Fabrication and characterization of a high performance polyimide ultrafiltration membrane for dye removal

Journal Pre-proofs Fabrication and characterization of a high performance polyimide ultrafiltration membrane for dye removal Chengyu Yang, Weixing Xu, Yang Nan, Yiguang Wang, Yunxia Hu, Congjie Gao, Xianhong Chen PII: DOI: Reference:

S0021-9797(19)31399-2 https://doi.org/10.1016/j.jcis.2019.11.075 YJCIS 25692

To appear in:

Journal of Colloid and Interface Science

Received Date: Revised Date: Accepted Date:

19 August 2019 12 November 2019 17 November 2019

Please cite this article as: C. Yang, W. Xu, Y. Nan, Y. Wang, Y. Hu, C. Gao, X. Chen, Fabrication and characterization of a high performance polyimide ultrafiltration membrane for dye removal, Journal of Colloid and Interface Science (2019), doi: https://doi.org/10.1016/j.jcis.2019.11.075

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Published by Elsevier Inc.

Fabrication and characterization of a high performance polyimide ultrafiltration membrane for dye removal Chengyu Yanga, Weixing Xub, Yang Nanb, Yiguang Wangc, *, Yunxia Hud, Congjie Gaoe, Xianhong Chene a Science

and Technology on Thermostructural Composite Materials Laboratory,

Northwestern Polytechnical University, Xi’an, Shaanxi 710072, P. R. China b

c Institute

Zhongfu Lianzhong Technology Co., Ltd 222006, P. R. China

of Advanced structure Technology, Beijing Institute of Technology, Haidian District Beijing, 100081, P. R. China

d State

Key Laboratory of Separation Membranes and Membrane Processes, School of

Materials Science and Engineering, Tianjin Polytechnic University, Tianjin 300387, PR China e Center

for Membrane Separation and Water Science & Technology, Ocean College, Zhejiang University of Technology, Hangzhou 310014, P. R. China

Abstract Membrane separation technology is one of the cost effective and most efficient technologies for treatment of wastewater from textile industry. However, development of membranes with better performance and thermal stability is still a highly challenging task. In this study, successful preparation of a novel thermally stable polyimide (PI) polymer was demonstrated using 2,4,6-trimethyl-1,3-phenylenediamine, 4,4′diaminodiphenylmethane and 1,2,4,5-benzenetetracarboxylic dianhydride components.

PI was selected as representative candidate because of its excellent thermal stability (decomposition temperature of 529 °C), as revealed by thermogravimetric analysis. Furthermore, PI polymer was used to fabricate ultrafiltration (UF) membrane by phase inversion process. This UF membrane is especially interesting as it allowed for almost complete penetration of monovalent (NaCl) and divalent (Na2SO4) inorganic salts because of its molecular weight cut off of 9320 Da. Moreover, the membrane exhibited very good surface hydrophilicity with the water contact angle of 67.6°. This PI-based UF membrane was found to be substantially effective as it showed high pure-water and dye-permeation fluxes of 345.10 and 305.58 L m−2 h−1 at 0.1 MPa, respectively. Besides, the membrane exhibited a rejection of 98.65% toward the direct red 23 dye (100 ppm) at 0.1 MPa. Thus, this PI-based UF membrane is highly beneficial and acts as a potential candidate for dye removal from wastewater produced by textile industry. Keywords: Synthesis of polyimide, ultrafiltration membranes, thermal stability, highflux, dye removal.

*Corresponding author. E-mail address: [email protected](YG Wang).

1. Introduction The wastewater generated from textile industry consists of various complex chemical substances including dyes, detergents, grease, oil, sulfates, solvents, heavy metals and other inorganic salts depending on the process regime [1]. Textile wastewater treatment is extremely challenging due to its high temperature, high pH, high chemical oxygen demand and low biodegradability. Moreover, it also contains high concentrations of inorganic salts and azo dyes [2], which pose serious hazards to human health and environment. Till date, several methods such as biological treatments, electrochemical oxidation, adsorption, dye degradation and membrane filtration processes have been developed for dye removal [3–7]. Nonetheless, most of the above mentioned approaches suffer from difficult recovery of the valuable residual dyes and salts, which is necessary for a truly sustainable operation. Among these technologies, membrane separation has been extensively used owing to its cost effectiveness, high efficiency and simplicity [8], long-term chemical stability and high mechanical strength. In addition to some commercial materials such as NF-70 and UTC-60 nanofiltration (NF) membranes, novel lab-scale NF membranes have also been developed [9,10] and utilized for wastewater treatment. Liu et al. fabricated an NF membrane with positive charges by an ultraviolet (UV)-induced photografting polymerization method. This membrane showed a 99.4% rejection toward congo red (CR) and 90% permeation of Na2SO4 [11]. However, these NF membranes required high operating pressures, leading to significant increase in energy consumption. In order to overcome these issues, ultrafiltration (UF) membranes have been fabricated and tested for treating textile

