water of polydimethylsiloxane composite membranes

water of polydimethylsiloxane composite membranes

Materials and Design 34 (2012) 732–738 Contents lists available at ScienceDirect Materials and Design journal homepage: www.elsevier.com/locate/matd...

983KB Sizes 1 Downloads 50 Views

Materials and Design 34 (2012) 732–738

Contents lists available at ScienceDirect

Materials and Design journal homepage: www.elsevier.com/locate/matdes

The structure and pervaporation properties for acetic acid/water of polydimethylsiloxane composite membranes Housheng Hong ⇑, Longxiang Chen, Qingwen Zhang, Feng He State Key Laboratory of Materials-Oriented Chemical Engineering, Nanjing University of Technology, Nanjing 210009, PR China

a r t i c l e

i n f o

Article history: Received 17 June 2010 Accepted 2 June 2011 Available online 28 June 2011 Keywords: Polymers Performance indices Material selection charts

a b s t r a c t Polydimethylsiloxane (PDMS)/organic montmorillonite (OMMT)/polyether polyethersulfone (PES) composite membranes were prepared by in situ anionic polymerization using 3-aminopropyltrimethoxy (AMEO) as a crosslinker. The morphology, thermal properties and interaction of PDMSAMEO/OMMT membranes were characterized by using a scanning electron microscope (SEM), Fourier transform infrared spectroscopy (FT-IR) and a thermal gravimetric analysis (TGA). The swelling behavior of membranes without PES support was investigated. The effects of AMEO content and OMMT content on separation properties were also studied. The results show that the addition of appropriate OMMT could improve the hydrophobic and pro-acetic acid properties of a membrane. The acetic acid selectivity of membranes was best when AMEO content was 0.1. The membrane, loading 2 wt.% OMMT, exhibited the highest separation factor for a feed concentration of 10 wt.% at 313 K. An increase in feed concentration resulted in the enhancement of flux and selectivity. When the feed concentration was above 20 wt.%, the separation factor of a filled membrane was larger than for an unfilled membrane. With increases in the feed temperature, the permeation flux of membranes increased. However, the acetic acid selectivity of an unfilled membrane decreased but for filled membranes initially increased before decreasing. Crown Copyright Ó 2011 Published by Elsevier Ltd. All rights reserved.

1. Introduction Silicone rubber is a compound that has alternate silicon and oxygen atoms, with the particular properties of being both semiorganic and semi-inorganic. The silicone rubber membranes not only have hydrophobic ability, but also fine heat and cold resistance and chemical durability which have been widely applied in pervaporation separation of organic–water mixtures. Pervaporation is a good alternative to conventional processes due to the simplicity of operation and the low working cost and reduced amounts of chemicals required [1–3]. Additionally it offers an opportunity for the separation of organic liquid mixtures, especially in azeotropic mixtures or organic contaminated industrial wastewater [4]. However, the homogeneous membrane is not meeting the practical application requirements of separation factor and permeation flux. Polymers filled with organic particles or inorganic particles gained considerable interest owing to their enhanced mechanical, rheological properties, and carrier properties [5,6]. The mechanical properties are affected by filler type, filler concentration, and the interaction between filler and matrix. Fang et al. [7] prepared super-hydrophobic nanosilica which exhibited uniform dispersion in the PDMS matrix, and their composites also showed good ⇑ Corresponding author. Tel.: +86 25 83403940/83172063; fax: +86 25 83405355. E-mail address: [email protected] (H. Hong).

