SUN Li, YU Bo, YU Wei, TANG Xiaolin, HU Mingjie, HUANG Chi*

(1. College of Chemistry and Molecular Sciences, Wuhan University, Wuhan 430072, China; 2. Hubei Military-Civilian Integration and Co-Innovation Center of Aerospace Power and Materials Technology, Wuhan 430072, China)

Abstract: Natural rubber/styrene-butadiene-styrene (NR/SBS) membrane was prepared by solution blend, and then used for recovery of butanol from its dilute solution by pervaporation (PV). The thermodynamic and mechanical properties of NR/SBS blend membrane were characterized by TGA and tensile test, respectively. A layer of relatively dense blend membrane with the thickness of about 37 μm was closely cast on a layer of porous polyvinylidene difluoride (PVDF) support. And there was no obvious phase separation observed between the interface of two layers. Both flux and separation factor increased with increasing feed temperature. Butanol flux increased as feed concentration increased consistently. The blend membrane which got the best performance obtained membrane separation factor of 28.8 with total flux of 2695.2 g/m2h at 70 ℃ when feed concentration was 4.00wt%.

Key words: natural rubber; styrene-butadiene-styrene; butanol; pervaporation

1 Introduction

In the past few years, the unrest in major oil-producing countries and the annual decrease in fossil energy reserves have been forcing humans to find alternative new energy sources. Occupying the advantages of sources, energy density and compatibility with gasoline, bio-butanol produced by fermentation has drawn more and more attention compared with bio-ethanol. However, the low concentration of

n

-butanol in fermentation broth caused by the toxicity of

n

-butanol to fermentation strains has always puzzled researchers. Thus, how to quickly and efficiently separate butanol from fermentation broth has become the key to its large-scale industrialization.Pervaporation (PV), which possesses high selectivity and energy efficiency, has been proved to be one of the most promising method to

n

-butanol recovery. The low operating temperature possibility of modular design makes it easy to be coupled with a chemical or biological reaction, resulting in a continuous reaction and high conversion rate. The PV separation processes which is defined by solution-diffusion model often include three steps: (1) adsorption and dissolution of solute molecules upstream of the membrane, (2) diffusion of solute molecules across through the membrane, and (3) desorption of solute molecules downstream. Obviously, the core component of PV is the selective membrane. In practical application, a good PV membrane should have high processing capacity, good selectivity, and a long service life. Thanks to the development of polymer science, several hydrophobic polymeric membranes have been employed to recover

n

-butanol from its dilute aqueous solution by PV, including polydimethylsiloxane (PDMS),poly[1-(trimethylsilyl)-1-propyne] (PTMSP), polyurethane (PU), polyvinylidene difluoride (PVDF)and polyether block amide (PEBA),

etc

. By and large, the main materials of these membranes can be divided into two broad categories, namely elastomer polymers and crystalline polymers. Since the development of pervaporation separation technology, composite membranes have drawn more and more concentration because the PV performance of a single separation membrane has been unable to meet the actual industrial needs. Compared with crystalline polymers, elastomer polymers are more suitable to be used as the base material of composite membranes due to its rich source, variety, low price, and good processing performance. Natural rubber (NR) is one of the most widely used general-purpose rubber (elastomer polymer). With good solvent resistance, inexpensive NR has been employed as selective PV membranes for separation of organic mixtures. However, although unvulcanized natural rubber usually exhibits high flux and separation factors, its poor strength results in a short service life. Although vulcanized natural rubber has good mechanical strength, it often exhibits low flux due to the high degree of cross-linking when removing or recovering trace substances from their dilute aqueous solution. Singha

et al

used crosslinked and filled natural rubber membranes to remove pyridine from water by PV. The modified NR membranes showed good stability and higher selectivity, but their flux decreased. Interpenetrating polymer network (IPN) is an effective method to improve polymer properties by changing the structure of polymer network. It can well combine the dominant properties of the two polymers. Amnuaypanich

et al

prepared natural rubber/poly(vinyl alcohol) semi-interpenetrating polymer network (NR/PVA semi-IPN)and natural rubber/poly(acrylic acid) semi-interpenetrating polymer network (NR/PAA semi-IPN)membranes to separate water and ethanol by PV. The water flux of modified NR membranes increased much more than that of NR membrane.

With good hydrophobicity, styrene-butadiene-styrene (SBS) triblock copolymer has been widely used in recovering trace organics from water. The microphase separation structure of SBS which consists of “hard segment” (polystyrene chains) and “soft segment” (polybutadiene chains) endows it good mechanical properties. And SBS also possesses good compatibility with unvulcanized NR due to the similar solubility parameter, which is beneficial for forming IPN-like structure with unvulcanized NR. The obtained blend polymer should exhibit good PV performance and mechanical properties.

