Pervaporation of dilute alcoholic mixtures using PDMS membrane

Pervaporation of dilute alcoholic mixtures using PDMS membrane

Chemical Engineering Science 60 (2005) 1875 – 1880 www.elsevier.com/locate/ces Pervaporation of dilute alcoholic mixtures using PDMS membrane Toraj M...

418KB Sizes 0 Downloads 81 Views

Chemical Engineering Science 60 (2005) 1875 – 1880 www.elsevier.com/locate/ces

Pervaporation of dilute alcoholic mixtures using PDMS membrane Toraj Mohammadia,∗ , Abdolreza Aroujalianb , Ali Bakhshia a Research Laboratory for Separation Processes, Faculty of Chemical Engineering, Iran University of Science and Technology, Narmak, Tehran, Iran b Department of Chemical Engineering, Amirkabir University of Technology, Tehran Polytechnic, No. 424, Hafez Avenue, 15875-4413 Tehran, Iran

Received 31 August 2004; received in revised form 2 November 2004; accepted 25 November 2004 Available online 28 January 2005

Abstract Pervaporation (PV) of methanol/water and ethanol/water mixtures through PDMS membrane was investigated using a PV cell (in laboratory scale). PDMS membrane is a well-known hydrophobic membrane for removing organics from aqueous mixtures. Experimental results were obtained at different initial alcohol (methanol and ethanol) concentrations (0.3–3 wt%) and temperatures (30–50 ◦ C). Recirculation flow rate was kept constant at a value of 15.6 l/h. Average permeation flux (j ), separation factor () and activation energy of permeation (EP ) were calculated. Separation factor of PDMS membrane for methanol was greater than that for ethanol. Total flux for methanol/water and ethanol/water mixtures was observed to vary from 0.37 to 0.56 (kg/m2 h) and 0.52 to 0.90 (kg/m2 h) at 30 ◦ C, respectively, as alcohol concentration changed from 0.3 to 3 wt%. Separation of alcohols depends on both their selective sorption in polymeric membrane and their diffusivity. The most important observation was that separation factor of methanol/water mixtures is greater than that of ethanol/water mixtures and it is because of different molecular size of alcohols. Different behavior of alcohol/water mixtures can also be explained in the entire concentration range studied using relative values of solubility parameters of the alcohols. It can be due to the fact that activation energy of alcohol permeation increases as solubility parameter difference between alcohol and membrane increases. 䉷 2005 Elsevier Ltd. All rights reserved. Keywords: Pervaporation; Methanol/water mixture; Ethanol/water mixture; Solubility parameter; PDMS membrane

1. Introduction Partial vaporization of a liquid through a dense polymeric membrane is called pervaporation (Favre, 2003). Compared with distillation, pervaporation can often be considered a better candidate for separation of close boiling, azeotropic or isomeric mixtures (Verkerk et al., 2001). For removal of volatile organic compounds, other separation technologies such as distillation, liquid–liquid extraction, carbon adsorption and air stripping are not applicable because of feed condition limitations, large volume of byproducts or high cost of post-treatments. However, pervaporation can be applied without these limitations (Kim et al., 2002). Currently, industrial applications of pervaporation are grouped into two: one is dehydration of alcohols and other organic solvents using hydrophilic or charged polymeric membranes ∗ Corresponding author. Tel.: +98 21 789 6621; fax: +98 21 789 6620.

E-mail address: [email protected] (T. Mohammadi). 0009-2509/$ - see front matter 䉷 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.ces.2004.11.039

and the other is removal of small quantities of volatile organic compounds from water using hydrophobic membranes (Zhang et al., 1992). For polymeric pervaporation membranes, extensive research was performed to find an optimized membrane material having selective interaction with a specific component of feed mixture to maximize separation performance in items of separation factor, flux and stability. However, performance of these membranes is strongly influenced by process conditions such as feed concentration and temperature (Verkerk et al., 2001). Siliconecontaining polymers were generally found to exhibit good organophilicity and silicone rubber (mainly polydimethylsiloxane) based membranes were most investigated for separation of organic aqueous mixtures such as alcohols, ketons, phenols and chlorohydrocarbones (Zhang et al., 1992; Jiraratananon et al., 2002; Watson and Payne, 1990). Permeation of molecules through a dense non-porous polymer matrix is generally governed by the sorption–diffusion mechanism. Relative sorption of permeants in membrane