mechanical properties and distinct advantage with thermal stability compared with those of the pure silica-filled PDMS composites. The influence of the surface nature (hydrophobic and hydrophilic) and concentration of silica nanoparticles on the coalescence behavior of immiscible polydimethylsiloxane (PDMS)/polyisobutylene (PIB) which was blended under simple low-rate shear flow were investigated via optical shear techniques [8]. Yu et al. [9] prepared the hexadecyl trimethyl ammonium bromide (CTAB) filled membrane used for the separation of dilute acetic acid (ethanol) solution. The results indicated that hybrid membrane had better separation performance. Tang et al. [10] prepared fumed-silicafilled polydimethylsiloxane–polyamide composite membranes by the introduction of hydrophobic fumed silica into a PDMS skin layer. Their pervaporation performance was tested with aqueous ethanol solutions. Increasing the amount of the fumed silica resulted in significantly enhanced ethanol permeability of the membranes. On the other hand, composite membranes are of practical importance in order to meet a sufficiently high flux requirement by reducing the thickness of the active layer. Integral composite membranes are made of one polymer forming a thin active layer upon a porous support layer. The active layer generally determines the separation performance of the membrane including the separation factor and flux, while the support layer endows the composite membrane with necessary mechanical properties [11,12]. Direct

0261-3069/$ - see front matter Crown Copyright Ó 2011 Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.matdes.2011.06.041

H. Hong et al. / Materials and Design 34 (2012) 732–738

coating of a polymer solution onto a microporous support is the current widely used method to prepare composite membranes [13]. Acetic acid, one of the top 50 chemicals, is important in the chemical industry. It is widely used in industry of medical, synthetic fiber, coating, pesticide, food additives, dyeing weaving and so on. It is an important component of China’s national economy. With the consumption of petroleum resources and the prices of petrochemical materials rising, renewable bio-resources are a worldwide concern. Using pervaporation coupled with fermentation for continuous fermentation of acetic acid, the acid can be separated from fermentation liquid. The inhibitory effect of the product would be significantly decreased to improve conversion rate and acid production rate. It can simultaneously increase the concentration of acetic acid to reduce the cost of stockpiles, transit and separation. In this study, using 3-aminopropyltrimethoxy (AMEO) as a cross-linker, the PDMS–AMEO composite membranes were prepared using polyethersulfone (PES) as a support layer. In addition, the organic montmorillonite (OMMT) was used as filler to improve the performance of the PDMS–AMEO composite membrane by in situ anionic polymerization. 2. Experimental 2.1. Materials Specification and suppliers of the material and reagent, used for the preparation of membranes, are shown in Table 1. 2.2. Membrane preparation PDMS homogenous solution was cast on the PES support to form the active layer of composite membrane. Firstly, PDMS and AMEO were dissolved in n-heptane under stirring to form a homogenous solution at room temperature. The catalyst, dibutyltin dilaurate, was then added into the above solution, with a weight ratio of 0.01 to PDMS. After degassing, the solution was cast onto the PES support (the holes of which were blocked with polyethylene glycol) using a casting knife, then dried in open air and then placed in a vacuum drying oven at the desired temperature to complete crosslinking and to evaporate the residual solvent. Finally, the membranes was soaked in deionizer water making the polyethylene glycol to dissolve from the holes, dried in air for 24 h and then placed in a vacuum drying oven for 24 h at 313 K. All membrane samples were stored in a dust free and dry environment before being used in the pervaporation experiments. For simplicity, the AMEO contents were indicated the weight ratio of AMEO to PDMS. The OMMT filled PDMS–AMEO/PES composite membranes were also prepared following an identical procedure. By following the similar procedure, PDMS–AMEO homogeneous membranes and OMMT filled PDMS–AMEO membranes (only the active layer without PES support) were fabricated for characterization purposes, except that the solutions were cast on the glass plate instead of the PES support membrane.


The preparation process of membranes is shown in Fig. 1. It shows that the AMEO not only could participate in the cross-linking reactions, but its amino-group could produce the PDMS chain onto the surface of the support layer with Van Der Waals Force. The superfluous amino-group would affect the mass transfer separation process as well. 2.3. Membrane characterization 2.3.1. The top and cross-sectional views The morphology and OMMT distribution of the composite membranes were observed and recorded by a QUANTA200 scanning electron microscope (SEM). 2.3.2. FT-IR Fourier transform infrared (FT-IR) spectra were obtained by a Nexus 670 (Nicolet Instrument Corporation, USA) instrument equipped with a scanning range of 400–4000 cm1. The experiments were run in open air. 2.3.3. TGA Thermal degradation measurements of the homogeneous membranes and OMMT filled PDMS–AMEO membranes samples were performed using a thermogravimetric analyzer (TGA) (STA 449C, Netzsch, Germany) at a heating rate of 10 °C/min and temperature programs were run from 40 °C to 800 °C in a nitrogen environment. A nitrogen flow of 20 ml/min was utilized in order to remove all corrosive gas involved in the degradation. 2.3.4. Swelling behaviour The PDMS–AMOE membranes and OMMT filled 4 wt.% PDMS– AMOE without PES support were weighed carefully before being immersed in the feed mixture at 303 K. The swollen membrane samples were taken out from the feed mixture after a certain period of time and wiped to remove the surface liquid before being weighed. The degree of swelling (D) was calculated by:

