The use of solvent-soaking treatment to enhance the anisotropic mechanical properties of electrospun nanofiber membranes for water filtration

Abstract

Cellulose acetate derived from bamboo cellulose (B-CA) was prepared via a typical acetylation process and used to fabricate the aligned electrospun nanofibrous membranes (ENMs). A solvent post-treatment by soaking the B-CA ENMs in a mixture solution with different ethanol/acetone volume ratios was also performed to improve the mechanical properties of the membranes. In this study, we showed that this approach can enhance the bonding in the fiber membranes by solvent-induced fusion of inter-fiber junction points and change the degree of molecular orientation of the polymer, which are both considered to have important influences on the mechanical properties of ENMs. The treated membranes exhibited a significant enhancement in mechanical strength while retaining high hydraulic permeability. Increased performances will improve the prospects of ENMs as an emerging material in the filtration space for handling and scale-up manufacturing.

---

1. Introduction

The electrospinning process is an easy and versatile technique for fabricating non-woven mats containing nano- to microscale fibers.^1–4 Recent research has focused on the use of this technique to fabricate nanofibers for various applications in extensive areas, such as filtration systems, biomedical tissue templates, and drug delivery membranes.^5–7 Given their several advantageous attributes, such as a high-porosity and interconnected open-pore structure, submicron pore sizes, and a large surface area-to-volume ratio, electrospun nanofibrous membranes (ENMs) show great potential in membrane filtration technology.^8–11 These advantages allow ENM-based filter media to present high permeability while offering high filtration efficiency.¹²

However, ENMs generally possess low mechanical properties because of their small fiber diameters, non-woven and highly porous structure, and weak bonding among the fibers, thus presently limiting their applications.¹¹ Considering the relatively low mechanical properties of ENMs, considerable research has recently been conducted. Reinforcing materials, such as carbon nanotubes, graphite nanoplatelets, or layered silicate, are added to polymer solution and electrospun into nanofibers.^13–16 The obtained composite nanofibers usually show enhanced Young's modulus; however, the problems of complicated and costly preparing process limit their widespread application. Another method is to collect electrospun nanofibers into bundles or yarns, and the aligned fibers exhibited significantly improved strength.¹³ The yield of electrospinning is presently relatively small. Although many efforts have been made in this area, preparing nanofiber yarn-based materials for wide applications is time consuming and costly. The yarn structure also has lower porosity than the traditional non-woven structure and is less suitable for filtration application. Increasing the mechanical properties of electrospun nanofiber via post-drawing treatment has recently attracted the attention of several researchers.^17–19 Post-treatments, such as stretching, twisting, and annealing, are also efficient methods.^20–22 The method to achieve this goal involves heating the mat above the glass transition temperature of the electrospun polymer but below its melting temperature, which can cause inter-fiber fusion. Control of post-drawing temperature in the vicinity of the glass transition temperature (T_g) could also achieve the maximum effect in reinforcing.²³ A possible drawback of this approach, however, is the axial shrinkage of heat-treated membranes caused by entropic relaxation of stretched polymer chains.²⁴ Meanwhile, stretching the non-woven nanofibrous membranes is proven to be impractical owing to their complicated texture and relatively low strength. In our previous study, we demonstrated that mixed solvent post-treatment could lead to a dense and compact fibrous structure with a large fiber diameter, which resulted in the increased tensile strength of nanofibrous mats.²⁵ However, mixtures with different solvent volume ratios were used to treat the nanofibrous mats and their mechanical behaviour has not been reported in the literature.

