Application of PDMS pervaporation membranes filled with tree bark biochar for ethanol/water separation

Abstract

In order to improve separation performance of pervaporation membranes, we explored, for the first time, using tree bark biochar (BB) as filler for preparing polydimethylsiloxane (PDMS) composite membranes to separate ethanol from water. The composite membranes prepared by adding BB fillers obtained from different temperatures had different pervaporation performances. It was found that the addition of the three types of modified BB particles could all significantly improve pervaporation performance of the composite membranes. However their effects on permeation flux and separation factor were varied. Compared to BB obtained at 300 °C and 900 °C, BB obtained at 600 °C resulted in the membrane with the highest pervaporation performance with a separation factor of 9.92 and a permeation flux of 219.20 g m⁻² h⁻¹ at 40 °C in 10% ethanol solution. This study demonstrates the promise of biochar as a novel type of biomass-derived nanoparticles suitable for applications in pervaporation separation membranes.

---

Introduction

There is a strong interest worldwide in producing bioethanol from renewable biomass resources through fermentation.¹ However, during the fermentation process, if the concentration of ethanol reaches above 8%, the activities of the yeast cells are suppressed leading to termination of the fermentation process as a whole.² As a result, ethanol has to be continuously removed from the fermentation broth in order to improve the bioethanol yield.

Pervaporation is a membrane separation technique commonly used to separate and remove ethanol from the fermentation broth. Its mechanism is based on taking advantage of different permeation rates of the components through a nonporous polymeric or inorganic membrane.³ Polydimethylsiloxane (PDMS) has been reported to be one of the best choice materials for preparing pervaporation membranes.⁴ According to a review paper,⁵ when removing ethanol by pervaporation, the permeation flux of unmodified PDMS membranes were ranged from 0.001 to 1 kg m⁻² h⁻¹ and the selectivity were ranged from 4.4 to 10.8. Some fillers, such as silicalite-1,^6–8 polyphosphazene nanotubes,⁹ and carbon black,¹⁰ etc. were added into PDMS membranes to improve the permeation ability of ethanol. Among these, carbon black has some particular advantages in terms of its porous structure, adsorptive property, cost, and ease in production.¹¹

However, hydrophobicity is a key factor in order for fillers to impart high separation factor in pervaporation membranes. Since carbon black surfaces are hydrophilic with very little organic functional groups, it is difficult to obtain high selectivity nanocomposite membranes using carbon black as the filler. On the other hand, tree bark biochar (BB) is prepared by pyrolyzing bark biomass under anoxic conditions. The resulting biochar particles have unique pore structures with enriched organic functional groups on the surfaces.¹²

Compared to carbon black, the remaining organic functional groups on the biochar surfaces have potentials to improve the compatibility between the biochar particles and PDMS, as these organic groups can be easily modified to enhance hydrophobicity. Besides favorable surface chemistry, biochar also has attractive adsorptive properties. However, no study in the literature has explored biochar for applications in pervaporation membranes.

In this study, tree bark biochar was added in PDMS to prepare the separation layer of the membranes. The PDMS/biochar separation layers were then supported by cellulose acetate (CA) microfiltration membranes. The resulting composite membranes (PDMS/CA) were used to separate ethanol by pervaporation separation. The effects of the type of BB on separation performance of the resulting nano-composite PDMS pervaporation membranes were investigated and discussed.

Experimental

Materials

PDMS (107#RTV) was purchased from Sigma-Aldrich (USA). Tetraethylorthosilicate, dibutyltin dilaurate, n-hexane, and ethanol were obtained as analytical reagents from Sigma-Aldrich (USA). NH₂(CH₂)₃Si(OC₂H₅)₃ (KH-550) was purchased from Sigma-Aldrich in chemically pure form (China). Cellulose acetate microfiltration membranes with an average pore size of 0.45 μm were purchased from Fisher (USA) and used as supports. BB were made at our laboratory at 300 °C, 600 °C and 900 °C using an experimental pyrolysis system (GSL-1100X, USA). The particle sizes of BB prepared at 300 °C, 600 °C and 900 °C were around 100 nm, 50 nm, and 50 nm, respectively, according to SEM images.