D ¼ ðms  md Þ=md


where md and ms were the weights of the dried and swollen membranes, respectively. 2.4. Pervaporation Pervaporation experiments were carried out on a self-regulating pervaporation membrane module. The feed solution was pumped into the membrane cell. The permeate vapour was collected in liquid nitrogen traps. The weight of permeate collected in the cold trap was measured to give the permeation flux, J:

J ¼ Q=ðA  tÞ


where Q is the total amount permeated during the experimental time slot, t, and A is the effective area of the membrane. The feed and permeate compositions were analyzed by gas chromatography (6890 N).

Table 1 Specification and suppliers of the materials used in the research. Material



PDMS AMEO n-Hexane Dibutyltin dilaurate OMMT PES membrane

4000 mPa s Analytical grade Analytical grade Analytical grade Trademark DK1 Pore diameter 0.45 lm

Jinan Zhonghao Chemical Co., LTD Nanjing Xinghui Industry and Trade Co., LTD Shanghai Chemical Co., LTD Beijing Chemical Company Zhejiang Fenghong New Materials Co., LTD Beijing Haicheng Shijie Clean Filtering Equipment Co., LTD


H. Hong et al. / Materials and Design 34 (2012) 732–738

Fig. 1. The preparation of PDMS–AMEO membrane.

The separation selectivity of the membrane was expressed by the separation factor (a):

a ¼ ðyHAc =ywater Þ=ðxHAc =xwater Þ


where xHAc is the weight fraction of acetic acid in the feed and xwater is the weight fraction of water in the feed. yHAc is the weight fraction of acetic acid in the permeation product and ywater is the weight fraction of acetic acid in the permeation product. 3. Results and discussion 3.1. Membrane characterization 3.1.1. The top and cross-sectional views The scanning electron micrographs of the top and cross-sectional views of the resulting membranes are shown in Fig. 2. As for the PDMS–AMEO/PES membrane, it is generally believed that it has a dense structure without pores. Also, no splits were found at the interface between the PDMS active layer and the PES support. Correspondingly, it can be seen from (c) and (d) that, in the PDMS–AEMO/OMMT/PES membrane with the OMMT loading of 4 wt.%, OMMT particles were distributed uniformly throughout the polymer matrix. Therefore, it could improve the absorption of membrane to liquid components. 3.1.2. FT-IR Fig. 3 shows the FT-IR spectra of OMMT, PDMS–AMEO, and PDMS–AMEO/OMMT. Two peaks are observed near 3432 cm1 and 3623 cm1 in the spectrum of OMMT. The peaks are attributed to the water in the interlayer of OMMT and the stretching of –OH vibration. The characteristic peaks at around 1470 cm1 and 2851 cm1 represented the –CH2 vibration, and the peak at 2923 cm1 is stretching of the C–H vibration. This is the characteristic peak of organic–organic of quaternary ammonium salt. It indicates that quaternary ammonium salt had entered into the MMT layers. On the other hand, the peaks of C–H stretching vibration around 2920– 2850 cm1 and bending vibration around 1470 cm1 are observed. This demonstrates that the MMT has been organized and modified. It can be determined from the disappearance of the absorption band around 3432 cm1 and 3623 cm1, that the –OH group of OMMT reacted with the crosslinker. It can also be seen from