Recently, our group and other researchers' studies have indicated that nanofibers prepared via electrospinning using renewable cellulosic precursors have been receiving increasing attention because of their abundance and low cost.^26–31 Cellulose is a cheap, abundant, sustainable, and readily available natural material that has a long history in fiber manufacturing.³² Bamboo is a fast-growing plant abundantly found in tropical countries.³³ This plant is enriched with bamboo cellulose (BC) that can provide an inexhaustible source to prepare nanofibers. However, similar to normal cellulose, BC has low solubility in conventional solvent systems because of the strong inter- and intra-molecular interactions that originate from the hydrogen bonds and rigid backbone structure. Therefore, cellulose derivatives, such as cellulose acetate, methyl cellulose, hydroxyl propyl cellulose, and ethyl cellulose,^34–39 have been extensively exploited to improve the solubility of cellulose to enhance its electrospinning capability.

In this study, cellulose acetate derived from BC (B-CA) was prepared via the typical acetylation process and then used to fabricate electrospun-aligned ultrafine cellulose-based nanofibrous membranes by using a high-speed collection drum. We demonstrated a minimally post-treatment approach to improving the mechanical properties of B-CA ENMs. This post-treatment involved the use of the solvent treatment method by soaking ENMs in a mixture solvent with different ethanol/acetone volume ratios. Solvent-soaking measurement can enhance the bonding at junction points in the fiber mat by welding or soldering the fibers together and change the degree of molecular orientation of polymer, which are considered having an important influence on the mechanical properties of the nanofiber. Scanning electron microscopy (SEM), polarized Fourier transform infrared (FTIR) spectroscopy, and dynamic mechanical analysis (DMA), were used to determine the membrane morphology, molecular orientation, and mechanical properties after solvent treatment. The hydraulic permeability test was also carried out on the treated B-CA ENMs to determine the treatment affected the membrane performance. Finally, the treated membranes exhibited significantly enhanced tensile strength and modulus while retaining high hydraulic permeability.

2. Experimental

2.1. Materials

N,N-Dimethylacetamide (DMAc, Sigma-Aldrich), acetone (Chem-Supply), and ethanol (Chem-Supply) were used without further purification. Bamboo cellulose (BC) was prepared and purified from commercial bleaching bamboo pulps provided by the Guangxi Jia Yu Paper Pulp Co.

2.2. Preparation of cellulose acetate form bamboo cellulose

BC acetylation was carried out using a typical process as follows (see Fig. 1A):^26,40 the dried bamboo cellulose was impregnated in a solution of acetic acid glacial and acetic anhydride containing sulfuric acid as a catalyst at a fixed temperature of 50 °C. After reaction for about 2.5 h, the solution system was then added with the hydrolysis mixture (acid catalyst and acetic acid aqueous solution) while stirring for hydrolysis to obtain the desired degree of substitution. Followed by precipitation with distilled water, the produced B-CA powders were separated by vacuum filtration and dried in a vacuum oven at 60 °C.

Fig. 1 Schematic of (A) acetylation reaction, (B) electrospinning, (C) solvent-soaking treatment process, (E) setup for measuring the hydraulic permeability of fibrous membranes, and (D) FTIR spectra of BC and B-CA.

2.3. Fabrication of B-CA ENMs

B-CA ENMs were prepared using a custom-built electrospinning setup (see Fig. 1B).^25,26 Details for the electrospinning system were showed as below. A homogeneous solution B-CA in acetone–DMAc (2/1, m/m) was prepared by stirring at ambient temperature. The concentration (37 wt%) was chosen based on previous studies which yielded relatively uniform fibers with fewer beads.²⁶ In this work, the electric potential and distance from syringe-tip to the collector were fixed at 22 kV and 15 cm, respectively. A grounded drum served as the collector. The collected B-CA nonwoven membrane was used for further treatment and characterization.

2.4. Solvent soaking post-treatment

The process for solvent soaking the spun B-CA ENMs is illustrated in Fig. 1C according to previous studies.²⁵ The treatment process began by carefully removing the samples from substrate (they are spun onto aluminum foil). And then they were cut into smaller coupons and hung up in a glass well-closed container filled with the mixed solution of ethanol and acetone (the ratio from 100 : 0 to 85 : 15). The coupons were soaked into mixed solution and then air dried in fume hood to keep solvent evaporation from the treated samples.