1876

T. Mohammadi et al. / Chemical Engineering Science 60 (2005) 1875 – 1880

depends on their relative solubility in the membrane. Extents of solubility or miscibility of a component in or with polymeric membranes can be explained by the solubility parameter theory. Solubility parameter was defined by Hidebraned and Scott using “cohesive energy density” which is a measure of cohesive force that holds molecules together in liquid phase (Mandal and Pangarkar, 2002). Silicone is recognized as one of pressure-sensitive key industry materials for design of release coating and pressure-sensitive adhesive. Semi-organic molecular siloxane bond provides a highly flexible backbone with large bond angles, long bond lengths and extreme freedom of rotation. Energy required for bond rotation is nearly zero. Freedom of rotation allows for siloxane polymer orientation, a helix polymeric structure consisting of an inorganic Si–O–Si backbone (high surface energy) with a pendant methyl group (low surface energy). Low solubility parameter of PDMS ([– Si(CH3 )2 – O– ]n ) shows immiscibility to most organic polymers and organic-substituted methylsiloxane. However, substitution of methyl with other organic groups provides modified reactivity, adhesion, surface energy, thermostability, hydrophilicity, etc. (Kuo, 2003). Water/alcohol separation is a well-known example of pervaporation process in chemical industry (Xu et al., 2003). Separation of alcohol/water mixtures by pervaporation is important for obtaining liquid fuel from biomass sources. Many authors reported pervaporation process principals and experimental results using different hydrophobic membranes (Molina et al., 2002). For example, Huang et al. (2001) used a sulfonated poly (ether ether keton) membrane and Molina et al. (2002) used CMG-OM-010 and 1060-SULZER membranes (alcohol concentration was varied between 13 and 20 wt% in water). In this research, an attempt was made to report effect of concentration and temperature on pervaporation of methanol and ethanol mixtures (alcohol concentration was varied between 0.3 and 3 wt% in water) using a hydrophobic flat-plate composite membrane with a dense skin layer of polydimethylsiloxane. 2. Experimental 2.1. Membrane preparation PDMS, crosslinking agent TAOS and catalyst dibutyltin dilaurate were mixed according to a 10/1/0.2 weight ratio in n-Heptane. Prior to coating, the CA support was laid and spread out on surface of water in a basin. The membrane system containing some crosslinked PDMS, was kept under ambient temperature for 2 h, and then introduced into an vacuum oven at 60 ◦ C for 4 h to complete crosslinking. The detailed procedure was described elsewhere (Li et al., 2004). 2.2. Degree of swelling The degree of swelling was measured by immersing the membrane in 0.3–5 wt% aqueous alcoholic mixtures

Fig. 1. Schematic diagram of the experimental set-up.

at 30 ◦ C for 72 h. Degree of swelling =

Ws − Wd × 100, Wd

(1)

where, Ws denotes weight of the swollen membrane and Wd denotes weight of the same membrane before immersion. 2.3. Materials and methods Methanol and ethanol (99.5%) was purchased from Merck Co. Ltd. . De-ionized laboratory water was used for making aqueous mixtures. Feed solutions, methanol/water and ethanol/water mixtures containing 0.3–3 wt% alcohols were prepared. Temperature was adjusted at 30, 40 and 50 ◦ C, meanwhile, recirculation flow rate velocity was kept constant at 15.6 l/h using a bypass pump to achieve a Reynolds number of 117. PDMS membrane employed in this study had a thickness of 128 m (support layer of 120 m and dense silicone rubber skin layer of 8 m). All experiments in this study were performed using the same membrane. PV experiments were carried out using an apparatus as shown in Fig. 1. The membrane was housed in a PV cell that consisted of two detachable stainless-steel parts. The membrane had an effective area of approximately 0.0024 m2 . Rubber O-rings were used to provide a pressure tight seal between the membrane and the PV cell. Physical dimensions, length, height and width of PV cell were 0.13, 0.06 and 0.09 m, respectively. A pump was employed to recirculate feed solution and feed temperature was controlled within 3 ◦ C using a thermostat. Volume of feed tank was 7 l, which was very big compared with permeation volume; therefore, variation of feed concentration during a period of 1 h was negligible. In all experiments, feed was kept at atmospheric pressure, whereas permeate pressure was maintained in a range of 8–10 mbar by an oil sealed vacuum pump (MOTO GEN 80-4B with RPM 1380, Iran). Permeate samples were condensed and collected in a Pyrex glass condenser kept inside a cryogenic trap at −35 ◦ C. An accurate refractometer (DR-A1) was employed to analyze alcohol concentration in