Preparation and modification of biochar

For preparation of biochar, tree bark was first grounded into powder and oven dried at 103 °C for 24 h, and then was used to prepared BB at 300 °C, 600 °C, and 900 °C temperatures under nitrogen environment, respectively.

The preparation steps for grafting KH-550 on BB are listed below:

• Bark was dried at 120 °C for 24 h in an vacuum oven.

• BB was extracted via Soxhlet extraction with toluene as an extractant under N₂ atmosphere for 48 h to remove surface oil coming from the pyrolysis process.

• 10 g of toluene-extracted BB was mixed with 100 mL methanol (37 wt%) and 10 mL sodium hydroxide (20 wt%), and stirred at 50 °C for 5 h. Then, the reaction was terminated by cooling to room temperature. Subsequently the BB was retrieved and washed with distilled water and dried in a vacuum oven.

• The dried BB obtained after the previous step was grafted with KH-550 in an n-hexane solution.

The grafting mechanism is shown in Scheme 1.

Scheme 1 Schematic representation of the coupling reaction between BB and KH-550 as well as its subsequent coupling with PDMS.

Preparation of composite membranes

The schematic of the composite membranes preparation process is shown in Fig. 1. The thickness of the separation layer of the composite membranes was approximately 20 μm.

Fig. 1 Schematic illustration of preparation steps for the composite membranes.

BB and membrane characterization

FT-IR analysis. Fourier transform infrared (FT-IR) spectroscopy (Spectrum400, Perkin-Elmer) analysis was performed to investigate the chemical characteristics of the modified BB obtained at 300 °C, 600 °C and 900 °C. The FT-IR spectra were obtained in the wavelength range from 400 to 4000 cm⁻¹.

Contact angles analysis. The contact angles of the PDMS composite membranes were measured at room temperature to characterize hydrophobicity of the membranes using a contact angle instrument (Mitutoyo519-109, USA). The liquids used for the sessile drop tests were distilled water and anhydrous ethanol, respectively. In the case of anhydrous ethanol, the measurement was carried out quickly right after the placement of sessile drop to minimize potential experimental error.

Scanning electron microscopy. Surface morphology and cross-sectional images of the prepared PDMS/BB composite membranes were obtained using a XL30 Scanning Electron Microscopy (FEI, USA). The membranes were fractured in liquid nitrogen to prepare cross-sectional sample of PDMS/BB composite membranes. The BB powders and surface morphology and cross-sectional samples of PDMS/BB composite membranes were gold-coated before the measurements.

Thermogravimetric analysis. The separation layer of the composite membranes was heated from room temperature up to 900 °C at a heating rate of 10 °C min⁻¹ in nitrogen, with a flow rate of 20 mL min⁻¹. All thermal degradation data were obtained from the TGA curves. Thermogravimetric analysis (TGA) was performed using a TGA-Q500 instrument (TA Instruments, USA).

Liquid absorption. Liquid absorption ability of the separation layer of the composite membranes was measured for understanding the interactions between ethanol molecular and the membranes. The separation layers were immersed in an aqueous solution containing 10 wt% ethanol at 40 °C. The separation layers were then removed from the solutions when they were saturated. The membranes were weighed within ±10⁻⁴ g accuracy immediately.

The curves of absorption were plotted. The average swelling degree of the active layer of the composite membrane was calculated based on the weight difference between the saturated wet film (W_s) and the weight of the dry separation layer (W_d) as shown in eqn (1):

Measurement of ethanol diffusion coefficient. The ethanol diffusion coefficient through the separation layer of the composite membranes were measured by a similar technique that was reported by Marais et al.¹³ In this experiment, the samples were measured using a microbalance with 1 mg accuracy. First, the separation layers were immersed in ethanol for 48 h to reach the equilibrium state, then the layers were taken out, quickly wiped with filter papers, and weighed rapidly.