This work aims to develop a stable PV membrane with high performance for the recovery of

n

-butanol from water. Styrene-butadiene-styrene (SBS) was used to prepare a blend membrane with NR by solution blend. The membrane was casted on a PVDF support to recover butanol from its dilute solution by PV. The thermodynamic and mechanical properties of NR/SBS blend film were characterized by TGA and tensile test, respectively. The morphologies of PV composite membranes were characterized by SEM. The PV performance was represented by the flux and the separation factor. The effects of feed temperature and feed concentration on PV performance were also studied. When feed concentration was 4.00wt%, NR/SBS blend membrane which got the best performance obtained membrane separation factor of 28.8 with total flux of 2695.2 g/mh at 70 ℃.

2 Experimental

2.1 Materials

Toluene, methanol (MeOH) and

n

-butanol were analytical pure and purchased from Shanghai Chemical Reagent Company. SBS (YH-791H, S/B=3/7) was supplied by Baling Company, SINOPEC. Unvulcanized NR was supplied by Hainan Rubber Company. PVDF supported membrane was purchased from Jiangsu Peier Membrane Company.

2.2 Preparation of membranes

The membranes were prepared by solvent evaporation. Firstly, solution A was obtained by dissolving 2 g of unvulcanized NR in 48 g of toluene at 80 ℃ for 12 h, and solution B was obtained by dissolving 1 g of SBS in 9 g of toluene at 50 ℃ for 6 h. For the unvulcanized NR membrane, a certain amount of solution A was cast onto the PVDF supported membrane after sonication for 2 minutes. For the SBS membrane, a certain amount of solution B was cast onto the PVDF supported membrane after sonication for 2 minutes. For the NR/SBS blend membrane, solution A and solution B were mixed and stirred at 80 ℃ for 2 h, and then a certain amount of the mixture solution was cast onto the PVDF supported membrane after sonication for 2 minutes. All the membranes were dried at room temperature for 24 h, and at 70 ℃ for another 24 h.

2.3 Thermal gravimetric analysis (TGA) measurement

TGA was conducted using a TG 209 F1 instrument (NETZSCH, Germany) under nitrogen atmosphere with a flow capacity of 20 mL/min. Scanning was performed at a heating rate of 10 ℃/min from ambient temperature to 800 ℃. The samples that were analyzed using an AlOcrucible weighed approximately 5 mg after drying.

2.4 Scanning electron microscopy (SEM) measurement

The morphology of the PV membrane was examined by SEM (Zeiss Merlin Compact), which was operated at 5 kV. The composite membrane was coated with a thin layer of sputtered gold after quenching with liquid nitrogen.

2.5 Degree of swelling measurement

The NR/SBS film was cut into 2 cm × 2 cm × 0.5 mm (length × width × thickness) and then dipped into a swelling solution for 10 days at 30 ℃. The weight of the films before and after absorption was measured within ± 0.1 mg at room temperature. The degree of swelling (

S

) was determined using Eq.(1):

where,

m

refers to the mass (in grams) of the dry polymer and

m

refers to the mass (in grams) after dipping. The film with the larger

S

has larger solution absorption

2.6 Mechanical properties measurements

Mechanical tensile tests were performed on a universal testing machine (Shimadzu, AGS-J, capacity 10 kN) at 25 ℃. The tensile strength and elongation at break were measured according to ASTM C1184-2000a, and an average of at least five effective measurements for SBS/NR film was recorded.

2.7 PV experiments

The PV experiments were conducted as previously reported. The feed flow rate of 30 L/h was sufficient to minimize concentration polarization. The permeate solution was collected in a trap immersed with liquid nitrogen on the downstream side at a fixed time of intervals, which the vacuum system was about 0.1 atmospheric pressure. The concentration of the feed mixtures varied from about 1.6wt% to 4wt% butanol and the temperature of PV experiments was in the range of 30 to 70 ℃. The composition of the permeate solution was determined by a gas chromatography with the addition of a proper amount of methanol and then calculated by Eq.(2):

where,

X

refers to the butanol concentration (%) in the permeate;

m

and

m

refer to the weight (in grams) of the permeate and MeOH added to the permeate;

A

and

A

refer to the peak areas (in millivolt minute) of butanol and MeOH peaks;

λ

refers to the response factor of butanol to MeOH that was measured using butanol and MeOH mixture of four known different concentrations.Based on the PV data, the separation performance of the membranes was assessed in terms of the flux of production (

J

, stands for the processing capacity of the membrane) and separation factor (

α

stands for the selectivity of the membrane).