Fig. 3 that the characteristic peaks of PDMS–AMEO and PDMS– AMEO/OMMT are almost the same indicating that the polymer structure was not changed by OMMT. 3.1.3. TGA The thermal decomposition expressed in terms of weight loss as a function of the temperature for the PDMS–AMEO sample with OMMT loading of 0%, 4% and 10% is shown in Fig. 4. From 40 °C to 320 °C, there was a subtle weight loss of about 5%, indicating that the membrane material was stable at high temperature. As shown in Fig. 4, the TGA plot of the pure membrane presented three steps of decomposition. The first decomposition step between 40 °C and 320 °C can be assigned to solvent and water loss. In the second step of decomposition there was the high weight loss of about 55% from 330 °C to 500 °C. This may be attributed to the decomposition of the forked chain. The third decomposition step started above 500 °C. After heating the material up to 750 °C there was still a weight residue of about 2.5%. This is due to the breakage of the Si–O bond. It also can be seen that the decomposition temperature of OMMT filled membrane only has two steps due to the second step being inhibited. The thermal stability improved owing to the OMMT being filled, accordingly: firstly, a silicone active terminal was inactivated when it touched the OMMT filler and impeded the decomposition of the backbone. Secondly, the polymer molecule chains were inserted into the layer of silicate. This molecule chains is not only the link to the layer of silicate, but also to the positive ion within the silicate layer. A large amount of molecular chains became entangled which slowed the speed of backbone decomposition. 3.2. Swelling behaviour The swelling behaviour of PDMS–AMEO homogeneous membrane and PDMS–AMEO/OMMT membrane without a PES support layer were investigated. As shown in Figs. 5 and 6, membranes swelled expeditiously in 10 wt.% acetic acid solution in the first 40 h. The swelling degree increased slowly, remaining virtually unchanged with increasing time. Also, the swelled degree of OMMT filled membrane was higher than for unfilled membrane because there was more vacuum in the filled membrane. With the increasing acetic acid, however, the swelled degree of pure membrane decreased. This shows that the membrane was difficult to swell in

H. Hong et al. / Materials and Design 34 (2012) 732–738

Fig. 2. SEM pictures of PDMS/PES composite membrane.

Fig. 4. TGA curves of membrane material. Fig. 3. IR spectrums: a. OMMT, b. PDMS–AMEO, and c. PDMS–AMEO/OMMT.

acetic acid at the aggregation state, while water molecules are smaller than acetic acid and enter more easily into the membrane. The increase in the OMMT filled membrane indicated that the addition of OMMT improved the absorption capability of acetic acid molecules and decreased the absorption of water due to the hydrophobic of OMMT, and shows that OMMT could restrain membrane swelling when acetic acid content is less than 6 wt.%. 3.3. Pervaporation 3.3.1. Effect of AMEO content on separation performance Fig. 7 shows the variation of permeation flux and separation factor as a function of AMEO content with 10 wt.% acetic acid/water at 303 K. The permeation flux decreased monotonically with AMEO content and a considerable increase in separation factor was found when AMEO content reached 0.1. Furthermore, the separation performance was better than that when the AMEO content exceeded

Fig. 5. Influence of immersing time on swelling degree.



H. Hong et al. / Materials and Design 34 (2012) 732–738

Fig. 8. Effect of OMMT content on pervaporation performance. Fig. 6. Influence of acetic acid concentration on swelling degree.

Fig. 7. Effect of AMEO content on pervaporation performance.

this range up to 0.2. With a comparatively low AMEO content of 0.06, the PDMS active layer was much looser than for the content of 0.1, which resulted in a higher permeation flux and a lower separation factor. When excessive AMEO was added, some of AMEO molecules did not perform their cross-linking function, but stayed in the PDMS active layer and were enriched around the interface of PDMS and PES by forming hydrogen bonds between the –NH2 groups of AMEO and the –SO2– groups of PES support [12]. This kind of enrichment blocked the separation of acetic acid, just as illustrated in Fig. 7, both the total permeation flux and the separation factor decreased. 3.3.2. Effect of OMMT content on separation performance The effect of OMMT content on the separation performance of the blended membranes is shown in Fig. 8. These experiments were carried out for a feed composition of 10 wt.% acetic acid at 313 K. As shown in Fig. 8, a maximum is reached with an increase in OMMT content. This indicates that the OMMT played at least two roles in the membrane’s separation performance. On the one hand, the OMMT was inserted by PDMS chains and the silicate layer was distributed in polymer. So, the hydrophobic passage among silicate layer provides a shortcut for acetic acid ro pass. On the other hand, the continuity of the membrane was destroyed when the OMMT content was high, producing voids without selectivity resulting in separation factor decrease and permeation flux increase.

