant. For this purpose, structural characterization of the LTA membrane before and after the VP process is very important. However, most of the past works have not dealt with this kind of problem from the view point of structural change and fouling of the membrane during the VP process. The main reason for this insufficient analysis lies in the difficulty of accurate physicochemical analysis of tubular-form LTA membrane.

The deterioration of the membrane is mainly caused by crystalline structure change and physical and chemical fouling of the tubular membrane. For the analysis of crystalline structure change, the methods we applied are high-resolution parallel-beam optics based X-ray diffraction analysis (XRD) and grazing incidence one (GIXRD) [11, 12]. For the analysis of physical and chemical fouling, we mainly applied Fourier Transform Infrared Attenuated Total Reflectance method (FTIR-ATR) with a diamond prism as the waveguide [13-15], transmission electron microscopy (TEM) combined with a focused ion beam (FIB) cross-section thin-layer specimen preparation technique [16-18], scanning electron microscopy (SEM) equipped with an energy dispersive X-ray spectrometer (EDX) and X-ray fluorescence analysis (XRF). Based on these analytical studies, this work first demonstrates the universal deterioration mechanism of LTA membrane, which is commonly observed in the VP dehydration process of sugarcane-based bio-ethanol.

In the dehydration process of hydrous fermented ethanol to match fuel grade specification, we observed two kinds of deteriorations by different reactions. There are two deteriorations: (1) Specific reaction in the crystal structure of LTA zeolite. (2) Fouling caused by the deposition of raw materials-derived substances.

(1){Crystal structure}

LTA zeolite is a typical ion-exchanger, so that sodium ion in the crystal structure was easily exchanged with other cation in the early stage of the VP running to become equilibrium with the feed (up to 50~300 hrs from the start of the experiment), and then reach equilibrium. The initial reaction took place through the sodium exchange with proton in the feed solution; the amount of exchanged sodium was found to be around 10 %. Preferential dealumination in the framework did not occur during the long term VP process [19]. Lattice constants increased as a consequence. As a result, a small decline in water permeance (Kw) was observed in the early stage of the VP process. Through the detailed examination of the Kw and crystalline structure change, the strong correlation between these two was observed. Although use of calcium-exchanged LTA zeolite may improve the initial deterioration, its Kw value is much lower than that of sodium-type LTA.

(2){Fouling}

Prior to distillation and VP dehydration processes, fermented hydrous ethanol is normally neutralized by using sodium hydroxide, so that large amounts of sodium acetate/sulfate sediment in bottom of feed tank. Usually, these insoluble salts are removed by demister or super-heater to some extent, but insufficiently. Through the long-term VP operation, the sodium acetate particles make many dimples on the membrane surface, and sodalite-like substances or sodium sulfate secondarily grows from surface of LTA crystal. Also, water condensation on the membrane surface may cause damage of amorphous grain boundary [20]. Crack and/or hole on the membrane surface are caused by these fouling phenomena. From the detailed analysis of these deteriorated membranes, the importance of pretreatment before membrane process became very clear.

In conclusion, the universal deterioration mechanism of the LTA membranes used in the commercial-scale plants was first reported.

[References] 1) Y. Morigami et al., Sep. Purif. Tech., 25, 251 (2001). 2) K. Okamoto et al., Ind. Eng. Chem. Res., 40, 163 (2001). 3) M. Kondo, J. Vac. Soc. Jpn., 49, 225 (2006). 4) S.-L. Wee et al., Sep. Purif. Tech., 63, 500 (2008). 5) K. Sato et al., J. Membr. Sci., 301, 151 (2007). 6) T. Mizuno et al., ZMPC2006, OA109, Yonago, Japan. 7) K. Izumi et al., Kagakukougaku, 71, 812 (2007). 8) K. Aoki et al., MEMBRANE, 32, 234 (2007). 9) K. Sato et al., Micropor. Mesopor. Mater., 115, 184 (2008). 10) M. Kondo, Zeolite News Letters, 25, 93 (2008). 11) T. Kyotani, Anal. Sci., 22, 961 (2006). 12) T. Kyotani et al., Bunseki Kagaku, 57, 339 (2008). 13) T. Kyotani et al., Anal. Sci., 21, 321 (2005). 14) T. Kyotani et al., Anal. Sci., 22, 325 (2006). 15) T. Kyotani, Catalysts & Catalysis, 51, 239 (2009). 16) Z. Liu et al., Chem. Mater., 18, 922 (2006). 17) T. Kyotani et al., Anal. Sci., 22, 1031 (2006). 18) T. Kyotani et al., J. Membr. Sci., 296, 162 (2007). 19) T. Kyotani et al., submitted for publication. 20) Y. Li et al., J. Membr. Sci., 297, 10 (2007).