wastewaters owing to their capability to provide complete salt passage. Chen et al. fabricated various charged regenerated cellulose UF membranes [12]. This membrane showed a 98.5% rejection of reactive red ED-2B under low pressures condition. Nonetheless, its dye rejections were reduced at higher operating pressures and ionic strengths. Besides, the flux of currently employed membranes is extremely low, which leads to higher operating costs thus making them practically inappropriate. Therefore, development of efficient membrane technology for textile wastewater treatment is urgently required. Comparative studies indicate that UF membranes provide good salt permeation and also maintain high fluxes resulting from their low osmotic pressure, leading to high separation efficiencies. Till date, a large number of polymeric materials including poly(vinylidene fluoride) (PVDF) [13], polyacrylonitrile (PAN) [14], cellulose acetate (CA) [15], polysulfone (PSf) [16,17] and polyethersulfone (PES) [18,19] have been used to prepare UF membranes. However, it is well known that operating temperature affects the performance of the membrane. In general, with the increase in the temperature, the flux of the membrane increases but the rejection rate decreases. Noteworthy, the actual operating temperature (as high as 90 °C) of the wastewater often exceeds the stable operating temperature of most polymeric membranes [20]. This leads to the improper working of the membranes. The properties of the membrane are closely related to the material used for its fabrication. Thus, development of thermally stable materials for fabricating UF membranes is highly desirable. In this study, a novel polyimide (PI) polymer was successfully synthesized from

2,4,6-trimethyl-1,3-phenylenediamine

(TMPDA),

4,4′-diaminodiphenylmethane

(DDM) and 1,2,4,5-benzenetetracarboxylic dianhydride (PMDA) in N-methyl-2pyrrolidone (NMP). The as-synthesized PI polymer showed a decomposition temperature of 529 °C as revealed by the thermogravimetric analysis (TGA), indicating that this composite material is well suited for fabricating thermally stable membranes. Then the Pi-based UF membrane was fabricated by the phase separation method. Furthermore, the performance of the PI membrane toward the filtration of negatively charged dyes was systematically investigated. Finally, the effects of the presence of salts and temperature on the filtration performance were also comprehensively investigated. 2. Experimental 2.1. Synthesis of polyimide Schematic illustration of the synthesis process of PI is shown in Fig. 1. First, TMPDA (0.075 g, 5 mmol), DDM (1.983 g, 10 mmol) and NMP (50 mL) were taken in a three-neck flask (100 mL), which was pre-washed and dried. After stirring at room temperature for 2 h, PMDA (3.272 g, 15 mmol) was added and the contents were stirred at 25 °C for ca. 10 h. Finally, pyridine and acetic anhydride were added and the mixture was stirred for 2 h at 80 °C. The entire experiment was carried out under nitrogen flow to maintain inert atmosphere. After methanol titration, yellow pulverized PI polymer precipitated out as the final product. 2.2. Membrane preparation The preparation process of the membrane is as follows: First, the PI was dissolved

