Oct 28, 2014 - In order to achieve a higher permeation flux, PBI single-layer hollow fiber membranes of thinner selective layers were developed and studied ...
Article pubs.acs.org/IECR
Polybenzimidazole (PBI) Membranes for Phenol Dehydration via Pervaporation Yan Wang,*,† Michael Gruender,‡ and Sheng Xu† †
Key Laboratory for Large-Format Battery Materials and System, Ministry of Education, School of Chemistry and Chemical Engineering, Huazhong University of Science and Technology, 430074, Wuhan, China ‡ PBI Performance Products, Inc., 9800-D Southern Pine Boulevard, Charlotte, North Carolina 28273, United States S Supporting Information *
ABSTRACT: Phenol is an important commodity in chemical industries and dewatering is a critical process in its application. In this work, polybenzimidazole (PBI) membranes with various morphologies are employed for phenol dehydration via pervaporation, including flat-sheet dense membranes, single-layer and dual-layer hollow fiber membranes. Effects of cross-linking modification and post-thermal treatment on the performance of PBI flat-sheet dense membranes were investigated; effects of the operation temperature and feed composition are also studied, not only in terms of flux and separation factor, but also of the intrinsic permeance and selectivity of the membrane. In order to achieve a higher permeation flux, PBI single-layer hollow fiber membranes of thinner selective layers were developed and studied with the effect of different spinning parameters. The preliminary study of dual-layer PBI/PBI and PBI/poly(vinylidene fluoride) (PVDF) dual-layer hollow fiber membranes are also carried out to explore the potential of high-performance composite membranes for phenol dehydration. The promising separation performance of PBI membranes exhibited via benchmarking shows its great potential for phenol dehydration, and it may open new perspectives for the development of high-performance membranes for the pervaporation dehydration of phenol or other corrosive organics.
■
INTRODUCTION Phenol is an important raw material in industry for the synthesis of phenolic resin, bisphenol A, caprolactum, alkyl phenols, and many other useful materials and compounds. The global demand of phenol has been steadily increasing over the last 10 years: it stood at 7 934 218 tons in 2010, and is expected to reach 11 576 620 tons by 2020.1 During its usage and production process, a large quantity of wastewater containing toxic phenol is generated and causes severe water pollution. Therefore, an effective disposal technique to recover phenol from the wastewater is highly desired to avoid its toxic effect on the environment and living organisms. Typical physicochemical properties of phenol and water are listed in Table S1 of the Supporting Information.2,3 The waterphenol system can form miscible phase if the phenol composition is approximately 70 wt % at somewhat higher than room temperature. The formation of the azeotropic mixture and high volatile property of phenol (despite its high boiling point of 181.7 °C) makes it difficult to separate the phenol−water system by normal distillation. Alternatively, membrane-based pervaporation technique is therefore gaining much popularity, because of its low energy consumption, low capital cost, and, most importantly, its high separation efficiency in breaking azeotrope, since the separation is mainly based on the differences of the components in the sorption and diffusion abilities across the membrane. In industry, the wastewater containing phenol can be separated in two stages, as shown in the schematic diagram presented in Figure 1.4 In the first stage, the phenol solution is concentrated up to 80 wt % by the organic-selective process, and then the concentrated solution is handled by the water© 2014 American Chemical Society
Figure 1. Schematic diagram of phenol separation in industry (reproduced with permission from ref 4).