2.5. Morphology characterization and property measurements

The morphology of the electrospun nanofibers and changes in surface morphology between untreated and treated ENMs were observed under a NeoScope bench-top Scanning Electron Microscope (SEM, JCM-5000). Before examination, the samples were all sputter-coated with a thin layer of gold. Attenuated total reflectance Fourier transform infrared (ATR-FTIR) spectroscopy (Bruker VERTEX 70) was used to determine the vibration frequency changes in the adsorbents. Molecular orientation was evaluated by polarized FTIR spectrophotometer equipped with an MIR polarizer. Two hundred and fifty-six scans were accumulated with a resolution of 4 cm⁻¹. The mechanical properties were measured by using TA instrument Dynamic Mechanical Analysis (DMA, TA Q800) along fibre direction.

2.6. Membrane characterization

Hydraulic permeability was measured by using a lab setup shown in Fig. 1E.⁴¹ The ENMs was mounted with a purpose-made clamp, which was connected with a liquid supply system. A liquid was then passed through the matrix sample at a fixed flux (q = 20 mL h⁻¹), and the liquid pressure generated was recorded. The hydraulic permeability (f) of the porous ENMs was calculated according to the Darcy's law,⁴²where S is the flushing area of the porous ENMs, Δt is the thickness of the ENMs, and Δp is the differential pressure of the fluid between the two sides of the ENMs.

2.7. Statistical analysis

Data was analyzed by one-way analysis of variance (ANOVA) using the SPSS 19.0 (SPSS Inc., USA), followed by Duncan's multiple range test. Statistical significance was set at the 5% significance level (p < 0.05). All results are reported as mean ± standard deviation.

3. Results and discussion

3.1. Morphology analysis

B-CA ENMs (see Fig. 2B) were prepared from B-CA using an electrospinning technique. According to previous studies, aligned and randomly distributed nanofibers can be obtained by varying the rotation speeds of drum collector.²⁶ The SEM micrographs of electrospun B-CA nanofibers fabricated at different collection speeds are presented in Fig. 2C and D. When collected at 15.2 m s⁻¹, the electrospun B-CA membranes exhibited distinct fiber alignments (Fig. 2D) compared with those disordered fibers (Fig. 2C). To explain these findings, the ejected electrospun fibers were taken up on the rotating drum surface tightly, which resulted in a good fiber alignment at high collection speeds. In this work, we examined the effect of ethanol–acetone mixture with different ethanol–acetone ratio (e.g., 100/0, 95/5, 90/10, and 85/15, v/v) treatments on the mechanical properties of B-CA nanofibrous membranes at an optimal collection speed of 15.2 m s⁻¹.

Fig. 2 Photographs of (A) B-CA powders and (B) electrospun B-CA ENMs, and SEM micrographs of B-CA nanofibers prepared at different collection speeds: (C) 0 m s⁻¹ and (D) 15.2 m s⁻¹. Arrow in (D) indicates the axis of aligned fibers.

The surface morphologies of untreated and treated B-CA ENMs are displayed in Fig. 3. The results indicated that the solvent-soaking treatment method was effective in treating B-CA nanofibers. The SEM images show that the electrospun B-CA fibers were accumulated tightly after ethanol/acetone mixture solvent soaking, and a more compact fibrous membrane could be obtained, which were in good agreement with a previous study.²⁵ Compared with the morphology of the non-treated nanofibers, that of the treated B-CA nanofibers exhibited a significant morphology change after solvent-soaking treatment, and most of the nanofibers were connected to one another. This approach could enhance the bonding in the fiber mat by the solvent-induced fusion of inter-fiber junction points, as indicated by the arrows (see Fig. 3B). The effect might be caused by the fact that B-CA, a visco-elastic polymer, was swollen by the acetone solvent. The solvent could then condense at these junctions and slightly dissolve the polymer to facilitate fusion.