in NMP at 16 wt.% and the mixture was stirred until the solution was homogeneous. After degassing by allowing the mixture to stand for 12 h, the casting solution was coated on the polyester (PET, 150 μm) membranes and rapidly immersed in deionized water for at least 24 h before the tests. 2.3. Characterization Fourier transform infrared (FTIR, TENSOR 27, Bruker, Germany) spectroscopy and proton nuclear magnetic resonance (1H NMR, Bruker Avance, 400 MHz) spectroscopy (deuterated chloroform as solvent) were employed to characterize the structure of the PI polymer. TGA (STA449 F5, NETZSCH, Germany) was conducted on the PI polymer under nitrogen flow at a heating rate of 10 °C min−1. Gel permeation chromatography (GPC) measurements (Shimadzu 10A apparatus) were carried out to analyze the molecular weight of the polymer (DMF was used as an eluent). Scanning electron microscopy (SEM, SU8010, Hitachi, Japan), atomic force microscopy (AFM, Multimode 8, Bruker, USA, tapping mode), and Zeta potential analysis (SurPASS 3, Anton Paar, Austria) were utilized for analyzing the microstructure and surface charge of the PI-based UF membrane. 2.4. Filtration experiments In the experiments, a total of four dyes was used: CR (696.08 g mol−1) Coomassie brilliant blue (BBR, 840.27 g mol−1), Direct red 23 (DR23, 813.72 g mol−1) and Evans blue (EB, 959.98 g mol−1). All experiments were based on three samples (the effective membrane area was 19.6 cm−2) and all results were averaged. Before the experiments, all the membranes were pre-compacted for 1 h (0.1 MPa). The separation performance

of membrane toward different dyes was tested on dye solutions (100 ppm) with different concentrations of salts (5, 10, 20, 30 and 40 g L−1) or under different pH (1, 3, 5, 7, 9, 11 and 13), respectively. The fluxes (Js) of the aqueous dye solution are calculated as follows: 

   

(1)

where V is the permeation volume (L), A represents the effective membrane area (m2) and t denotes the permeation time (h). The dyes rejection (Rd) is calculated by using the following equation:







       

(2)

where Aps is the dye absorbance of the permeate and Afs represents the dye absorbance of feed solutions. UV–visible (UV–vis) spectrophotometer was used to measure the dye concentration by measuring the absorbance at peak maximum for each dye. 2.5. Molecular weight cut off and pore size distribution Molecular weight cut off (MWCO) was measured by polyethylene glycol (PEG, MW of 2000, 4000, 6000, 8000 and 10000 Da, analytical purity, Sigma-Aldrich) tests and analyzed using a total organic carbon (TOC) analyzer [21]. The calculation of PEG rejection rate could still be applied to Eq. (2). When the PEG rejection rate was 90%, the MWCO was equal to the molecular weight of PEG. According to MWCO, the Stokes radius can be calculated as follows [22]:         

(3)

The mean effective pore size (μp) is defined as the Stokes radius (d) of the PEG

having a 50% rejection rate. Moreover, the mathematical fitting analysis of the

relationship between them is as follows [23,24]: ! "# "#

"



#$% &#



()#  '

$% "#  $% *#  $% &#



(4)

where rp is the pore radius and σp denotes the geometric standard deviation, which is the ratio of rp at 84.13% rejection rate to over that at 50%. 3. Results and discussion 3.1. Characterization of polyimide 3.1.1. Structural characterization Fig. 2 shows the FTIR spectrum of the PI polymer, exhibiting that the main characteristic absorption bands of the imide ring appear in the following positions: 1779 cm−1 (asymmetric stretching of the carbonyl group), 1727 cm−1 (asymmetric stretching of the carbonyl group) and 1359 cm−1 (C–N bond). Fig. 3 shows the 1H NMR spectrum of the PI polymer, exhibiting the assignment of all the signals. Moreover, the nuclear magnetic analysis software (MestReNova) was used to calculate the integration of peaks from a to g as follows: 24.22, 12.13, 3.97, 24.10, 33.09, 31.79 and 15.87 (in the ratio 6:3:1:6:8:8:4), which confirms the expected copolymer structure [25,26]. 3.1.2. Thermal analysis and molecular weight Fig. 4 shows the differential scanning calorimetry (DSC) and TGA curves of PI polymer. The results indicate that PI is a material with relatively high thermal stability. It shows glass transition and initial decomposition temperatures of 258 and 529 °C, respectively. Table 1 presents the comparative analysis of the thermal stability of PIbased UF membrane with those of some reported membranes. The pores of PI-based UF membrane do not collapse or expand with temperature due to the outstanding

