Microsoft Word - 476hernandez.docx CHEMICAL ENGINEERING TRANSACTIONS VOL. 43, 2015 A publication of The Italian Association of Chemical Engineering Online at www.aidic.it/cet Chief Editors: Sauro Pierucci, Jiří J. Klemeš Copyright © 2015, AIDIC Servizi S.r.l., ISBN 978-88-95608-34-1; ISSN 2283-9216 Adsorption Properties and Permeation Performances of DD3R Zeolite Membranes Alessio Caravella*a, Pasquale F. Zitoa, Adele Brunettib, Enrico Driolia,b, Giuseppe Barbierib aUniversity of Calabria, Department of Environmental and Chemical Engineering (DIATIC), Via Pietro Bucci, Cubo 44A, Rende (CS), 87036, Italy. bNational Research Council - Institute on Membrane Technology (ITM-CNR), Via Pietro Bucci, Cubo 17C, Rende (CS), 87036, Italy. alessio.caravella@unical.it This modelling paper is focused on the evaluation of the performance of DD3R zeolite membranes for light gases separation. First, the required adsorption properties are calculated by means of a multivariate regression. Then, the Maxwell-Stefan multicomponent transport equations are implemented considering a quaternary equimolar gas mixture (CH4, CO2, CO and N2) on the high-pressure feed side to evaluate the actual permeance and selectivity of the considered gases keeping the low-pressure permeation side at 100 kPa. After validating the model through experimental data taken from the literature, the simulation is performed in a certain range of temperature and pressure. It is found that in the range of considered conditions, the CO2 / CH4 selectivity is higher than 100, increasing with decreasing temperature and pressure. Therefore, the most convenient temperature and pressure conditions where permeance and CO2 / CH4 selectivity are good enough to be of interest for industrial applications are identified. 1. Introduction Membrane technology represents a powerful alternative to more traditional and more energy consuming gas separation processes like, for example, Pressure Swing Adsorption (PSA). Among all types of membranes, the zeolite ones have been nowadays gaining more and more importance in gas separation thanks to their particular crystalline ceramic nanostructures, which are able to selectively adsorb molecules and separate them by adsorption equilibrium, diffusion kinetics and/or molecular sieving. Furthermore, zeolite membranes show a good thermal and mechanical stability. A crucial factor affecting the separation performance of a zeolite membrane is the composition of the framework, with a particular regard to the Si-Al ratio. For example, zeolites with low Si-Al ratios are suitable for water separation, whereas membranes with high Si-Al ratios are suitable for gas separation thanks to their low content of defects (van den Bergh, 2010a). In this paper, DD3R (deca-dodecasil 3R) membranes are considered because of their promising performance in gas separation. In particular, they can be used to separate several gas mixtures like CO2/CH4, propane/propene and water/ethanol (Zhu et al., 1999; Gascon et al., 2008; Tomita et al., 2004; van den Bergh et al., 2008a, 2008b, 2010a; Kuhn et al., 2008). The pore size of DD3R (0.36 x 0.44 nm), which is smaller than that of other zeolites like Silicalite and FAU (Faujasite), makes this zeolite very interesting for light gases separation. The transport through zeolite membranes is mostly driven by the surface diffusion, which can be effectively described by the multicomponent Maxwell-Stefan approach, as reported in several literature works (van den Bergh et al., 2007, 2008b, 2010a, 2010b; Lee, 2007; Kangas et al., 2013; Kaptejin et al., 2000; Wirawan et al., 2011; Bakker et al., 1996; Krishna and Wesselingh, 1997; Li et al., 2007; Krishna and Paschek, 2000). DOI: 10.3303/CET1543180 Please cite this article as: Caravella A., Zito P.F., Brunetti A., Drioli E., Barbieri G., 2015, Adsorption properties and permeation performances of dd3r zeolite membrane, Chemical Engineering Transactions, 43, 1075-1080 DOI: 10.3303/CET1543180 1075 The implementation of such an approach requires the knowledge of several adsorption properties, like saturation loading, heat of