Fig. 3 SEM images of (A) untreated and (B) treated B-CA ENMs. The arrows point at fiber junction points.

3.2. FTIR spectra characterization

As illustrated in Fig. 1D, the FTIR spectrum proved the formation of B-CA after cellulose was acetylated in the acetylation medium (glacial acetic acid and acetic anhydride). Compared with the infrared spectrum of BC, the evidence of acetylation of the cellulose in the medium was provided by the increase in the intensity of the bands related to the acetyl group at ∼1745 cm⁻¹ attributed to C=O stretching, ∼1236 cm⁻¹ assigned to the C–O–C stretching of the acetyl group, and ∼1372 cm⁻¹ attributed to the –C–H blending in the methyl section of the acetyl group. Moreover, a blue shift of ∼19 cm⁻¹ observed for the band at ∼883 cm⁻¹ was assigned to the glucosidic linkages among the sugar units (ESI Table S1†). Another important aspect observed in the infrared spectroscopy, namely, the lowering of the absorbance of the OH stretching band in the region between ∼3013 and ∼3631 cm⁻¹, also presented further evidence of acetylation. The wavenumber at the maximum absorbance of the cellulose in this region shifted from ∼3332 to ∼3490 cm⁻¹ after acetylation in the medium, thereby indicating that OH stretching was induced primarily by water molecules.

Fig. S1 (in the ESI†) depicts the attenuated total reflection FTIR spectra of the untreated and treated B-CA NFMs. The spectrum of B-CA NFMs was typical, in which the absorption band assigned to the hydroxyl group and the hydrogen bond was observed at approximately 3500 cm⁻¹. Similarities were observed in the treated B-CA NFMs. The results indicated that the solvent-soaking treatment did not change the structure of nanofibers. However, the OH group in the treated NFMs led to a broad and weak band, which might be due to that the intramolecular hydrogen-bonds were broken partly after solvent-soaking treatment.

3.3. Molecular orientation

Apart from surface feature and structural changes, the molecular orientation within nanofibers was also examined. Polarized FTIR spectroscopy was used to determine the orientation of B-CA ENMs in the various samples (Fig. 4), and the results showed the polarized FTIR spectrum of solvent-soaking-treated before and after B-CA ENMs of angles 0° and 90°. For the untreated B-CA ENMs, the spectra showed almost no dependence on the sample angle (Fig. 4A). For the treated B-CA ENMs, the initial angle was set to arrange the fiber alignment in the same direction as that of the polarized FTIR light source (Fig. 4B). The FTIR spectrum changed the profile with polarized angle. After the solvent-soaking treatment, this difference in the B-CA ENMs became more obvious with the untreated B-CA ENMs as the reference substance.

Fig. 4 Variations in FTIR absorbance with polarized light angle relative to B-CA membranes: (A) untreated and (B) treated.

3.4. Mechanical properties

The mechanical properties of B-CA ENMs, such as tensile strength, modulus, and strain at break, as a function of ethanol/acetone ratio in mixed solvent, are shown in Fig. 5. An abrupt increase was observed in both the strength and modulus for the treated B-CA membranes with 95/5 (v/v) ethanol/acetone mixture solutions (Fig. 5A and B, p < 0.05). Further increasing the ratio of acetone in solution resulted in a decline in the mechanical property of the nanofiber. The mechanical property results could be correlated to the SEM observations in Fig. 3, in which the increase in tensile property and modulus in B-CA membranes might be attributed to the increased fibrous density and solvent-induced fusion at the fiber junction points. In other words, the small amount of acetone in the mixture solution during the post-treatment could keep the electrospun membrane relatively “wet” and “swollen”, thereby leading to the fusion among inter-fibers. When a specific amount of acetone was added, however, the dissolution of B-CA fibers was accelerated, which resulted in a slight decrease in the mechanical property of the treated ENMs. Although the post-treatment effect might decline when the ENMs were soaking in a mixture solution with a high acetone content, fiber–fiber fusion would still exist. This effect would ensure that the strength enhancement was maintained when the treated membranes ware placed into the solutions (the ratio of ethanol/acetone: 90/10 and 85/15, v/v) compared with untreated membranes. Moreover, the macromolecular chain orientation of the nanofiber was changed after mixture solution soaking, which might also help improve the mechanical properties of electrospun nanofibers.