J

and

α

were calculated by Eqs.(3) and (4):

where,

J

and

m

refer to the flux and weight (in grams) of component

i

in the permeation;

s

and

t

refer to the effective area (m) of the membrane and operating time (h) respectively;

X

,

X

,

Y

, and

Y

refer to the molar fraction of butanol in the permeation and feed, molar fraction of water in the permeation and feed, respectively. All the experiments were repeated at least three times and then the results were averaged.

3 Results and discussion

3.1 Thermodynamic property of membrane

The TG curves are shown under nitrogen atmosphere in Fig.1. The initial decomposition temperature of all the films was above 280 ℃. Pure NR film shown the fastest thermal decomposition rate and almost completely decomposed at about 480 ℃. The thermal decomposition rate of NR/SBS film was between that of pure SBS and NR film in the range of 280-480 ℃ and was almost the same as that of SBS after 480 ℃. The results of TGA shown excellent thermodynamic stability of SBS/NR film.

Fig.1 TG curves of films: (a) NR; (b) NR/SBS; (c) SBS

3.2 Mechanical property of membrane

The results of mechanical tests are shown in Table 1. The blend film, combining the elasticity of NR with the rigidity of SBS, had strong impact resistance, and exhibited excellent mechanical properties.

Table 1 The results of mechanical tests

Sample Tensile strength/MPa Elongation at break,×102/%NR 2.65 9.42 NR/SBS 10.04 9.94 SBS 26.11 22.31

Table 2 The swelling degree of SBS/NR film in pure water and butanol

Butanol concentration /wt% S/% Solubility parameter /((cal/m3)1/2)0 (pure water) 2.10 23.4 100 (pure butanol) 25.01 11.4

3.3 Morphology of membrane

The Morphologies of NR/SBS blend membrane was characterized by SEM. As is shown in Fig.2, the PV membrane consisted of two layers of polymer with different morphologies. A layer of relatively dense blend membrane (in the lower layer) with the thickness of about 37 μm was casted on a layer of porous PVDF supported membrane (in the upper layer). There was no obvious phase separation observed between the interface of two layers of membranes.

Fig.2 SEM image of NR/SBS blend membrane at 1000 times magnification

3.4 Swelling property of membrane

The swelling degree of NR/SBS film in different mass fractions of butanol aqueous solution are shown in Fig.3 and Table 2. The swelling degree of SBS/NR film in butanol aqueous solution increased with increasing butanol concentration, and the swelling degree of SBS/NR film in butanol was far more than that in water. According to Flory-Higgins Eq.(5), the interaction between solute molecules and polymers is directly related to the difference in their solubility parameters. The smaller the difference in solubility parameters between the solute molecule and the polymer, the stronger the interaction between the two and the better their compatibility.

Fig.3 The swelling degree of NR/SBS film in different concentrations of butanol aqueous solution

Table 3 Comparison of PV performance with other open literatures

Membrane Temperature/℃ Concentration/wt% Flux/(g/m2h) Separation factor PSI×10-3/(g/m2h) Reference PDMS/HFA 68 2.75 1519.6 64.4 96.3 [12]PDMS/ZIF-8 55 1.5 1039.04 56 57.1 [13]PTMSP 37 1 587 46.3 26.6 [16]HTPB-PEIL2-PU 70 3 160 29.2 4.5 [18]PEBA/MCM-41 35 2.5 500 25 12 [19]NR/SBS 70 4.00 2695.2 28.8 74.9 This work

where,

X

refers to the interaction parameter between the solute molecules and the polymer;

V

refers to the volume fraction of the solute molecule;

δ

and

δ

refer to the solubility parameter of the polymer and the solute molecule.

The solubility parameter of NR is about 8.2 (cal/m), and the solubility parameter of SBS lies in the range of 8 to 9 (cal/m)for the difference of block ratio between butadiene and styrene. Therefore, the solubility parameter of blend polymer which can be calculated by weighting lies in the range of 8 to 9 (cal/m). The difference of solubility parameter between the blend polymer and butanol is much smaller than that between the blend polymer and water, meaning that the SBS/NR blend film had a good affinity with butanol and weak interaction force with water.

3.5 PV performance

The flux and separation factor were used to characterize the pervaporation properties of all the membranes.

3.5.1 The influence of membrane composition on PV performance

The PV performance is based on the membrane properties. In order to study the effect of membrane composition on PV performance, taking the feed concentration of about 4wt% as an example. As is shown in Fig.4, pure NR membrane exhibited high total flux and separation factor. However, the water flux increased fast and the separation factor decreased much as temperature increased because of the low mechanical strength of unvulcanized NR. When the membrane was swollen by

n

-butanol, the structure of unvulcanized NR was permanently deformed and its internal pores became larger under negative pressure, which is good for water transport. For pure SBS membrane, it exhibited low flux and high separation factor due to its tight arrangement of polymer chains and super hydrophobicity. As PV experiment goes on, its structure can remain stable for a long time due to the good mechanical properties. For NR/SBS blend membrane, it exhibited high flux and considerable separation factor, and the PV performance get better with increasing temperature, owing to the formation of IPN-like structure which endowed the blend membrane a stable mass transfer channel.