The diffusion coefficients of ethanol in the separation layer of the composite membrane were calculated using the following eqn (2):

where ΔM is the ethanol absorbed in the membrane at time t, ΔM_eq is the ethanol absorbed at the equilibrium state. D is the diffusion coefficient of ethanol diffused into the film and L is the initial thickness of the membrane. The diffusion coefficients of ethanol were calculated from the slopes of semi-log plots of the left side values of eqn (2) versus time.

Pervaporation. The schematic of the pervaporation testing apparatus is shown in Fig. 2. The composite membrane was sealed in a stainless steel module and dipped directly in the feed solution. Above the membrane module, a small submerged pump was used to reduce the concentration polarization. The effective membrane area was 39.6 cm². The temperature of the feed tank was controlled at 40 °C with a water bath. The concentration of ethanol aqueous solution was 10 wt%. A vacuum pump was used to maintain the downstream pressure at 20 kPa. In addition, nitrogen gas was used to flash out the permeation vapor. After reaching a steady state, the permeated solution was collected in a liquid nitrogen trap. The permeation mass was measured with an analytical balance. The mass concentrations of the feed and permeated solution were analyzed with an Abbe refractometer (WAY-2W, China).

Fig. 2 Schematic illustration of the pervaporation measurement apparatus.

Permeation flux and selectivity are two key parameters to evaluate the pervaporation performance of the membranes. The permeation flux (J g⁻¹ m⁻² h⁻¹) is defined as follows:

where Q, A, and t represent the permeation mass (g), the membrane area (m²) and the operation time (h), respectively. Separation factor (α) expresses the membrane's selectivity to ethanol and is defined by the following equation:where X_A and X_B are the mass fractions of ethanol and water in the feed, and Y_A and Y_B are the mass fractions of ethanol and water in the permeation, respectively.

Results and discussion

FT-IR analysis of the modified BB

The chemical bonding between three kinds of BB nanoparticles and silane coupling agents was characterized by FT-IR as shown in Fig. 3(a)–(c). It can be seen that in Fig. 3, from 300 °C to 900 °C, there were less CH₂ stretching (2800–3000 cm⁻¹) and C–C stretching (1500–1600 cm⁻¹), so the organic groups reduced gradually with the increase in preparation temperature. Silane-coupled BB showed that the characteristic peaks at 2930 cm⁻¹ (CH₂ stretching) and 1470 cm⁻¹ (CH₂ bending) became higher. Additionally, a sharp characteristic peak of the Si–O at 1081 cm⁻¹ can be observed for the modified BB. From the FT-IR analysis, it can be concluded that the functional groups of silane coupling agents were successfully introduced onto the BB surface.

Fig. 3 FT-IR spectra of the unmodified and modified BB particles obtained at 300 °C (a), 600 °C (b) and 900 °C (c).

Morphology of the composite membranes

Fig. 4(a) shows that the particle size of BB prepared at 600 °C is around 50 nm. The surface and cross-sectional morphologies of 3 wt% BB-filled PDMS/CA membranes are shown in Fig. 4(b) and (c). The active layer was tightly adhered to the surface of the CA support layer. It can be seen that the modified BB particles were well dispersed in the PDMS matrix.

Fig. 4 SEM images of BB (600 °C) and PDMS/CA nano-composite membranes filled with 3 wt% of grafted BB (600 °C). (a) Biochar morphology of BB prepared at 600 °C (b) surface morphology of composite membranes (c) cross-sectional morphology of the composite membranes.

Contact angle analysis of the PDMS composite membranes

The Fig. 5(a) shows the water contact angle values at the surface of composite PDMS membranes containing grafted BB (0%, 1.5%, 3%, 4.5%) prepared at 300 °C, 600 °C, and 900 °C, respectively. It can be seen that for the modified membranes with BB prepared at three temperatures, the water contact angles were larger than those of the unmodified membranes. For each type of the modified membranes, with an increase in the BB content in the membranes, the water contact angle also increased. This indicated that the surface of the composite membrane became more hydrophobic due to the addition of BB particles. These modified BB particles contained hydrophobic groups after grafted with KH-550 coupling agent (NH₂(CH₂)₃Si(OC₂H₅)₃). Therefore, the surface hydrophobicity of the composite membranes was effectively improved.