3.3.3. Effect of acetic acid concentration on separation performance Figs. 9 and 10 show the effect of feed concentration on the pervaporation performance of membranes for OMMT loadings: 0 and 2 wt.% with the feed flow rate set at 26 L h1 at 313 K. As shown in Fig. 9, the separation factor of unfilled membrane was increased with increasing feed concentration. When acetic acid content was above 20 wt.%, the separation factor of the OMMT filled membrane was higher than the unfilled PDMS membrane. The Fig. 10 shows that the permeation flux of the membranes increased with the increasing of feed acetic acid concentration, the permeation flux of OMMT filled membrane was higher than for unfilled membranes. The process of pervaporation includes the action of components on the membrane and the interaction between components. According to the solution-diffusion theoretical model, there is an equilibration between absorption and dissolution of feed components on the surface of a membrane. With feed acetic acid concentration increasing, the quantity of acetic acid adsorbed on the membrane surface increased and the acetic acid concentration of separating layer and support layer would increase. The size of tunnel in the membranes for feed through become larger due to the increase of membrane swelling, and water molecules are smaller than acetic acid, so water molecules are easier to spread through the membrane. Thereby, the flux of acetic acid and water both increased, and the change of separation factor depended on the relative increased strength of acetic acid to water. Deng et al. [14] studied the pervaporation of PDMS for an acetic acid/water mixture and found that the permeation flux and the

Fig. 9. Effect of acetic acid concentration on separation factor.

H. Hong et al. / Materials and Design 34 (2012) 732–738

Fig. 10. Effect of acetic acid concentration on flux.


Fig. 12. Effect of operating temperature on pervaporation flux.

4. Conclusion PDMS–AMEO/OMMT/PES composite membranes were prepared by in situ anionic polymerization using AMEO as a cross-linker. The OMMT was distributed equitably resulting in improved thermal stability. The addition of appropriate OMMT could improve the adsorption ratio of the membrane for acetic acid and water molecules and restrain membrane swelling. In the range of 0.06–0.20, the separation performance of PDMS–AMEO/PES was best when the AMEO content was 0.1. The PDMS–AMEO/OMMT/PES composite membrane showed a high separation performance with the permeation flux of 98.36 g/(m2 h) and the separation factor of 2.23 when the OMMT content was 2 wt.% at 313 K. The flux and selectivity increased with increasing acetic acid concentration since the flux increase of acetic acid was faster than that of water. With the increasing of feed temperature, permeation flux increased significantly. However, the acetic acid selectivity of the unfilled membrane decreased while initially increasing before decreasing for the filled membrane. Fig. 11. Effect of operating temperature on separation factor.

Acknowledgments separation factor increased with increasing feed acetic acid concentration for feed acetic acid concentrations less than 59 wt.%. The finding agrees with the research results presented here. The PDMS membrane investigated by Lu et al. [15] gave a separation factor of 1.9 and a permeation flux of 92 g m2 h1 for a feed acetic acid concentration of about 40 wt.%. Liu et al. [16] prepared silicalite membranes on stainless steel support which gave a separation factor of 0.3 and a permeation flux of 200 g m2 h1 for a feed acetic acid concentration of 5 wt.%. 3.3.4. Effect of feed temperature on separation performance The effect of feed temperature on the pervaporation performance of membranes with OMMT loadings of 0, 2 wt.% was investigated with the feed acetic acid concentration set at 10 wt.% and feed flow rate at 26 L h1. Figs. 11 and 12 show the effect of feed temperature on separation performance. As shown in Fig. 11, the separation factor of the unfilled PDMS membranes decreased with increasing temperature, and the separation factor of the OMMT filled membrane increased initially and then decreased with increasing temperature. The separation factor of filled membrane was higher than that of unfilled membrane when the temperature was from 305 K to 320 K. As shown in Fig. 12, the permeation flux of membranes rose with increasing temperature and the permeation flux of the filled membrane was higher than that of the unfilled membrane. The maximum separation factor of the OMMT filled membrane was 2.17, with the permeation flux of 130.32 g m2 h1 at 313 K.

