Research Article www.acsami.org
Multiblock Copolymer Grafting for Butanol Biofuel Recovery by a Sustainable Membrane Process Shankarayya Vijay Kumar, Carole Arnal-Herault, Miao Wang, Jérôme Babin, and Anne Jonquieres* Laboratoire de Chimie Physique Macromoleculaire, Université de Lorraine, CNRS UMR 7375, 1 rue Grandville, BP 20451, 54 001 Nancy Cedex, France S Supporting Information *
ABSTRACT: Biobutanol is an attractive renewable biofuel mainly obtained by the acetone−butanol−ethanol (ABE) fermentation process. Nevertheless, the alcohol concentration has to be limited to a maximum of 2 wt % in ABE fermentation broths to avoid butanol toxicity to the microorganisms. The pervaporation (PV) membrane process is a key sustainable technology for butanol recovery in these challenging conditions. In this work, the grafting of azido-polydimethylsiloxane (PDMSN3) onto a PDMS-based multiblock copolymer containing alkyne side groups led to a series of original membrane materials with increasing PDMS contents from 50 to 71 wt %. Their membrane properties were assessed for butanol recovery by pervaporation from a model aqueous solution containing 2 wt % of n-butanol at 50 °C. The membrane flux J50μm for a reference thickness of 50 μm strongly increased from 84 to 192 g/h m2 with increasing PDMS content for free-standing dense membranes with thicknesses in the range of 38−95 μm. At the same time, the intrinsic butanol permeability increased from 1.47 to 4.68 kg μm/h m2 kPa and the permeate butanol content was also strongly improved from 38 to 53 wt %, corresponding to high and very high membrane separation factors of 30 and 55, respectively. Therefore, the new grafted copolymer materials strongly overcame the common permeability/selectivity trade-off for butanol recovery by a sustainable membrane process. KEYWORDS: grafted copolymer materials, membranes, butanol, biofuel, pervaporation membrane process, structure−property relationships
1. INTRODUCTION With an ever increasing demand for energy, the development of green fuels has been growing rapidly for the past ten years to reduce dependence on oil by promoting renewable energy resources. Currently, fossil fuels are still dominating (∼80%) the world energy economy but biofuels are already offering new interesting alternatives.1 In particular, bioalcohols (ethanol, butanols, ...) are receiving increasing attention from the petrochemical industry.2,3 Although bioethanol remains a leading biofuel worldwide,1,2 biobutanols are considered as promising next-generation biofuels with several specific advantages compared to bioethanol. Their higher energy content, lower vapor pressure, improved solubility for fuel formulations, higher hydrophobicity, and blending capacity are particularly interesting for the formulation of gasoline blends.3,4 With this respect, acetone−butanol−ethanol (ABE) fermentation is a major process to produce biobutanols from renewable resources. In this process, butanols are mainly produced by anaerobic fermentation with Clostridium acetobutylicum bacteria. Nevertheless, butanol concentration has to be limited to a maximum of 2 wt % in ABE fermentation broths to avoid butanol toxicity to the microorganisms.5 In these conditions, the traditional distillation method is highly energy-intensive and costly for butanols recovery due to their © XXXX American Chemical Society
low concentration and high boiling point and their costeffective recovery from fermentation broths remains critical for sustainable biobutanols production.6,7 Considerable efforts have been developed to overcome the distillation limitations with different separation alternatives.8−15 In particular, pervaporation (PV) with organophilic membranes, alone or coupled to gas stripping, required only a small feed fraction be vaporized through the membrane material and offered important energy and cost savings for the recovery of volatile organic compounds (VOCs) from dilute aqueous solutions.8−10,12−21 Nevertheless, most membrane materials are limited by a permeability/selectivity trade-off and the search for highly organophilic membranes remains very active worldwide. Different organophilic membrane materials have been reported for butanol recovery, mainly polydimethylsiloxane (PDMS),16,22−30 and, to a much less extent, polyoctylmethylsiloxane (POMS), 31−33 poly(1-trimethylsilyl-1-propyne) (PTMSP),34−36 polytetrafluoroethylene (PTFE),23,37 polypropylene (PP),38 polyvinylidene fluoride (PVDF),39 and hydrohexafluoroisopropylnorbonene block and random copolyReceived: February 16, 2016 Accepted: June 6, 2016
A
DOI: 10.1021/acsami.6b01900 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
Figure 1. Synthesis of the precursor PDMS-based PUU containing alkyne side groups.