thermal stability of PI polymer. This is attributed to the fact that the rigidity of the polymer chain hinders longitudinal sliding displacements between inter-chains. Notably, the average molecular weight is used to describe the molecular weight size of the polymer. The average molecular weight can be divided into number average molecular weight (Mn), weight average molecular weight (Mw) and viscosity average molecular weight (Mv). The Mn of the PI polymer is 68035 Daltons and the Mw is 151274 Daltons. The ratio of Mw to Mn is denoted as polydispersity index (PDI). The smaller the PDI, the more uniform the molecular weight distribution. The PDI of the PI polymer is 2.22, which may be due to the presence of small amount molecular impurities that make the reaction environment impure. In condensation polymerization, the removal of small molecular products is undeniably necessary, because the impurity widens the molecular weight distribution of the products. 3.2. Characterization of the membrane 3.2.1. Morphological characterization Figs. 5a and 5b demonstrate that the membrane surface is relatively dense. Fig. 5c shows cross-section morphology, exhibiting typical asymmetric structure of the membrane, which is composed of a dense surface layer and a completely developed finger-like macroporous structure. This structure also indicates that the membrane may have high flux [30]. Fig. 5d shows three-dimensional (3D) AFM image, which reveals a ridge-and-valley structure (3.55 nm (Ra) and 4.58 nm (Rq)). 3.2.2. Analysis of surface performance Surface charge and hydrophilicity of the membrane can aid in determination of its

permeability, rejection and antifouling performance [31,32]. Obviously, the PI-based membrane is electronegative at the pH range of 3.0–10.0 (Fig. 6a). Thus, the membrane shows higher retention rates for electronegative molecules. Electrostatic attraction leads to the aggregation of electropositive molecules on the surface of the membrane, thus resulting in membrane fouling. Fig. 6b exhibits that the contact angle is 67.6°, revealing an inherent hydrophilicity of the material. 3.3. Molecular weight cut off, permeation and filtration performance Fig. 7 shows that the MWCO of the PI-based UF membrane is 9320 Da. According to the mathematical fitting analysis, this membrane shows a mean effective pore radius of 1.785 nm. Moreover, Fig. 8 exhibits variation of the membrane permeance under different pressures. Clearly, the pure water flux increases in direct proportion to the operating pressure. Fig. 9 shows that the rejection is above 95.0% for all dyes tested in this study, while the pure water flux (WF) is 345.1 L m−2 h−1 at 0.1 MPa. The order of change of dye flux (DF) is as follows: FDR23 ≈ FCR > FBBR > FEB which can be explained in terms of the relationship between dye charges and osmotic pressure. The osmotic pressure produced by dye solution increases with the increase of dye charges so that the flux decreases. However, BBR (one negative charge) does not fit this pattern because it has higher molecular weight compared to other dyes. Thus, the electrostatic interaction among the dye molecules of this compound and the dye molecules and the membrane surface becomes low. As a result, dyes tend to aggregate and get adsorbed on the membrane surface, resulting in higher membrane pollution and lower permeate flux

[33]. Table 2 lists and compares the properties of some dye separation membranes reported in literature. The PI-based membrane prepared herein showed excellent dye rejection, salt permeation and very high flux. This may be due to the upright macroporous structure at the bottom of the membrane, which makes its flux quite high. Results of dye retention indicate that the membrane skin layer has a strong negative charge and the electrostatic interaction plays a dominant role, thus the rejection rate for negative charge dye is fairly high. Table 2 presents that the flux of the PI membrane was found to be dozens of times higher than those of other membranes reported previously. The result indicates that this material has potential for decolorizing wastewater or to fractionate dye/salt mixtures for effective dye removal. 3.4. Effect of salt concentration on membrane performance Fig. 10 demonstrates that the permeate flux decreases with the increase in the salt concentration. This can be attributed to the fact that the osmotic pressure between the permeate streams the retentate. With the increase in the salt concentration, the osmotic pressure increases and the resistance of the solution through the membrane pores also increases, thus resulting in decrease of membrane flux [37]. On the other hand, the figure exhibits that the rejection of BBR first increases and then becomes stable at higher salt concentrations. This result can be attributed to the aggregation of the anionic dyes affected by metal ions because counterions offset the Coulombic repulsive forces [38,39]. 3.5. Effect of operating temperature on membrane performance