Fig. 5 Contact angle of water (a) and ethanol (b) at the surfaces of PDMS composite membranes containing 3 wt% of grafted BB prepared at three different temperatures.

The hydrophobicity level of the modified membranes with BB obtained at the three temperatures followed the following order 300 °C > 600 °C > 900 °C. The reason was that BB prepared at lower temperatures had more hydrophobic organic groups after modification and with the increase in temperature, its organic groups were reduced.

Fig. 5(b) shows the values for ethanol contact angle at the surfaces of the modified PDMS membranes. With an increase in the BB content in the membranes, the ethanol contact angles decreased significantly for all membranes, which implied that the presence of the modified BB made the membrane more compatible with ethanol. However, there is almost no difference in ethanol contact angle values among the membranes contained different percentages of BB in this study.

Effect of BB on the ethanol absorption of the membranes

The active layers of the composite membranes were immersed into 5 wt%, 10 wt%, 15 wt%, 20 wt% and 25 wt% ethanol aqueous solutions separately for testing their swelling degrees. In Fig. 6, with the temperature of biochar preparation increased, the swelling degree of the composite membranes increased. The reason is that with the increase in pyrolysis temperature, the biochar particles became smaller. And it can be seen that the swelling degree of the composite membranes increased with the increase of the feed ethanol concentration. With a higher ethanol concentration in the feed liquid and a higher biochar preparation temperature, PDMS membranes had more opportunities to make contact and dissolve ethanol, thus, the swelling of the composite membranes increased. The increase was caused by a larger amount of dissolved ethanol, the weakening of the link in the PDMS membrane due to the interactions with ethanol, and the increased local free volume of the polymer. As a result of the increased swelling, the solubility of ethanol and water molecules in the membrane was improved. However, an excessive degree of swelling is not desirable since it will reduce the mechanical strength of the active layer of the membrane and the service life of the membrane itself. So, it is necessary to control the degree of swelling in membrane to an optimum.

Fig. 6 Swelling degrees of the grated BB modified composite membranes at different feed ethanol concentrations.

According to eqn (2), ln(1 − ΔM/ΔM_eq) is linearly dependent on time t as shown in Fig. 7. Thus, the slope for different BB could be obtained from the semi-log plot. According to the value of the slope, the diffusion coefficient of ethanol in pure PDMS membrane is 4.80 × 10⁻⁸ cm² s⁻¹. For the PDMS composite membranes containing 3 wt% modified BB prepared at 300 °C, 600 °C, and 900 °C, respectively, the diffusion coefficients were 3.82 × 10⁻⁷ cm² s⁻¹, 4.31 × 10⁻⁷ cm² s⁻¹, and 5.62 × 10⁻⁷ cm² s⁻¹, respectively. Adding modified BB nanoparticles improved the diffusion rate of ethanol in the PDMS membranes and improved PDMS membrane performance in terms of pervaporation separation factor. The calculated results showed that the four types of membranes had the diffusion coefficients in the following order: 900 °C > 600 °C > 300 °C > pure PDMS membrane.

Fig. 7 Diffusion coefficient of the grafted BB modified PDMS membranes.

Thermal stability

According to Fig. 8, thermal degradation curves of all composite membranes filled with BB showed a two-stage degradation response. From room temperature to 200 °C, the absorbed water and the hydroscopic moisture were removed. From 200 °C to 900 °C, the grafted component and volatiles were released from the composite membranes.

Fig. 8 TGA curves of the modified BB filled PDMS membranes.

CA support layers degraded at the first stage that was followed by the degradation of PDMS active layers filled with BB in the second stage. This observation implied that the PDMS matrix was well mixed with the BB particles and the addition of the modified BB improved the thermal stability of the composite membranes.

Addition of BB significantly increased the permeation fluxes of the PDMS membranes as shown in Fig. 9. As shown by other literature studies,¹⁴ the diffusion coefficients of permeation molecules will increase when the free volume of the composite membranes becomes higher. BB prepared at 300 °C presented the composite membrane with the largest permeation flux, followed by BB prepared at 600 °C and 900 °C, respectively. This is mainly due to the differences in ethanol absorption ability. The particle size of BB prepared at 300 °C is the largest among the three types of BB. Consequently, the BB obtained at 300 °C was more effective in disrupting the chain packing to increase the free volume of PDMS than the BB prepared at other temperatures. And that caused the permeation flux of the corresponding composite membrane to be the highest.