mers.40 A few linear multiblock copolymers (PDMS-based polyureas,41 poly(ether-urethane),22 poly(butadiene-urea-urethane),42 and poly(ether-block-amide)s (PEBA)18,22,43−45) have also been rarely reported for butanol recovery. Beyond the scope of this work, related hybrid polymer membranes have also been developed to further increase their separation properties by adding inorganic fillers, such as silicate, zeolite, activated carbon, and metal−organic frameworks (MOFs), in mixed matrix membranes (MMMs). In a related field, we have recently reported a new strategy consisting in grafting a multiblock copolymer for strongly improving the membrane flux for the separation of the azeotropic mixture ethanol/ethyl tert-butyl ether (ETBE) involved in the purification of another important biofuel (ETBE) widely used in the European Union.46 In the latter work on organoselective membranes, the grafting of a poly(ether-urea-imide) multiblock copolymer with a soft polymethacrylate oligomer enabled to strongly increase membrane permeability (×3), while the membrane selectivity remained in the very high range for this application. In this work, this new grafting strategy is considered for designing original organophilic polymer membranes for butanol recovery from dilute aqueous solutions. The new grafted copolymer materials were PDMS-based to provide a strong hydrophobic character. Compared to the former works47−53 on
multiblock copolymer membranes for different separations, which have investigated the influence of the structure and length of the soft and hard blocks, this approach explores a new way for varying the polymer architecture and offers new opportunities for membrane design. Compared to our former work on organoselective grafted multiblock copolymers for ETBE purification,46 this work reports original PDMS-based materials with a new architecture. The grafting strategy had to be specifically adapted to avoid the formation of highly hydrophilic ammonium groups reported in our former work and to overcome the usual solubility limitations encountered with PDMS-based copolymers. Furthermore, the influence of PDMS grafting on the membrane properties observed in this new work was very different from that of grafting hydrophilic grafts onto a hydrophilic multiblock copolymer reported in our former work. The last part of this paper will show the new potential of the grafting strategy for improving both permeability and selectivity simultaneously for butanol recovery by pervaporation with original organophilic membranes displaying very high selectivity. Since the pioneering work of Du Prez et al.,54 combining step-growth polymerization with the copper(I)-catalyzed 1,3dipolar azide−alkyne cycloaddition (CuAAC) “click” chemistry has become one of the best approaches to polyurethane multiblock copolymers with complex architectures and B
DOI: 10.1021/acsami.6b01900 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
deionized water, the polymer fibers were dried under vacuum at 60 °C overnight to get final product. Yield: 98%. SEC-MALLS: Mn = 58 290 g/mol, Mw = 73 990 g/mol, Đ = 1.3. 1H NMR (THF-d8): δ (ppm) 8.64 (s, 4H, Hf), 7.70 (bs, 2H, Hj), 7.45 (bs, 2H, Hl), 6.96- 7.37 (m, 32H, Ha, Hb, Hd, He), 5.45 (bs, 2H, Hk), 4.21 (s, 8H, Hg), 4.10 (m, 4H, Ht), 3.79 (m, 8H, Hc), 3.14 (m, 4H, Hp), 2.51 (s, 8H, Hh), 2.42 (s, 4H, Hi), 1.53 (m, 8H, Hq, Hu), 0.58 (m, 8H, Hr), 0.90 (s, 104 H, Hs) with 16.4 siloxane units according to 1H NMR analysis of the α,ω-aminopropyl-terminated PDMS (not shown). 2.2.2. Grafting of a Poly(Urea-urethane) Copolymer Containing Alkyne Side Groups. An azido-terminated PDMS oligomer (PDMSN3, Mn = 1169 g/mol) was grafted onto the multiblock poly(ureaurethane) (PDMS−PUU) with alkyne side groups by CuAAC “click” chemistry with different grafting rates. The detailed synthesis and characterization of the PDMS-N3 oligomer are described in Supporting Information (Figure S2). As way of example, the following procedure describes the grafting of the multiblock copolymer with a PDMS grafting rate of 50%. In a Schlenk tube, 1.6 g of PDMS−PUU copolymer (corresponding to 2.30 mmol of alkyne side groups) and 1.17 g (1.15 mmol) of PDMS-N3 dissolved in 25 mL of dry