Fig. 11 exhibits the effect of the operating temperature on the membrane flux and rejection of CR. With the increase in the operating temperature from 20 to 95 °C at 0.1 MPa, the dye permeate flux increases from 305.58 to 539.85 L m−2 h−1; however, the rejection decreases from 99.09 to 91.82%. Compared to that at 20 °C, the dye permeate flux increases by 76.66%. However, the rejection rate decreases by 7.27% below 95 °C. With the increase in the operating temperature, the dye permeate flux increases due to higher mobility of water molecules and lower viscosity of the dye solution [40,41]. Correspondingly, the rejection is also reduced [42]. Stability of the rejection confirms the excellent thermal stability of PI. Thus the membrane performance in the operating temperature range could meet the requirements (normally 20 to 90 °C) in the textile dyeing industry. 3.6. Effect of the solution pH on membrane performance Fig. 12 shows low dye permeation fluxes and high rejection under acidic conditions, which can be attributed to the sulfonated acid groups of CR. These groups can lead to the decrease in the polarity and solubility of CR due to protonation. Therefore, dyes aggregate and precipitate, causing membrane fouling. Eventually, this phenomenon leads to the decrease in flux and the increase in retention, finally resulting in increase of the rejection rate. Besides, the dye permeation flux increases and the rejection decreases slightly with the increase in pH to 11, and then it decreases significantly when the pH is increased to 13. The results indicated that the imide ring was easily broken down by hydroxyl ions at high temperatures of the solution and in the presence of alkali catalysts [43]. At pH ≤ 11, the imide ring was more stable as it

was only slightly hydrolyzed, showing a slight decrease in the rejection rate. However, the membrane degraded and its structure collapsed at pH = 13, leading to very high flux and low retention. 4. Conclusions In this study, a new polyimide (PI) polymer material was synthesized using TMPDA, DDM and PMDA. Further, ultrafiltration (UF) membrane was prepared by phase inversion process. Compared to other reported membranes [11, 27–29], the hydrophilicity and surface smoothness of this PI-based UF membrane were improved. MWCO and surface Zeta potential test showed that the membrane is suitable for the removal of negatively charged dyes. The rejection rate was higher than 95.0% for all dyes tested in this study, while the dye permeate flux was 270.5 L m−2 h−1 (0.1 MPa, 100 ppm). Moreover, owing to the excellent thermal and chemical stability, the PIbased UF membrane is thermally stable under extreme operating conditions including high temperature (95 +) or wide pH range (1–11). The results indicate that due to the

electrostatic interaction at the membrane interface, the developed new PI-based UF membrane is suitable for decolorizing wastewater or fractionating dye/salt mixture to effectively remove dyes. Thus it acts as a potential candidate for a wide range of engineering and environmental applications. This study also provides new insight into the design and development of novel UF membranes for dye removal.

References [1] A.N. Kulkarni, A.D. Watharkar, N.R. Rane, et al. Decolorization and detoxification of dye mixture and textile effluent by lichen, Dermatocarpon vellereceum, in fixed bed upflow bioreactor with subsequent oxidative stress study[J]. Ecotoxicology and Environmental Safety, 2018, 148:1725. [2] A.K. Verma, R.R. Dash, P. Bhunia, A review on chemical coagulation/flocculation technologies for removal of colour from textile wastewaters, J. Environ. Manag. 93 (2012) 154–168. [3] J. Dasgupta, J. Sikder, S. Chakraborty, S. Curcio, E. Drioli, Remediation of textile effluents by membrane based treatment techniques: a state of the art review, J. Environ. Manag. 147 (2015) 55– 72. [4] H. E. Emam, H. N. Abdelhamid, R. M. Abdelhameed, Self-cleaned photoluminescent viscose fabric incorporated lanthanideorganic framework (Ln-MOF), Dyes and Pigments 2018, 159, 491498. [5] H. N. Abdelhamid, Z. H. Huang, A. M. El-Zohry, H. Q. Zheng, X. D. Zou, A Fast and Scalable Approach for Synthesis of Hierarchical Porous Zeolitic Imidazolate Frameworks and One-Pot Encapsulation of Target Molecules, Inorganic chemistry 2017, 56 (15), 9139-9146. [6] A. S. Etman, H. N. Abdelhamid, Y. Y. Yuan, L.G. Wang, X.D. Zou, J. L. Sun, Facile WaterBased Strategy for Synthesizing MoO3−x Nanosheets: Efficient Visible Light Photocatalysts for Dye Degradation, ACS omega 2018, 3 (2), 2193-2201. [7] H. N. Abdelhamid, X. D. Zou, Template-free and room temperature synthesis of hierarchical porous zeolitic imidazolate framework nanoparticles and their dye and CO2 sorption, Green chemistry 2018, 20 (5), 1074-1084.