T. Mohammadi et al. / Chemical Engineering Science 60 (2005) 1875 – 1880

permeate samples. Some samples were also analyzed using a Varian gas chromatography (model STAR 3400 CX) equipped with a flame ionization detector for confirmation. Permeate flux was calculated using the following equation: j=

M , At

(2)

where j is total flux (kg/m2 h), M is permeate weight (kg), A is effective membrane surface area (m2 ) and t is PV time (h). Separation factor was also calculated using the following equation:

PV =

Y (1 − X) , X(1 − Y )

(3)

where PV is separation factor (dimensionless), X is weight fraction of alcohols in feed and Y is weight fraction of alcohols in permeate. 3. Results and discussion 3.1. Swelling of PDMS membrane in alcohol/water mixtures Swelling measurements of PDMS membrane in alcohol/water binary mixtures were presented in Fig. 2. Degree of swelling of the membrane in methanol/water and ethanol/water mixtures was measured as a function of alcohol concentration. Degree of swelling values in ethanol/water mixtures was higher than those in methanol/water mixtures. According to Table 1, solubility parameter difference between ethanol and PDMS membrane is smaller than that between methanol and PDMS membrane (solubility parameter of PDMS is 14.9 (M pa)1/2 ). Results show that the difference between degrees of swelling of PDMS membrane in both mixtures is small at low alcohol concentration (less than 1 wt%). However, at concentrations more than 1 wt%, the difference is slightly enhanced with increasing alcohol concentration. This may be attributed

Fig. 2. Degree of swelling of PMDS membrane.

1877

Table 1 Solvent properties (Mandal and Pangarkar, 2002; Shah et al., 2000) Solvent

Calculated molecular diameter (nm)

Solubility parameter (M pa)1/2

Water Methanol Ethanol

0.26 0.41 0.52

47.9 29.7 26.2

to different molecular size of methanol and ethanol. It can be said that increasing molecular size of alcohol (from methanol to ethanol) causes PDMS membrane to swell more. As a result, more water can permeate through the membrane when treating ethanol/water mixture. 3.2. Effect of feed concentration Fig. 3 presents results of PV of methanol/water and ethanol/water mixtures through PDMS membrane. This figure shows the effect of alcohol concentration on methanol and ethanol fluxes. As seen, as alcohol concentration increases, alcohol fluxes increase significantly. It is due to the fact that increasing alcohol concentration increases alcohol sorption into the membrane, and as a result, the membrane becomes more swollen. The results show that methanol flux is slightly higher than ethanol flux. According to Table 1, solubility parameter difference of methanol and the membrane is 14.8 compared with that of ethanol and the membrane which is 11.2, but diffusion coefficient of methanol is larger than that of ethanol because of its smaller molecular size. Solvents with larger molecules have more interaction with polymers and this results in the fact that diffusion coefficient of solvents through polymeric membranes decrease as their molecular size increases. It seems that diffusion effect is more significant than sorption effect on alcohol flux. As can be observed in Fig. 3, for both mixtures, increasing alcohol concentration increases permeation fluxes of both alcohols and water, however, water flux increases more significantly than alcohol fluxes resulting in a separation factor reduction. This can be due to the fact that increasing alcohol concentration increases membrane-free volume and simultaneously side chain mobility increases. Consequently, small-sized water cluster can permeate easily through the membrane-free volume. In other words, this phenomenon enhances diffusion of water into the membrane and as a result separation factor decreases. As can be seen, in the range of 0.3–3 wt%, separation factors of alcohols are greater than 1, therefore, the permeate alcohol concentration is more than the feed alcohol concentration (Fig. 4). It can also be seen that the reduction of separation factor at low alcohol concentrations (below 1 wt%) is more significant than that at high alcohol concentrations. In other words, for methanol/water mixtures while methanol concentration increases from 0.3 to 1 wt%, separation factor decreases up to 55%, but while it increases