THF: DMF (1:1) solvent mixture. The tube was degassed by three standard freeze−thaw cycles and flushed with dry argon. A solution of copper chloride (0.2 equiv/alkyne group; 22.79 mg, 0.23 mmol) and PMDETA (0.2 equiv/alkyne groups; 39.86 mg, 0.23 mmol) in 5 mL of degassed dry DMF was added to the reaction mixture by canula and the grafting reaction was carried out at 80 °C for 48 h under nitrogen atmosphere. After concentration under reduced pressure, the resulting viscous solution was precipitated in deionized water. After it was washed two times with deionized water, the grafted copolymer was dried under vacuum overnight at 60 °C. Yield: 95%. The grafted copolymers with different grafting rates were characterized quantitatively by 1H NMR in THF-d8 (see main text and Figure S3). 2.2.3. Characterization. A Bruker Avance 300 spectrometer at 300.15 MHz was used for recording the 1H NMR spectra in CDCl3, DMS-d6 or THF-d8 purchased from Euriso-top. The reference for the chemical shifts was TMS and the isotopic impurities of the deuterated solvent were used for chemical shift referencing. For FTIR analysis of the new materials, thin polymer films were obtained on KBr disks by vacuum solvent evaporation from a deposited polymer solution. A Bruker Tensor 127 FTIR spectrometer was then used for recording the FTIR spectra in transmission mode. Size exclusion chromatography using multiangle laser light scattering (SEC-MALLS) was performed to determine the polymer molecular weights. Two detectors from Wyatt Technology (i.e., a MALLS TREOS and an Optilab rEX differential refractometer) were used for SEC-MALLS analysis. On the basis of a calibration obtained from the differential refractometer, the measured value for dn/dC of the “clickable” multiblock copolymer was 0.089 mL/g at a wavelength of 654 nm. Two columns (PL gel 5 μm Mixed-D 300−7.5 mm) were used in series at 70 °C to ensure efficient separation in THF of HPLC analytical grade. Before injecting, the polymer solutions in THF (C = 0.15% w/v) were filtered with 0.2 μm PTFE membranes (Alltech). A TA Instruments DSC Q2000 was used to perform modulated differential scanning calorimetry (MDSC) on polymer samples (8−10 mg) under nitrogen flow. Each modulation was achieved over 40 s with an amplitude of 0.5 °C. A classical procedure of scanning in two cycles from −70 to 200 °C was used to erase the former thermal history of each polymer sample with heating and cooling rates of 5 °C/ min. 2.3. Membrane Preparation for Sorption and Pervaporation Experiments. The sorption and pervaporation membranes were prepared by adapting our formerly reported procedure.46 In this work, the solvent used for preparing the polymer solutions was THF (pure for synthesis) and the molds for membrane casting were made of glass. For assessing the average membrane thickness, the membranes were divided in eight different radial sectors and the thickness measurements were made for each sector with a Roch micrometer (Luneville, France). This procedure is widely applied to measure the thickness of industrial films in the thickness range of interest with a measurement
functionalities.55 In particular, polyurethanes with alkyne side groups have been developed as versatile polymer platforms for grafting fluorescent dyes,54,56 bactericidal moieties,57 highly fluorinated chains,54,58 sugar derivatives,59 various nonlinear optical chromophores,60 biocompatible PEO oligomers,61 and flame retardants.62,63 The first part of this work describes the grafting of a new PDMS-based poly(urea-urethane) multiblock copolymer containing alkyne side groups with different PDMS ratios by CuAAC “click” chemistry, leading to a series of highly organophilic membrane materials with original architecture for butanol recovery by pervaporation. An analysis of their chemical structure and morphology is then presented on the basis of different complementary techniques. The second part describes their membrane properties for the sorption and pervaporation of a model aqueous solution containing 2 wt % of n-butanol. The strong simultaneous increase in flux and selectivity obtained with the new organophilic membranes is discussed in terms of structure−property relationship.