adsorption and affinity constant. For this reason, the first part of this paper is dedicated to the evaluation of the required adsorption properties, obtained by a non-linear regression analysis of several experimental isotherms by an optimization routine using the Levenbergh-Marquardt algorithm to fit the data through the Langmuir model. By doing that, the functionality of the adsorption parameters with temperature is achieved. Then, the adsorption parameters are used as input data in the simulation of the permeation of a quaternary equimolar gas mixture (CH4, CO2, CO and N2) through a DD3R membrane adopting the multicomponent approach described below. 2. Description of the system The system considered in this investigation is a cross-section of a tubular membrane composed of a thin DD3R layer (60 μm) deposited on a porous support (Figure 1). The high-pressure side is located on the zeolite layer side, whereas the low-pressure one is on the support side. The characteristic pore size of support is supposed to be sufficiently large to provide a negligible effect to mass transfer. 3. Mathematical approach Langmuir model is used to fit the collected experimental isotherms. Eqs(1-5) provide the details of the single- component Langmuir isotherms (Do, 1998): ( ) ( )Pb Pb CC L L + = 1 μsμ ,       = ∞ RT Q bb LLL exp, , T b b L 0 , =∞ , MRk a b d π2 0 ∞ = ,               −= 0 μs,0μs 1exp T T CC χ (1-5) 3.1 Maxwell-Stefan approach to mass transport According to the Maxwell-Stefan theory, the driving force for mass transport, given by the chemical potential gradient, is balanced by the friction that a single species with a certain velocity has with zeolite and the other species in mixture (Kaptejin et al., 2000; Krishna and Wesselingh, 1997; Krishna and Paschek, 2000): iis i ijjsis jii n j i i DC N DCC NCNC RT j μμμ μμμ ρθ + − = ∇ −  =1 (6) The Maxwell-Stefan surface diffusivity Di describes the interactions between the ith species and zeolite, whereas the exchange diffusivity Dij describes the interactions between two generic species. Eq(6) can be represented using matrices, from which the form of the flux vector can be obtained as expressed in Eqs(7-9). )]([][)( 1 μ ρ μ ∇−= − CB RT N ,  = += n j ijs j i ii DC C D B j 1 1 μ μ , ijs j ij DC C B jμ μ −= ( )θ−= 10)(ii DD , sC C μ μθ = (7-11) For the mathematical closure of the system, the Maxwell-Stefan equations (Eqs(7-9)) are coupled to the mass balance of the ith species through the zeolite layer along with the appropriate boundary conditions. The system considered for simulation is shown in Figure 1. Notation Table Feed Permeate (100 kPa) Zeolite Layer Support F lu x Figure 1: Simulated System N Molar flux, mol m -2 s -1 P Pressure, Pa QL Heat of adsorption, J mol -1 T Temperature, K T0 Reference temperature, K χ Empirical parameter in Eq 5, - θ Fractional surface occupancy, - μ Chemical potential, J mol-1 ρ Density of zeolite, kg m-3 a Sticking coefficient, - bL Langmuir affinity constant, Pa -1 Cμ Molecular loading, mol kg -1 Cμs,0 Saturation molecular loading at T0, mol kg -1 Di(0) Zero-loading molecular diffusivity, m 2 s -1 Di Maxwell-Stefan self diffusivity, m 2 s -1 Dij Maxwell-Stefan exchange diffusivity, m 2 s -1 kd∞ Desorption rate constant, mol s -1 m -2 M Molar mass, kg mol-1 1076 The boundary conditions are given in terms of fixed pressures on both membrane sides. The simulation hypotheses are: (1) negligible effect of porous support; (2) negligible effect of external mass transfer; (3) single loadings described by Langmuir isotherm; (4) strong confinement scenario. The hypothesis of strong confinement scenario corresponds to a linear relation between the diffusivity Di with the occupancy degree θ (Lee, 2007). 4. Results and Discussion 4.1 Adsorption Properties The optimal values of some key-adsorption properties (saturation loading and heat of adsorption) are reported along with their respective confidence intervals at 95% (Figure 2a and b). Figure 2a shows quite narrow confidence intervals of the optimal values of the saturation loadings for all species, indicating the good calculation accuracy. The confidence intervals for the heat of adsorption (Figure 2b) are sufficiently narrow (within around ±10%) for all species but hydrogen, for which the large calculation uncertainty is due to the weak hydrogen-zeolite physical bond. In Table 1, optimal values of parameters and corresponding confidence intervals are reported for CO2. 