selective process in the second stage. Available literature shows that most studies on pervaporation separation of phenol/water mixture focus on the phenol recovery5−10 with a feed solution of low phenol content, using polyurethane (PU), poly(ether block amide) (PEBA 2533), polydimethylsiloxane (PDMS) membranes, etc. Relatively, very few pervaporation studies on phenol dehydration4,11 (second stage) were reported where the phenol concentration is generally >70 wt %. The toxic and Received: Revised: Accepted: Published: 18291
July 9, 2014 October 8, 2014 October 28, 2014 October 28, 2014 dx.doi.org/10.1021/ie502626s | Ind. Eng. Chem. Res. 2014, 53, 18291−18303
Industrial & Engineering Chemistry Research
Article
was prepared by diluting the original supplied PBI solution, and used for the fabrication of flat-sheet PBI dense membranes. The membrane was prepared by casting the dope solution onto a glass plate with a casting knife 100 μm thick, and then being placed together with the glass plate on a hot plate preset at 75 °C for 15 h, to allow the solvent DMAc to evaporate slowly. The resultant membrane then was allowed to cool to room temperature and removed from the glass plate by being immersed in water. After the membrane separated from the glass plate automatically, the wet membrane was dabbed dry with a filter paper and dried in a vacuum oven between two wire meshes with the temperature gradually increasing to 250 °C at a rate of 0.6 °C/min and holding at that temperature for 24 h to remove the residual solvents before cooling naturally. The wire meshes helped to dry the membrane from both surfaces uniformly. A Mitutoyo micrometer was used to measure the membrane final thickness, which is ∼20 ± 2 μm. Cross-Linking Modification and Post-Thermal Treatment of Flat-Sheet PBI Membranes. The cross-linking modification of the as-fabricated PBI membrane was carried out by immersing it in 1 wt % p-xylene dichloride solutions of different solvents (methanol, ethanol, and heptane) for a fixed duration at room temperature. The modified PBI membrane then was rinsed with corresponding fresh solvent and blotted dry with a filter paper. Thermal treatment of PBI membranes was carried out after the cross-linking and before the module fabrication, if applied. All thermal treatments were conducted for 2 h in a programmable oven preset at a fixed temperature, followed by naturally cooling under vacuum. Spinning Process and Module Fabrication of Hollow Fiber Membranes. The schematic diagram of single-layer and dual-layer hollow fiber spinning systems has been described elsewhere.27 For single-layer PBI hollow fibers, 24.6/74/1.4 wt % PBI/DMAC/LiCl solution was used as the dope solution, and a mixture of 70/30 (w/w) DMAc/water was employed as the bore fluid in order to make a porous inner surface. The polymer solution was degassed for 24 h before being loaded into a syringe pump (ISCO 1000). Hollow fibers were spun by extrusion of the polymer solution and bore fluid out of the spinneret orifice and subsequent phase inversion in a coagulant bath with a preset air gap. Both the dope fluid and bore fluid were filtered through 15 μm sintered metal filters before spinning. Tap water was used as the external coagulant at room temperature. The nascent fibers were rolled up by a drum, cut into segments, and then rinsed in a clean tapping water bath for at least 3 days with water changed daily to remove the remaining DMAc. The as-spun hollow fibers were freeze-dried, and then stored in an ambient environment. The detailed spinning parameters of the single-layer coextrusion process with various bore fluid flow rates and take-up speeds are given in Table S2 in the Supporting Information. For the spinning of dual-layer hollow fibers, 24.6 wt % PBI dope solution was used for the fabrication of the selective outer-layer, while 21 wt % PBI solution and 17 wt % PVDF solution were used for the fabrication of supporting layers, respectively. Detailed spinning parameters of PBI/PBI and PBI/PVDF dual-layer coextrusion process are listed in Table S3 in the Supporting Information. As-spun fibers were also immersed in tapping water for at least 3 days, with the water changed daily. After that, they were solvent-exchanged by methanol and hexane for three times each (each time for 30
corrosive properties of phenol are the possible reasons. Its weak acidic and highly corrosive properties especially set a high standard for the test facilities and membrane materials to remain their good stability in the solution with high phenol concentration. In one study for the pervaporation dehydration of 80/20 wt % phenol/water feed solution,4 Rhim et al. used poly(vinyl alcohol) (PVA)/poly(acrylic acid) (PAA) (4:1 weight ratio) cross-linked flat-sheet membranes. A high separation factor of ∼3580 was reported with a relative low flux (66 g/(m2 h)). In another work, a polyimide (PI) flat-sheet membrane was employed,11 but the feed employed only contained 8 wt % phenol, and a low separation factor of 18 is reported. In both works, the membrane stabilities in the phenol-containing feed solution were not discussed. Compared to conventional polymeric materials, polybenzimidazole (PBI) is a more suitable candidate as the membrane material for pervaporation dehydration.12−20 PBI is a glassy polymer with a high Tg (417 °C)21 and has stable mechanical properties up to 350 °C. As a high performance aromatic polymeric material, it can survive in aggressive environments, because of its outstanding chemical resistance and thermal stability. In addition, PBI possesses both donor and acceptor hydrogen-bonding sites,22,23 which are capable of participating in specific interactions.24,25 It is also known to absorb 15 wt % water at equilibrium and the water in PBI is mobile.26 Thus, water can preferentially permeate through the PBI membrane, because of its stronger affinity with PBI molecules and smaller molecular size, relative to most organics. All the above characteristics make PBI a promising pervaporation membrane material for the dehydration of various organics, especially for applications in harsh environments. In this work, PBI material is employed as the pervaporation membrane for phenol dehydration. The effects of cross-linking and thermal treatment modifications, as well as operation conditions on the pervaporation performance of PBI flat-sheet dense membranes, are investigated. PBI membranes with various morphologies, i.e., PBI flat-sheet dense membrane, single-layer and dual-layer hollow fiber membranes, are also studied and compared among themselves, as well as with those in previous literature. Via this study, it is expected to facilitate the applications of PBI materials with an enhanced stability in harsh environments. It could also open new perspectives for future research and development on the purification and separation of those corrosive organics.