Fig. 9 Effect of the grafted BB content on the permeation flux of the composite membranes.

As shown in Fig. 10, with an increase of the modified BB content in the membranes, the separation factors first increased with the BB content and then decreased. Even though the filled BB particles disrupted the chain packing and increased the free volume, more BB particles would also inhabit the free volume cavities in the membranes. Therefore, the final result would be a balance between the above two competing effects. During the pervaporation process, the increase in the feed concentration caused more ethanol to dissolve into the membrane. Therefore, the swelling degree of PDMS membrane increased, and in turn, increased the free volume of PDMS. Thus, the diffusion rates of ethanol and water molecules through the membrane were higher which resulted in the larger flux as indicated in Fig. 10. However, the increase in the diffusivity of water was much larger than the increase in the diffusivity of ethanol since the molecular size of water is smaller than that of ethanol. Consequently, the separation factor decreased.

Fig. 10 Effect of modified BB content on the separation factor of the composite membranes.

From Fig. 10, it is also shown that the separation factor of PDMS composite membranes with BB obtained at 600 °C is similar to that of membranes with BB obtained at 900 °C, while the separation factor for PDMS composite membrane with BB obtained at 300 °C was much lower. The reason is closely related to the particle size. Because the particle size of BB at 300 °C was the biggest, the PDMS composite membranes with BB prepared at 300 °C had the worst dispersibility and the lowest separation factor among all membranes.

Fig. 11 illustrates the long-term stability of the composite membrane with 3 wt% BB (600 °C) in 10 wt% ethanol/water solutions at 40 °C. The permeation flux and separation factor of the composite membrane maintained almost constant values. It indicated that separation performance of these composite membranes would not be adversely influenced despite of the swelling of the separation layers.

Fig. 11 Effect of operation time on pervaporation performance of PDMS/CA composite membrane filled with 3% modified BB obtained at 600 °C.

Table 1 compares pervaporation performances of the PDMS composite membranes filled with BB prepared in this study with some published data on PDMS composite membranes filled with other types of fillers for the separation of ethanol from water. Compared to nano-silicalite, silicalite, zerolite, fumed silica, and carbon black, BB filled membranes had very promising performance.

Table 1 Ethanol–water separation performance of PDMS membranes with different fillers

Fillers	Filler content (wt%)	Ethanol feed concentration (wt%)	Temperature (°C)	Separation factor	Flux (g m^−2 h^−1)	Reference
Nano-silicalite-1	30	6	35	15.7	—	15
Silicalite-1	60	5	22.5	16.5	51	15
Silicalite-1	77	5	22	37	150	16
Silicalite-1	50	5	50	29.3	—	17
Silicalite-1	15	3	41	4.8	170	18
Fumed silica	20	5	40	7.0	—	19
USY	50	—	30	16.1	—	20
ZSM-5	30	—	35	5	250	21
ZSM-5	50	—	30	14	—	20
Zeolite Y	30	—	35	4.5	750	21
ALPO-5	50	—	30	5.2	—	20
[Cu^II2(bza)4(pyz)]n	3	5	25	2.3	23	22
Carbon black	4.5	13.7	20	10.1	127.32	23
Carbon black	10	6	35	9	49.8	24
Carbon black	10	—	35	9	—	11
Carbon black	1.5	13.73	30	9	189	24
Tree bark biochar	3	10	40	9.9	219	This study