Fig. 5 Mechanical properties of untreated and treated B-CA ENMs with different ethanol/acetone volume ratios in mixed solvent: (A) tensile strength; (B) modulus; and (C) strain at break (a–c values followed by different letters are significantly different at p < 0.05 using Duncan's multiple range test).

As shown in Fig. 5C, solvent-soaking-treated ENMs became less brittle, as indicated by the increased strain at break values. Similarly, an improvement in strain for B-CA was observed after treatment with 95/5 (v/v) ethanol/acetone solution (p < 0.05). This improvement was due to the fibers becoming plasticized by acetone, thereby making them more flexible. By contrast, the strain did not improve with increasing ratio of acetone for treated B-CA ENMs, which again verified that acetone facilitated the fiber dissolution and disorientation in the fiber network. This solvent-soaking treatment to enhance the mechanical strength and integrity of ENMs is comparable with previous methods reported by other researchers.^43–45

As discussed above, the solvent-soaking treatment plays an important role in determining the mechanical properties of B-CA ENMs. This solvent can facilitate the formation of a dense and compact fiber network structure in the fiber membrane after solvent immersion, which has effective points for hindering the slip of fibers by providing local physical or frictional entanglements. Mildly disoriented fibers with the solvent-induced adhesion at the fiber junction points during immersion can clamp together in the condensed fibrous membrane, which greatly enhances such hindrance and thus requires more stress to fail. A well structure formed in fibers does not allow the crack to propagate easily, which contributes to the stability of the obtained mechanical strength for B-CA ENMs after the solvent-soaking treatment. Given the lack of fusion for as-spun B-CA membranes, relatively poor mechanical properties are expected. Based on the mechanical strength results, the 95/5 (v/v) ethanol/acetone mixture solution used in the solvent-soaking method achieved relatively high strength and modulus for B-CA membrane, and was selected as the optimum treatment conditions. The treated B-CA at optimal treating conditions was used in the subsequent pure water permeability tests.

3.5. Hydraulic permeability

The pure water permeability of the untreated and treated B-CA ENMs is shown in Fig. 6. Such a hydraulic permeability was sufficient to allow the liquid separation, thereby showing potential as high-efficiency pre-filters for fibrous membranes. Under the same applied pressure, the pure water permeability of B-CA ENMs slightly decreased after solvent-soaking treatments. This finding might be due to the fact that the solvent-soaking treatment led to fiber cross-bonding and fiber deformation in the ENM, which decreased the porosity and the porosity interconnectivity (increasing tortuosity) by blocking pores. This blocking also reduced the pore size.

Fig. 6 Hydraulic permeability results.

In general, the solvent-soaking treatment did not greatly compromise the pure water permeability of B-CA ENMs while significantly improving the mechanical strength to make them suitable for water filtration applications. These effects would also improve prospects for handling and scale-up manufacturing. They presented additional benefits related to fouling resistance or as a primer layer for further modification and might also be used in applications of electrospun nanofibers beyond the field of membrane filtration. Furthermore, for aqueous filtration applications, a hydrophilic membrane surface usually favored the forecasted subsequent step of research.

4. Conclusion

In this study, B-CA was prepared via a typical acetylation process. By the use of a high-speed collection drum, B-CA was electrospun into aligned ultrafine fibrous membranes, and a solvent-soaking post-treatment was performed on the membranes to improve their mechanical properties. This study found that treating fibers via this solvent-soaking method resulted in the changes in morphology and molecular orientation degree, which could enhance the mechanical properties without significantly compromising the membrane hydraulic permeability. This method offers a practical and facile approach to improving the mechanical performances of other polymer nanofibrous membranes, which may be useable as self-supporting stand-alone materials without the need of an additional support for application in membrane filtration.