[8] R.W. Baker, Membrane Technology and Applications, Second ed., John Wiley & Sons Ltd, England, 2004. [9] B. Van der Bruggen, B. Daems, D. Wilms, C. Vandecasteele, Mechanisms of retention and flux decline for the nanofiltration of dye baths from the textile industry, Sep. Purif. Technol. 22 (2001) 519–528. [10] L. Meihong, Y. Sanchuan, Z. Yong, G. Congjie, Study on the thin-film composite nanofiltration membrane for the removal of sulfate from concentrated salt aqueous: preparation and performance, J. Membr. Sci. 310 (2008) 289–295. [11] F. Liu, B.-R. Ma, D. Zhou, L.-J. Zhu, Y.-Y. Fu, L.-X. Xue, Positively charged loose nanofiltration membrane grafted by diallyl dimethyl ammonium chloride (DADMAC) via UV for salt and dye removal, React. Funct. Polym. 86 (2015) 191–198. [12] X. Chen, Y. Zhao, J. Moutinho, J. Shao, A.L. Zydney, Y. He, Recovery of small dye molecules from aqueous solutions using charged ultrafiltration membranes, J. Hazard. Mater. 284 (2015) 58– 64. [13] F. Liu, N.A. Hashim, Y. Liu, M.R.M. Abed, K. Li, Progress in the production and modification of PVDF membranes, J. Membr. Sci. 375 (2011) 1–27. [14] W. Li, Z. Yang, Q. Meng, C. Shen, G. Zhang, Thermally stable and solvent resistant selfcrosslinked TiO2/PAN hybrid hollow fiber membrane fabricated by mutual supporting method, J. Membr. Sci. 467 (2014) 253–261. [15] A. Idris, L.K. Yet, The effect of different molecular weight PEG additives on cellulose acetate asymmetric dialysis membrane performance, J. Membr. Sci. 280 (2006) 920–927. [16] D.Y. Koseoglu-Imer, B. Kose, M. Altinbas, I. Koyuncu, The production of polysulfone (PS)

membrane with silver nanoparticles (AgNP): physical properties, filtration performances, and biofouling resistances of membranes, J. Membr. Sci. 428 (2013) 620–628. [17] R. Kumar, A.M. Isloor, A.F. Ismail, T. Matsuura, Performance improvement of polysulfone ultrafiltration membrane using N-succinyl chitosan as additive, Desalination 318 (2013) 1–8. [18] H. Rajabi, N. Ghaemi, S.S. Madaeni, P. Daraei, B. Astinchap, S. Zinadini, S.H. Razavizadeh, Nano-ZnO embedded mixed matrix polyethersulfone (PES) membrane: influence of nanofiller shape on characterization and fouling resistance, Appl. Surf. Sci. 349 (2015) 66–77. [19] G. Ouyang, A. Hussain, J. Li, D. Li, Remarkable permeability enhancement of polyethersulfone (PES) ultrafiltration membrane by blending cobalt oxide/graphene oxide nanocomposites, RSC Adv. 5 (2015) 70448–70460. [20] S. Barredo-Damas, M.I. Alcaina-Miranda, A. Bes-Piá, M.I. Iborra-Clar, A. IborraClar, J.A. Mendoza-Roca, Ceramic membrane behavior in textile wastewater ultrafiltration, Desalination 250 (2010) 623–628. [21] Y.P. Tang, S. Yuwen, T.S. Chung, M. Weber, C. Staudt, C. Maletzko, Synthesis of hyperbranched polymers towards efficient boron reclamation via a hybrid ultrafiltration process, J. Membr. Sci. 510 (2016) 112–121. [22] J. Lin, C.Y. Tang, C. Huang, Y.P. Tang, W. Ye, J. Li, J. Shen, R. Van den Broeck, J. Van Impe, A. Volodin, A comprehensive physico-chemical characterization of superhydrophilic loose nanofiltration membranes, J. Membr. Sci. 501 (2016) 1–14. [23] T.Y. Liu, L.X. Bian, H.G. Yuan, B. Pang, Y.K. Lin, Y. Tong, B. Van der Bruggen, X.L. Wang, Fabrication of a high-flux thin film composite hollow fiber nanofiltration membrane for wastewater treatment, J. Membr. Sci. 478 (2015) 25–36.