■
EXPERIMENTAL SECTION Materials. The PBI solution was supplied by PBI Performance Products, Inc.; it had a composition of 26.2 wt % PBI, 72.3 wt % N,N-dimethylacetimide (DMAc), and 1.5 wt % lithium chloride (LiCl). The LiCl ensures that PBI does not phase out of the solution. Poly(vinylidene fluoride) (PVDF) powder (Kynar HSV 900) was purchased from Arkema, Inc., and dried in a vacuum oven at 60 °C overnight before use. DMAc (analytical grade), which was employed as the solvent in the bore fluid and dope solutions, was supplied by Merck and used as-received. Phenol (analytical grade) was purchased from Sigma−Aldrich and mixed with deionized water to prepare the binary feed solution. Analytical grade p-xylene dichloride, from Aldrich Chemistry was used as a cross-linking agent. All other solvents (ethylene glycol, n-heptane, methanol, and ethanol) were of analytical grade and obtained from Merck. Fabrication of PBI Flat-Sheet Dense Membranes. The polymer dope solution of PBI/DMAc/LiCl (15/84.1/0.9 wt %) 18292
dx.doi.org/10.1021/ie502626s | Ind. Eng. Chem. Res. 2014, 53, 18291−18303
Industrial & Engineering Chemistry Research
Article
min). The fibers then were dried in air naturally and kept in an ambient environment. All hollow fiber membranes are exposed to the cross-linking modification by 1 wt % p-xylene dichloride/heptane for 1 h, followed by a post-heat treatment at 75 °C for 2 h in the vacuum oven. The pervaporation module with an effective length of ∼15 cm was prepared by loading two pieces of hollow fiber into a perfluoroalkoxy (PFA) tubing connecting with two Swagelok stainless steel male run tees on the two ends. Both ends were sealed by epoxy and cured for 24 h at ambient temperature. At least two pervaporation modules were tested for each membrane sample. Pervaporation Study. A static pervaporation cell, according to Prof. Matsuura’s design,28 was employed for the pervaporation test of PBI flat-sheet membranes. The schematic design has been shown elsewhere.12 A testing membrane with an effective surface area of 15.2 cm2 was placed in the stainless steel permeation cell. A 350-mL feed solution of binary phenol/ water mixture was used with a phenol content ranging from 0 to 90 wt %. The operating temperature was ranged from 20 °C to 80 °C. For the test of hollow fiber membranes, a laboratoryscale acid-resistant pervaporation unit was employed and the schematic diagram of the apparatus was similar as described in the previous work.29 To avoid the corrosive attack of the feed solution containing high phenol content, almost all stainless steel parts (tubings, valves, etc.) in the feed side were replaced by acid-resistant PFA or PVDF materials, and the feed temperature was controlled by the external thermal tape instead of the conventional heating bath. A 2-L feed solution of 70/30 wt % phenol/water and an operation temperature of 60 °C were used for the pervaporation tests of hollow fiber membranes. The feed flow rate was 0.5 L/min for each hollow fiber module. For all pervaporation tests in both setups, the feed composition was varied