1878

T. Mohammadi et al. / Chemical Engineering Science 60 (2005) 1875 – 1880

Fig. 4. Variation of alcohol concentration in permeate with alcohol concentration in feed.

and 37%, respectively. The most important observation was that separation factor of ethanol/water mixture was less than that of methanol/water mixture. It can be due to the fact that enlargement of the membrane-free volume, when it is in contact with ethanol/water mixture, is higher. 3.3. Effect of feed temperature

Fig. 3. Permeation fluxes and separation factor of methanol/water and ethanol/water mixtures at 30 ◦ C.

from 1 to 3 wt%, separation factor decreases (about 45%). It can be said that at higher alcohol concentrations (more than 1 wt%), alcohol permeation flux increases. As can be observed, at first, alcohol concentration range is 0.3–1 wt%, alcohol permeation flux increases by about 55% and water permeation flux increases by about 7%, but at the other range (1–3 wt%), these permeation fluxes increase 130%

Fig. 5 presents effect of temperature on PV of methanol/water and ethanol/water mixtures through PDMS membrane. As seen, increasing temperature increases permeation fluxes of alcohols and water. It can be due to the fact that during PV, permeating molecules diffuse through free volumes of the membrane. Thermal motions of polymer chains in amorphous regions randomly produce free volumes. As temperature increases, frequency and amplitude of polymer jumping chains increase. As a result, free volume of the membrane increases. Thus, at higher temperatures, diffusion rate of individual permeating molecules increases leading to high permeation fluxes (Sampraniboon et al., 2000). The results showed that total permeation flux of ethanol/water mixture was slightly higher than that of methanol/water mixture and the separation factor of ethanol was less than that of methanol because of their different molecular sizes and also solubility parameters (Kuo). From the results, it can be concluded that PDMS membrane is practically not selective at 3 wt% ethanol and methanol concentrations, however, the membrane is suitable for less alcohol concentration. Flux data were, respectively, fitted to Arrehenius equation (Feng and Huang, 1996) and membrane activation energy (EP ) was calculated (Fig. 6). Table 2 shows activation energy of PDMS membrane for PV of methanol/water and ethanol/water mixtures. It can be seen that activation energy of PDMS membrane for methanol is more than that for ethanol. It can be said that activation energy of alcohol permeation increases as solubility parameter difference (between alcohol and membrane) increases.

T. Mohammadi et al. / Chemical Engineering Science 60 (2005) 1875 – 1880

1879

Fig. 6. Arrhenius plot for PDMS membrane at 3 wt% alcohol concentration. Table 2 Activation energy of alcohol permeation through PDMS membrane Alcohol

EP (kJ/mol)

Methanol Ethanol

49.61 43.35

Fig. 5. Permeation fluxes and separation factors of methanol/water and ethanol/water mixtures at 3 wt% alcohol concentration.

3.4. Effect of duration time on permeation fluxes As seen in Fig. 7, effect of duration time on both alcohols and water fluxes was also studied. In all experiments, the total feed volume and the feed flow rate along the membrane surface were 7 l and 15.6 l/h, respectively. It can be

Fig. 7. Permeation fluxes of methanol/water and ethanol/water mixtures versus time at 3 wt% alcohol concentration.

seen that permeation fluxes decrease with time. At the beginning, alcohol fluxes decrease sharply because alcohol concentration in feed is high. As time goes on, the alcohol concentration decreases more slowly. It is due to the fact that