4.2 Simulation results Before simulation, the model was validated collecting some experimental data from the literature in terms of CO2 permeability and selectivity with respect to CH4 (van den Bergh et al., 2008b), running simulations in the same conditions as those reported in the paper. The model predicts quite well the experimental behaviour at a temperature higher than around 10°C (3), which are of direct interest for industrial applications. In particular, the CO2 permeability decreases with increasing temperature, this owing to the weaker adsorption strength on which surface diffusion is based. The calculated permeability and selectivity trends are reported in Figure -6. Figure a shows that the CH4 permeability decreases with increasing pressure, as well as that of CO2 (Figure b), CO (Figure a) and N2 (Figure b). Such a common trend is due to the increasing species-zeolite interactions caused by the higher concentration in the pores. However, for CH4, an increase of the feed pressure does not significantly affect permeability. Feed pressure has a stronger influence on the permeability of the other components, especially at the lower pressures (up to ≈ 600 kPa). Differently, permeability shows an increasing trend with temperature for CH4, CO and N2. As for CO2, its trend with temperature exhibits a maximum (Figure a), owing to the contrasting effects of adsorption and diffusion with temperature. Moreover, the temperature at the maximum is found to increase at higher pressure, changing from 293 K at 200 kPa to 313 K at 1000 kPa. Comparison among permeabilities shows that the less polar molecules (N2 and CH4) pass slower than the more polar ones (CO2 and CO). This can be due to the preferentially CO2 and CO adsorption into the zeolite layer, this representing a barrier for adsorption and consequent diffusion of N2 and CH4. Figure shows that the CO2/CH4 selectivity decreases with both feed pressure and temperature, keeping values between 100 and 1000. These results are also observed by van den Bergh et al. (2008b), but, additionally, our work indicates temperature and pressure range for an effective CO2/CH4 separation. H e a t o f A d so rp tio n, J m o l-1 Figure 2: Calculated values of (a) saturation loadings and (b) heat of adsorption. 1077 0 200 400 600 800 1000 Feed Pressure, kPa 0 2 4 6 8 10 CO2 333 253 K 293 K T, K Figure 5: (a) CH4 and (b) CO2 permeability vs. feed pressure at different temperatures. Figure 6: (a) CO and (b) N2 permeability vs. feed pressure at different temperatures. Figure 3: Model validation using literature data (van den Bergh et al., 2008b). Figure 4: Calculated CO2/CH4 selectivity vs. feed pressure at different temperatures C O 2 P e rm e a b ili ty , 1 0 -1 2 m o l s- 1 m -1 P a -1 1078 Table 1: Optimal value of the model parameters and corresponding confidence intervals for CO2 in DD3R Parameters Optimal Value Confidence Intervals Cμs0, mol kg-1 3.45 ± 0.04 χ, - 0.23 ± 0.04 b0, 10-8 K0.5 Pa-1 1.19 ± 0.36 QL, kJ mol-1 23.525 ± 0.421 5. Conclusions The separation performances of DD3R zeolite membranes were investigated using the multicomponent Maxwell-Stefan approach. For this investigation, the permeation of an equimolar gas mixture (CH4, CO2, CO, N2) was simulated as a function of temperature and feed pressures. The model adopted was preliminary validated using experimental data from the literature. The simulation results showed that the permeabilities of the less polar molecules (N2 and CH4) are lower than the more polar ones (CO2 and CO). The most important results consisted in predicting the membrane behaviour at high feed pressures and in a wide range of temperature, this providing the optimal usage conditions for the DD3R membranes. In particular, the CO2/CH4 selectivity is found