[24] S. Atchariyawut, C. Feng, R. Wang, R. Jiraratananon, D. Liang, Effect of membrane structure on mass-transfer in the membrane gas–liquid contacting process using microporous PVDF hollow fibers, J. Membr. Sci. 285 (2006) 272–281. [25] S. Park, K. Kim, J.C. Kim, W. Kwon, D.M. Kim, M. Ree, Synthesis and nonvolatile memory characteristics of thermally, dimensionally and chemically stable polyimides, Polymer 52 (2011) 2170–2179. [26] K.L. Mittal. Polyimides and other high temperature polymers: synthesis, characterization, vol. 2. Utrecht: VSP; 2003. [27] G. A. Polotskaya, T. K. Meleshko, I. V. Gofman, A. E. Polotsky, A. N. Cherkasov, Polyimide ultrafiltration membranes with high thermal stability and chemical durability, Separation Science and Technology 44.16 (2009): 3814-3831. [28] B. Pulido, S. Chisca, S. P. Nunes, Solvent and thermal resistant ultrafiltration membranes from alkynefunctionalized high-performance polymers, J Membr Sci 564 (2018) 361–371. [29] S. Xu, Y. Wang, Novel thermally cross-linked polyimide membranes for ethanol dehydration via pervaporation, J Membr Sci 496 (2015) 142-155/ [30] B. Chakrabarty, A.K. Ghoshal, M.K. Purkait. Effect of molecular weight of PEG on membrane morphology and transport properties. J Membr Sci 309 (2008) 209–221. [31] Y. Liu, Y.P. Ai, P.F. Huo, An effective approach to fabrication of antifouling ultrafiltration membrane based on zwitterionic polyimide, High Performance Polymers 2017, Vol. 29(9) 1006– 1015. [32] P.S. Liu, Q. Chen, S.S Wu, et al. Surface modification of cellulose membranes with zwitterionic polymers for resistance to protein adsorption and platelet adhesion. J Membr Sci 2010; 350: 387–

394. [33] J.Y. Lin, W.Y. Ye, Marian-Cornel Baltaru, Y. P. Tang, et al. Tight ultrafiltration membranes for enhanced separation of dyes and Na2SO4 during textile wastewater treatment, J Membr Sci 514 (2016) 217–228. [34] J. Zhu, M. Tian, J. Hou, J. Wang, J. Lin, Y. Zhang, J. Liu, B. Van der Bruggen, Surface zwitterionic functionalized graphene oxide for a novel loose nanofiltration membrane, J. Mater. Chem. A 4 (2016) 1980–1990. [35] J. Lin, W. Ye, H. Zeng, H. Yang, J. Shen, S. Darvishmanesh, P. Luis, A. Sotto, B. Van der Bruggen, Fractionation of direct dyes and salts in aqueous solution using loose nanofiltration membranes, J. Membr. Sci. 477 (2015) 183–193. [36] C.L. Liu, H.C Mao, L.F. Zheng, S.B. Zhang, Tight ultrafiltration membrane: Preparation and characterization of thermally resistant carboxylated cardo poly (arylene ether ketone)s (PAEKCOOH) tight ultrafiltration membrane for dye removal, J. Membr Sci 530 (2017) 1–10. [37] M. Nilsson, G. Trägårdh, K. Östergren, The influence of pH, salt and temperature on nanofiltration performance, J. Membr. Sci. 312 (2008) 97–106. [38] K.W. Yeung, S.M. Shang, The influence of metal ions on the aggregation and hydrophobicity of dyes in solutions, Color. Technol. 115 (1999) 228–232. [39] S.M. Mooi, S.N. Keller, B. Heyne, Forcing aggregation of cyanine dyes with salts: a fine line between dimers and higher ordered aggregates, Langmuir: ACS J. Surf. Colloids 30 (2014) 9654– 9662. H. Lim, W.J. Cho, C.S. Ha, S. Ando, Y.K. Kim, C.H. Park, K. Lee, Adv. Mater. 14 (2002) 1275 [40] J. Dasgupta, S. Chakraborty, J. Sikder, R. Kumar, D. Pal, S. Curcio, E. Drioli, The effects of