1880

T. Mohammadi et al. / Chemical Engineering Science 60 (2005) 1875 – 1880

alcohol concentration in feed decreases and also mass transfer resistance in boundary layer increases. As seen, water flux decreases slowly during the whole experiment. It is due to the fact that water concentration in feed is always high (Li et al., 2002). More studies are currently under investigation. 4. Conclusion PV of methanol/water and ethanol/water mixtures using PDMS membrane was studied at different concentrations and temperatures. Fluxes and separation factors of ethanol and methanol were compared. Total permeation flux of methanol/water mixture through PDMS membrane was found to vary from 0.37 to 0.56 (kg/m2 h) and that of ethanol/water mixture from 0.52 to 0.90 (kg/m2 h) over a concentration range of 0.3–3 wt% at 30 ◦ C. Little flux and high separation factor were obtained for methanol/water mixtures. It was observed that under the same experimental conditions, alcohol permeation flux and separation factor for methanol/water mixture were higher than those for ethanol/water mixture. According to Table 1, solubility parameter difference of methanol and the membrane is 14.8 compared with that of ethanol and the membrane which is 11.2. However, diffusion coefficient of methanol is higher than that of ethanol because of its smaller molecular size. Solvents with larger molecules have more interaction with polymers and this result that diffusion coefficient of solvents through polymeric membranes decrease as their molecular size increase. It seems that diffusion effect is more significant than sorption effect on alcohol flux. References Favre, E., 2003. Temperature polarization in pervaporation. Desalination 154, 129–138. Feng, X., Huang, R.Y.M., 1996. Estimation of activation energy for permeation in pervaporation processes. Journal of Membrane Science 118, 127. Huang, R.Y.M., Shao, P., et al., 2001. Pervaporation separation of water/isopropanol mixture using sulfonated poly (ether ether keton) (SPEEK) membranes: transport mechanism and separation performance. Journal of Membrane Science 192, 115–127.

Jiraratananon, R., Sampranpiboon, P., Uttapap, D., Huang, R.Y.M., 2002. Pervaporation separation and mass transport of ethylbutanoate solution by polyether block amid (PEBA) membranes. Journal of Membrane Science 210, 389–409. Kim, H.J., Nah, S.S., Min, B.R., 2002. A new technique for preparation of PDMS pervaporation membrane for VOCs removal. Advanced in Environmental Research 6, 255–264. Kuo, A.C.M., 2003. Silicone release coatings for the pressure sensitive industry—overview and trends. D.C. Corporation, pp. 1–4. Li, J., Chan, C., Han, B., Peng, Y., Zou, J., Jiang, W., 2002. Laboratory and pilot-scale study on dehydration of benzen by pervaporation. Journal of Membrane Science 203, 127–136. Li, L., Xiao, Z., Tan, S., Zhibing, L., 2004. Composite PDMS membrane with high flux for separation of organics from water by pervaporation. Journal of Membrane Science 243, 177–187. Mandal, S., Pangarkar, V.G., 2002. Separation of methanol–benzen and methanol–toluene mixtures by pervaporation: effects of thermodynamics and structural phenomenon. Journal of Membrane Science 201, 175–190. Molina, J.M., Vatai, G., Bekassy-Molnar, E., 2002. Comparison of pervaporation of different alcohols from water on CMG-OM-010 and 1060-SULZER membrane. Desalination 149, 89–94. Sampraniboon, P., Jiraratananon, R., Uttapap, D., Feng, X., Huang, R.Y.M., 2000. Separation of aroma compounds from aqueous solutions by pervaporation using polyoctylmethylsiloxane (PDMS) membranes. Journal of Membrane Science 174, 55–65. Shah, D., Kissick, K., Ghorpade, A., Hannah, R., Bhattacharyya, D., 2000. Pervaporation of alcohol–water and dimethylformamide–water mixture using hydrophilic zeolite NaA membranes: mechanism and experimental results. Journal of Membrane Science 179, 185–205. Verkerk, A.W., Male, P.V., Vorstman, M.A.G., Keurentjes, J.T.F., 2001. Description of dehydration performance of amorphous silica pervaporation membranes. Journal of Membrane Science 193, 227– 238. Watson, J.M., Payne, P.A., 1990. A study of a organic compound pervaporation through silicone rubber. Journal of Membrane Science 49, 171–185. Xu, Z., Dai, Q., Liu, Z., Kou, R., Xu, Y., 2003. Microporous polypropylene hollow fiber membranes part II. Pervaporation separation of water/ethanol mixtures by the poly (acrylic acid) grafted membranes. Journal of Membrane Science 214, 71–81. Zhang, S.Q., Fouda, A.E., Matsuura, T., 1992. A study on pervaporation of aqueous benzyl alcohol solution by polydimethylsiloxane membrane. Journal of Membrane Science 70, 249–255.