Chemical compounds studied in this article

Acetone (PubChem CID: 180).

Ethanol (PubChem CID: 702).

N,N-Dimethylacetamide (PubChem CID: 31374).

Acknowledgements

Funding support from the Science and Technology Research Project of Hubei Provincial Department of Education (No. Q20161701) is acknowledged.

References

1. T. Lin, G. S. Bhat and R. Hegde, J. Nanomater., 2014, 2014, 1–2.
2. G. Yazgan, A. M. Popa, R. M. Rossi, K. Maniura-Weber, J. Puigmartí-Luis, D. Crespy and G. Fortunato, Polymer, 2015, 66, 268–276.
3. A. Becker, Current Directions in Biomedical Engineering, 2015, 1, 104–107.
4. N. Wang, Y. Gao, Y. X. Wang, K. Liu, W. Lai, Y. Hu, Y. Zhao, S. L. Chou and L. Jiang, Adv. Sci., 2016, 1600013.
5. J. Choi, B. J. Yang, G. N. Bae and J. Jung, ACS Appl. Mater. Interfaces, 2015, 7, 25313–25320.
6. R. Seyedmahmoud, P. Mozetic, A. Rainer, S. M. Giannitelli, F. Basoli, M. Trombetta, E. Traversa, S. Licoccia and A. Rinaldi, J. Biomed. Mater. Res., Part A, 2015, 103, 103–114.
7. E. Kenawy, F. I. Abdel-Hay, M. H. El-Newehy and G. E. Wnek, Text. Res. J., 2008, 78, 45–52.
8. L. Huang, J. T. Arena, S. S. Manickam, X. Jiang, B. G. Willis and J. R. Mccutcheon, J. Membr. Sci., 2014, 460, 241–249.
9. R. Gopal, S. Kaur, C. Y. Feng, C. Chan, S. Ramakrishna, S. Tabe and T. Matsuura, J. Membr. Sci., 2007, 289, 210–219.
10. R. Gopal, S. Kaur, Z. W. Ma, C. Chan, S. Ramakrishna and T. Matsuura, J. Membr. Sci., 2006, 281, 581–586.
11. L. W. Huang, S. S. Manickam and J. R. McCutcheon, J. Membr. Sci., 2013, 436, 213–220.
12. K. Yoon, K. Kim, X. F. Wang, D. F. Fang, B. S. Hsiao and B. Chu, Polymer, 2006, 47, 2434–2441.
13. X. F. Wang, K. Zhang, M. F. Zhu, B. J. S. Hsiao and B. J. Chu, Macromol. Rapid Commun., 2008, 29, 826–831.
14. J. J. Mack, L. M. Viculis, A. Ali, R. Luoh, G. L. Yang, H. T. Hahn, F. K. Ko and R. B. Kaner, Adv. Mater., 2005, 17, 77–80.
15. G. M. Kim, R. Lach, G. H. Michler, P. Potschke and K. Albrecht, Nanotechnology, 2006, 17, 963–972.
16. R. Sen, B. Zhao, D. Perea, M. E. Itkis, H. Hu, J. Love, E. Bekyarova and R. C. Haddon, Nano Lett., 2004, 4, 459–464.
17. J. W. Lu, Z. P. Zhang, X. Z. Ren, Y. Z. Chen, J. Yu and Z. X. Guo, Macromolecules, 2008, 41, 3762–3764.
18. J. Y. Ji, G. Sui, Y. H. Yu, Y. X. Liu, Y. H. Lin, Z. J. Du, S. Ryu and X. P. Yang, J. Phys. Chem. C, 2009, 113, 4779–4785.
19. L. Q. Liu, D. Tasis, M. Prato and H. D. Wagner, Adv. Mater., 2007, 19, 1228–1233.
20. S. F. Fennessey and R. J. Farris, Polymer, 2004, 45, 4217–4225.