---

Conclusions

In this study, for the first time, tree bark biochar modified with a silane coupling agent (KH-550) was used as fillers in PDMS membranes for separation of ethanol and water. The effect of the amount of biochar in the membrane on the pervaporation characteristics of the nano-composite PDMS/BB membranes was investigated. It was found that the modified BB particles dispersed well in the PDMS matrix and the addition of the modified BB particles significantly improved pervaporation performance of the resulting PDMS membranes. The permeation flux of the composite membranes decreased with the biochar preparation temperatures ranging from the highest to the lowest from 300 °C, 600 °C to 900 °C. For the separation factor, trend was reversed with the membranes with the biochar prepared at the highest temperature level being the largest. The best overall performing membranes had 3 wt% of modified biochar (that was prepared at 600 °C) with a separation factor of 9.92 and the permeation flux maintained at 219.20 g m⁻² h⁻¹ in a 10 wt% ethanol/water solution at 40 °C. This study has shown that biochar is a novel type of biomass-derived nanoparticles highly promising for applications as fillers in ethanol-water pervaporation separation membranes.

Acknowledgements

Authors would like to acknowledge financial support from China Scholarship Council (CSC) of the Ministry of Education, P. R. China and Professor Yan's NSERC-Discovery Grant. This study was supported by “The Fundamental Research Funds for the Central Universities”, 2572014AB15. Assistance in preparation of biochar by Ms Sossina Gezahegn is appreciated. We would also like to thank Prof. S. Thomas for providing access to his lab pyrolysis facility.

References

1. L. Vane, V. Namboodiri and R. Meier, J. Membr. Sci., 2010, 364, 102–110.
2. F. Bai, W. Andersona and M. Moo-Young, Biotechnol. Adv., 2008, 26, 89–105.
3. X. Zhan, J. Li, J. Huang and C. Chen, Appl. Biochem. Biotechnol., 2010, 160, 32–42.
4. M. Rezakazemia, K. Shahidia and T. Mohammadi, Desalin. Water Treat., 2015, 54, 1542–1549.
5. L. Vane, J. Chem. Technol. Biotechnol., 2005, 80, 603–629.
6. M. Jia, K. Pleinemann and R. Behling, J. Membr. Sci., 1992, 73, 119–128.
7. X. Chen, Z. Ping and Y. Long, J. Appl. Polym. Sci., 1998, 67, 629–636.
8. L. Vane, V. Namboodiril and T. Bowen, J. Membr. Sci., 2008, 308, 230–241.
9. Y. Huang, P. Zhang and J. Fu, J. Membr. Sci., 2009, 339, 85–92.
10. I. Vankelecom, J. Kinderen and B. Dewitte, J. Phys. Chem., 1997, 101, 5182–5185.
11. S. Shi, Z. Du and H. Ye, Polym. J., 2008, 28, 949–955.
12. J. Lehmann, M. Rillig and J. Thies, Soil Biol. Biochem., 2011, 43, 1812–1836.
13. S. Marais, M. Métayer, T. Nguyen, M. Labbé, L. Perrin and J. Saiter, Polymer, 2000, 41, 2667–2677.
14. T. Merkel, B. Freeman and R. Spontak, Science, 2002, 296, 519–522.
15. B. Moermans, W. Beuckelaer and P. Jacobs, Chem. Commun., 2000, 24, 2467–2468.
16. M. Jia, K. Pleinemann and R. Behling, J. Membr. Sci., 1992, 73, 119–128.
17. X. Chen, Z. Ping and Y. Long, J. Appl. Polym. Sci., 1998, 67, 629–636.
18. A. Dobraka, A. Figolic and S. Chovaua, J. Colloid Interface Sci., 2010, 346, 254–264.
19. X. Tang, R. Wang and Z. Xiao, J. Appl. Polym. Sci., 2007, 105, 3132–3137.
20. B. Adnadjevic, J. Jovanovic and S. Gajinov, J. Membr. Sci., 1997, 136, 173–179.
21. I. Vankelecom, D. Depre and S. Beukelaer, J. Phys. Chem., 1995, 99, 13193–13197.
22. S. Takamizawa, C. Terajima and M. Kohbara, Chem.–Asian. J., 2007, 6, 837–848.
23. S. Shi, Z. Du and H. Ye, Polym. J., 2006, 38, 949–955.
24. V. Vankelecom, J. Kinderen and B. Dewitte, J. Phys. Chem., 1997, 101, 5182–5185.

---