Fig.4 PV performance of different membrane compositions: (a) Total flux; (b) Butanol flux; (c) Water flux; (d) Separation factor

3.5.2 Effect of feed temperature on PV performance of NR/SBS blend membrane

Temperature is an important parameter as it significantly affects both sorption and diffusion. As is shown in Fig.5, flux and separation factor always increased with increasing feed temperature.

Fig.5 PV performance of the NR/SBS blend membrane: (a) Total flux; (b) Butanol flux; (c) Water flux; (d) Separation factor

The vapor pressure of the components in the feed increased as feed temperature increased, resulting in increasing driving force of mass transfer. Meanwhile, more solute molecules were activated with the increase of the feed temperature, promoting the diffusion process. Therefore, all the flux increased as feed temperature increased.

Based on the Arrhenius Eq.(6), the calculated permeation activation energy of butanol and water is shown in Fig.6, which are 52.54 and 22.86 kJ/mol,respectively. The permeation activation energy of butanol is much larger than that of water. When feed temperature was low, the water molecules were more easily activated. However, as feed temperature increased, more butanol molecules were activated, resulting in increasing separation factor.

Fig.6 Arrhenius fit of butanol and water fluxes at the feed concentration of 2.99wt%

where,

J

refers to the permeate rate constant,

E

refers to the activation energy,

R

refers to the gas constant and

T

refers to the absolute temperature.

3.5.3 Effect of feed concentration on PV performance of NR/SBS blend membrane

As is shown in Fig.5, butanol flux increased as feed concentration increased consistently; water flux and total flux almost increased at feed concentration below 2.99wt%, and decreased when feed concentration increased from 2.99wt% to 4.00wt%. The permeation activation energy of

n

-butanol and water using NR/SBS blend membrane at different feed concentration was calculated (in Fig.7). The permeation activation energy of

n

-butanol decreased slightly with increasing feed concentration. While the permeation activation energy of water decreased at feed concentration below 2.99wt%, and increased when feed concentration increased from 2.99wt% to 4.00wt%.

Fig.7 The permeation activation energy of n-butanol and water using NR/SBS blend membrane at different feed concentration

Based on the results in the swelling experiment, the interaction between NR/SBS blend membrane and butanol is much larger than that between NR/SBS blend membrane and water, and the degree of swelling increased with increasing butanol concentration. Meanwhile, the driving force of butanol transfer increased as butanol concentration increased. Therefore, butanol flux increased with increasing butanol concentration. Although the membrane has a poor affinity for water, the interaction between butanol and water is a little bit strong. Water molecules will enter the membrane along with the butanol molecules. When the feed concentration was low, as the concentration of butanol increased, more butanol molecules entered the membrane, and the stronger interaction between the butanol molecules and water molecules caused more water molecules to enter the membrane, and the growth rate of water molecules entering the membrane was greater than that of butanol molecules. However, the number of channels in the membrane is limited. When more butanol molecules entered the membrane to occupy the channels, resulting in less water molecules entering the membrane, water flux decreased. The separation factor accordingly decreased at feed concentration below 2.99wt%, and then increased.

3.5.4 Comparison of PV performance with other literatures

Table 3 lists the PV performance of some different membranes for

n

-butanol recovery. In order to evaluate the overall performance of the membrane, the pervaporation separation index (PSI) is introduced and calculated by Eq.(7). Compared with the other membranes in Table 1, NR/SBS blend membrane possesses prominent performance in terms of separation of butanol from its dilute aqueous solution. It got the highest flux with 2 695.2 g/mh for the separation of 4.00% butanol at 70 ℃. The PSI (×10) with 74.9 g/mh indicated that the NR/SBS blend possessed good overall performance.

4 Conclusions

In order to develop a stable PV membrane with high performance for the separation of butanol and water, NR/SBS blend membrane was prepared by solution blend. The obtained membrane possesses good thermodynamic and mechanical properties, which was characterized by TGA and tensile tests. The flux and separation factor increased significantly with increasing feed temperature. And the butanol flux increased as feed concentration increased continuously. The blend membrane which got the best performance obtained membrane separation factor of 28.8 with total flux of 2 695.2 g/mh at the feed temperature of 70 ℃ and the feed concentration of 4.00wt%.