21. X. H. Zong, S. F. Ran, D. F. Fang, B. S. Hsiao and B. Chu, Polymer, 2003, 44, 4959–4967.
22. R. Jalili, M. Morshed, S. Abdolkarim and H. Ravandi, J. Appl. Polym. Sci., 2006, 101, 4350–4357.
23. W. Sun, Q. Cai, P. Li, X. L. Deng, Y. Wei, M. M. Xu and X. P. Yang, Dent. Mater., 2010, 26, 873–880.
24. K. Yoon, B. S. Hsiao and B. Chu, Polymer, 2009, 50, 2893–2899.
25. J. Cai, H. Niu, Y. Yu, H. Xiong and T. Lin, J. Text. Inst. DOI:10.1080/00405000.2016.1174456.
26. J. Cai, J. Chen, Q. Zhang, M. Lei, J. He, A. Xiao, C. Ma, S. Li and H. Xiong, Carbohydr. Polym., 2016, 140, 238–245.
27. O. Kousaku, Molecules, 2015, 20, 9139–9154.
28. R. Dashtbani and E. Afra, Int. J. Nano Dimens., 2015, 6, 1–9.
29. M. K. Joshi, A. P. Tiwari, B. Maharjan, K. S. Won, J. K. Han, H. P. Chan and C. S. Kim, Carbohydr. Polym., 2016, 147, 104–113.
30. Z. Zhou, W. Lin and X. F. Wu, Colloids Surf., A, 2016, 494, 21–29.
31. J. Cai, H. Niu, H. Wang, H. Shao, J. Fang, J. He, H. Xiong, C. Ma and T. Lin, J. Power Sources, 2016, 324, 302–308.
32. A. Rezaei, A. Nasirpour and M. Fathi, Compr. Rev. Food Sci. Food Saf., 2015, 14, 269–284.
33. S. C. Lakkad and J. M. Patel, Fibre Sci. Technol., 1981, 14, 319–322.
34. K. Y. Lee, L. Jeong, Y. O. Kang, S. J. Lee and W. H. Park, Adv. Drug Delivery Rev., 2009, 61, 1020–1032.
35. S. Tungprapa, T. Puangparn, M. Weerasombut, I. Jangchud, P. Fakum, S. Semongkhol, C. Meechaisue and P. Supaphol, Cellulose, 2007, 14, 563–575.
36. S. O. Han, W. K. Son, J. H. Youk, T. S. Lee and W. H. Park, Mater. Lett., 2005, 59, 2998–3001.
37. J. Y. Park, B. W. Han and I. H. Lee, J. Ind. Eng. Chem., 2007, 13, 1002–1008.
38. A. Frenot, M. W. Henriksson and P. Walkenstrom, J. Appl. Polym. Sci., 2007, 103, 1473–1482.
39. S. L. Zhao, X. H. Wu, L. G. Wang and Y. Huang, Cellulose, 2003, 10, 405–409.
40. J. Cai, P. Fei, Z. Y. Xiong, Y. J. Shi, K. Yan and H. G. Xiong, Carbohydr. Polym., 2013, 92, 11–18.
41. Y. W. Tang, C. Wong, H. X. Wang, A. Sutti, M. Kirkland, X. G. Wang and T. Lin, Tissue Eng., Part C, 2011, 17, 209–218.
42. J. S. Li and A. F. T. Mak, J. Biomater. Appl., 2005, 19, 253–266.
43. H. Q. Hou, J. J. Ge, J. Zeng, Q. Li, D. H. Reneker, A. Greiner and S. Z. D. Cheng, Chem. Mater., 2005, 17, 967–973.
44. J. Li and A. F. Mak, J. Biomater. Appl., 2005, 19, 253–266.
45. Z. Ma, M. Kotaki and S. Ramakrishna, J. Membr. Sci., 2006, 272, 179–187.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra11020g

---