to be very high (100-1000), especially at lower temperatures and feed pressures, whereas a gradual decrease is observed with increasing both variables. Acknowledgements The research under this project is co-funded by the European Union Seventh Framework Programme (FP7/2007 - 2013) under DEMCAMER project (NMP3-LA-2011-262840). References Bakker W.J., Kaptejin F., Poppe J., Moulijn J.A., 1996. Permeation characteristics of a metal supported silicalite-1 zeolite membrane. J. Membrane Sci., 117, 57-78. Do D.D., 1998. Adsorption Analysis: Equilibria and Kinetics. Imperial College Press, London. ISBN 1-86094137-3. Gascon J., Blom W., van Miltenburg A., Ferreira A., Berger R., Kapteijn F., 2008. Accelerated synthesis of all- silica DD3R and its performance in the separation of propylene/propane mixtures. Micropor. Mesopor. Mat., 115, 585-93. Himeno S., Tomita T., Suzuki K., Yoshida S., 2007. Characterization and selectivity for methane and carbon dioxide adsorption on the all-silica DD3R zeolite. Micropor. Mesopor. Mat., 98, 62-9. Kangas J., Sandstrom L., Malinen I., Hedlund J., Tanskanen J., 2013. Maxwell–Stefan modeling of the separation of H2 and CO2 at high pressure in an MFI membrane. J. Membrane Sci., 435, 186-206. Kaptejin F., Moulijn J.A., Krishna R., 2000. The generalized Maxwell-Stefan model for diffusion in zeolites: sorbate molecules with different saturation loadings. Chem. Eng. Sci., 55, 2923-30. Krishna R., Wesselingh J. A., 1997. The Maxwell-Stefan approach to mass transfer. Chem. Eng. Sci., 52, 861. Krishna R., Paschek D., 2000. Separation of hydrocarbon mixtures using zeolite membranes: a modelling approach combining molecular simulations with the Maxwell–Stefan theory. Sep. Purif. Technol., 21, 111. Kuhn J., Yajima K., Tomita T., Gross J., Kapteijn F., 2008. Dehydration performance of a hydrophobic DD3R zeolite membrane. J. Membrane Sci., 321, 344-9 1079 Lee S.C., 2007. Prediction of permeation behavior of CO2 and CH4 through silicalite-1 membranes in single- component or binary mixture systems using occupancy-dependent Maxwell-Stefan diffusivities. J. Membrane Sci., 306, 267-76. Li S., Falconer J.L., Noble R.D., Krishna R., 2007. Interpreting Unary, Binary, and Ternary Mixture Permeation Across a SAPO-34 Membrane with Loading-Dependent Maxwell-Stefan Diffusivities. J. Phys. Chem. C, 111, 5075-82. Tomita T., Nakayama K., Sakai H., 2004. Gas separation characteristics of DDR type zeolite membrane. Micropor. Mesopor. Mat., 68, 71-5. van den Bergh J., Zhu W., Groen J.C., Kapteijn F., Moulijn J.A., Yajima K., Nakayama K., Tomita T., Yoshida S., 2007. Natural gas purification with a DDR zeolite membrane; permeation modelling with Maxwell- Stefan equations. In: Xu R., Gao Z., Chen J. and Yan W. Editors, Zeolite to Porous MOF Materials - the 40th Anniversary of International Zeolite Conference, 1st Edition. van den Bergh J., Zhu W., Kapteijn F., Moulijn J.A., Yajima K., Nakayama K., Tomita T., Yoshida S., 2008. Separation of CO2 and CH4 by a DDR membrane. Res. Chem. Intermediat., 34, 467-74. van den Bergh J., Zhu W., Gascon J., Moulijn J.A., Kapteijn F., 2008. Separation and permeation characteristics of a DD3R zeolite membrane. J. Membrane Sci., 316, 35-45. van den Bergh J., 2010. DD3R zeolite membranes in separation and catalytic processes: Modelling and Application. PhD Thesis. Technische Universiteit Delft, Delft, The Netherlands. ISBN: 978-90-5335-298-4. van den Bergh J., Tihaya A., Kapteijn F., 2010. High temperature permeation and separation characteristics of an all-silica DDR zeolite membrane. Micropor. Mesopor. Mat., 132, 137-47. van den Bergh J., Mittelmeijer-Hazeleger M., Kapteijn F., 2010. Modeling Permeation of CO2/CH4, N2/CH4, and CO2/Air Mixtures across a DD3R Zeolite Membrane. J. Phys. Chem. C, 114, 9379-89. Wirawan S.K., Creaser D., Lindmark J., Hedlund J., Bendiyasa I.M., Sediawan W.B., 2011. H2/CO2 permeation through a silicalite-1 composite membrane. J. Membrane Sci., 375, 313-22. Zhu W., Kapteijn F., Moulijn J.A., 1999. Shape selectivity in the adsorption of propane/propene on the all-silica DD3R. Chem. Commun., 2453-4. 1080