The authors gratefully acknowledge the support of the State Key Development Program for Basic Research of China (973 Program) (No. 2007CB714305) and the State Key Laboratory of Materials-Oriented Chemical Engineering for their financial support and scholarships. References [1] Peng FB, Pan FS, Li D, Jiang ZY. Pervaporation properties of PDMS membranes for removal of benzene from aqueous solution experimental and modeling. Chem Eng J 2005;114:123. [2] Dobrak A, Figoli A, Chovau S, Galiano F, Simone S, Vankelecom IFJ, et al. Performance of PDMS membranes in pervaporation: effect of silicalite fillers and comparison with SBS membranes. J Colloid Interf Sci 2010;346:254. [3] Chovau S, Dobrak A, Figoli A, Galiano F, Simone S, Drioli E, et al. Pervaporation performance of unfilled and filled PDMS membranes and novel SBS membranes for the removal of toluene from diluted aqueous solutions. Chem Eng J 2010;159:37. [4] Hasanoglu A, Salt Y, Keleser S, Dincer SOS. Pervaporation separation of organics from multicomponcnt aqueous mixtures. Chem Eng Process 2007;46:300. [5] Dongling M, Siegel RW, Jung-Il H, Schadler LS, Mårtensson E, Önneby C. Influence of nanoparticle surfaces on the electrical breakdown strength of nanoparticle-filled low-density polyethylene. J Mater Res 2004;19:857. [6] Grunlan JC, Gerberich WW, Francis LF. Electrical and mechanical property transitions in carbon-filled poly(vinylpyrrolidone). J Mater Res 1999;14:4132. [7] Fang X, Huang X, Lu Z, Chen L, Chen S. Synthesis of new superhydrophobic nanosilica and investigation of their performance in reinforcement of polysiloxane. Polym Compos 2010;31:1628. [8] Tong W, Huang YJ, Liu CL, Chen XL, Yang Q, Li GX. The morphology of immiscible PDMS/PIB blends filled with silica nanoparticles under shear flow. Colloid Polym Sci 2010;288:753.


H. Hong et al. / Materials and Design 34 (2012) 732–738

[9] Yu J, Li H, Liu HZ. Recovery of acetic acid over water by pervaporation with a combination of hydrophobic ionic liquids. Chem Eng Commun 2006;193:422. [10] Tang XY, Wang R, Xiao ZY, Shi E, Yang J. Preparation and pervaporation performances of fumed-silica-filled polydimethylsiloxane–polyamide (PA) composite membranes. J Appl Polym Sci 2007;105:3132. [11] Henis LMS, Triboldi MK. Composite hollow fiber membranes for gas separation: the resistance model approach. J Membr Sci 1981;8:233. [12] Wu H, Zhang XF, Xu D, Li B, Jiang ZY. Enhancing the interfacial stability and solvent-resistant property of PDMS/PES composite membrane by introducing a bifunctional aminosilane. J Membr Sci 2009;337:61.

[13] Feng XS, Huang RYM. Liquid separation by membrane pervaporation: a review. Ind Eng Chem Res 1997;36:1048. [14] Deng S, Sourirajan S, Matsuura T. A study of polydimethylsiloxane/aromatic polyamide laminated membranes for separation of acetic acid/water mixtures by pervaporation process. Sep Sci Technol 1994;29:1209. [15] Lu SY, Chiu CP, Huang HY. Pervaporation of acetic acid/water mixtures through silicalite filled polydimethylsiloxane membranes. J Membr Sci 2000;176:159. [16] Liu Q, Noble RD, Falconer JL, Funke HH. Organics/water separation by pervaporation with a zeolite membrane. J Membr Sci 1996;117:163.