thermally stable titanium silicon oxide nanoparticles on structure and performance of cellulose acetate ultrafiltration membranes, Sep. Purif. Technol. 133 (2014) 55–68. [41] C. Wu, S. Zhang, C. Liu, D. Yang, X. Jian, Preparation, characterization and performance of thermal stable poly(phthalazinone ether amide) UF membranes, J. Membr. Sci. 311 (2008) 360– 370. [42] A. Mehta, A.L. Zydney, Permeability and selectivity analysis for ultrafiltration membranes, J. Membr. Sci. 249 (2005) 245–249. [43] D.J. Liaw, K.L. Wang, Y.C. Huang, K.R. Lee, J.Y. Lai, C.S. Ha, Prog, Advanced polyimide materials: Syntheses, physical properties and applications, Polymer. Sci. 37 (2012) 907.

Appendices Fig. 1. Synthesis route of the PI polymer. Fig. 2. FTIR spectrum of the PI polymer. Fig. 3. 1H NMR spectrum of the PI polymer. Fig. 4. DSC and TGA curves of the PI polymer. Fig. 5. Characterization of the membrane morphology: (a) and (b) surface morphology, (c) cross-section morphology and (d) 3D AFM image. Fig. 6. Analysis of surface performance of the membrane. Fig. 7. (a) MWCO and (b) pore size distribution of the membrane. Fig. 8. Effect of the operation pressure on the pure water flux. Fig. 9. Membrane permeation of pure water and rejection of dye solutions. (Dye concentration: 0.1 g L−1) Fig. 10. Effect of salt concentration on membrane performance. (Dye: BBR, operation pressure: 0.1 MPa, background salt: Na2SO4). Fig. 11. Effect of the operating temperature on membrane performance. (Dye: CR, operation pressure: 0.1 MPa) Fig. 12. Effect of the solution pH on membrane performance. (Dye: CR, operation pressure: 0.1 MPa, pH range: 1–13, and the pH was adjusted with HCl and NaOH and stabilized before the tests) Table 1 Comparison of thermal stability of PI-based UF membrane with those of some reported membranes. Table 2 Comparison of the performance between PI-based UF membrane and some

reported membranes.

Table 1 Comparison of thermal stability of PI-based UF membrane with those of some reported membranes. Material poly[(4,4’oxydiphenylene)pyromel liteimide] POX-N-PR PTA-O-PR PBI-N-PR FMD BMD

Initial decomposition temperature (°C) 517

PI

Reference [27]

497 497 443 450 470

[28] [28] [28] [29] [29]

529

This study

Table 2 Comparison of the performance between PI-based UF membrane and some reported membranes. Membrane

PWF L m−2 h−1

GOPSBMA/PES NF UH004 Sepro NF 2A Sepro NF 6 PAEKCOOH UF PI UF

12.5

DF Dye −2 −1 rejection Lm h (%) 8.8 95

pH range

Max. temp. (°C)

Ref

-

-

[31]

27.5 10.5 13.7 29.5

27 9.6 13.2 25.0

98.9 99.98 99.95 95

0–14.0 3.0–10.0 3.0–10.0 1.0–10.0

95 50 50 95

[11] [32] [32] [33]

345.1

270.5

95

1.0–11.0

95

This study

Operation pressure: 0.1 MPa.

Declaration of interests  The authors declare that they have no known competing financial interests or personal  relationships that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Chengyu Yang Validation, Formal analysis, Investigation, Data Curation, Writing - Original Draft, Writing - Review & Editing Weixing Xu: Conceptualization, Methodology Yang Nan: Conceptualization, Methodology Yiguang Wang: Resources, Supervision, Project administration, Funding acquisition Yunxia Hu: Writing - Review & Editing Congjie Gao: Conceptualization, Methodology, Funding acquisition Xianhong Chen: Resources, Writing - Review & Editing, Funding acquisition