2. EXPERIMENTAL SECTION 2.1. Solvents and Reagents. α,ω-Aminopropyl-terminated PDMS (Mn of 1385 g/mol, as determined by 1H NMR) and monocarbinol terminated PDMS (Mn of 1015 g/mol, as determined by 1H NMR) were purchased from Gelset, ABCR GmbH & Co. KG, Germany and were used without further purification. All other reagents and solvents were purchased from Sigma-Aldrich unless otherwise specified. Dimethylmalonate, sodium ethoxide, propargyl bromide, N,N,N′,N″,N″-pentamethyldiethylenetriamine (PMDETA), 1,4-butanediol (BD), copper(I) chloride, 4-toluenesulfonyl chloride (TsCl), sodium azide (NaN3), propargyl bromide 80 wt % in toluene were used as received. 4,4′-methylene-bis-phenylisocyanate (MDI) was distilled under vacuum and stored at −20 °C. N,Ndimethylformamide (DMF, pure for synthesis) and N,N-dimethylacetamide (DMAc, pure for synthesis) were fractionally distilled under reduced pressure over calcium hydride to remove any water and residual hydrolysis products. Tetrahydrofuran (THF) (pure for synthesis, Carlo-Erba) was purified by distillation over sodium/ benzophenone under argon atmosphere. The distilled solvents were stored over 4 Å molecular sieves. All reagents and solvents were stored under dry argon to avoid contamination by atmospheric moisture. 2.2. Synthesis and Grafting of a Poly(Urea-urethane) Copolymer Containing Alkyne Side Groups. 2.2.1. Synthesis of a Poly(Urea-urethane) Copolymer Containing Alkyne Side Groups. The synthesis of a PDMS-based multiblock poly(urea-urethane) (PDMS−PUU) with alkyne side groups for the CuAAC “click” chemistry was carried out in three successive steps by using theoretical number equivalents of 2 and 1 for the dialkyne diol monomer (DPPD) and α,ω-aminopropyl terminated PDMS respectively (Figure 1). The detailed synthesis and characterization of DPPD are described in Supporting Information (Figure S1). First, in 250 mL three-neck round-bottom flask fitted with nitrogen inlet, MDI (4.505 g; 18.0 mmol), DPPD (1.370 g; 9.0 mmol), and dibutyltin dilaurate as catalyst (0.10 g ; 0.16 mmol) were dissolved in 60 mL of THF:DMAc (70:30) solvent mixture. The mixture was stirred for 2 h at 60 °C. At the end of this first step, the OH band at 3288 cm−1 disappeared in the FTIR spectrum as expected after the formation of the DPPD-based diisocyanate. In the second step, the reaction mixture was cooled down to room temperature and α,ω-aminopropyl-terminated PDMS (6.233g; 4.50 mmol) dissolved in 30 mL of THF:DMAc (70:30) solvent mixture was slowly added and stirrer for another 2 h to obtain the corresponding macrodiisocyanate. FTIR analysis showed the presence of the corresponding characteristic NCO stretching band at 2270 cm−1 and the appearing of the urea stretching band at around 3280 cm−1. In third step, the macrodiisocyanate was reacted with one equivalent of 1,4-butanediol (0.406g; 4.50 mmol) at 70 °C for 3 h. The reaction mixture was cooled to room temperature, concentrated and precipitated in deionized water. After washing twice with C
DOI: 10.1021/acsami.6b01900 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
Figure 2. Grafting of the precursor PDMS-based PUU with azido-functionalized PDMS. error of 1 μm. The thicknesses of the sorption and pervaporation membranes were typically ∼200 and 50 μm except for the highest grafting rate, respectively. In the latter case, the PV membrane thickness had to be almost doubled owing to the high material softness, but it still remained lower than 100 μm, which was the limit formerly reported for free-standing dense PDMS membranes.41 Nevertheless, all the free-standing membranes obtained in these conditions were mechanically resistant enough to withstand the pervaporation operating conditions at 50 °C. 2.4. Sorption Measurements. After the sorption membranes were dried under vacuum, the initial weight of each dry membrane, wD, was determined. These membranes were then equilibrated in a model dilute aqueous solution containing 2 wt % of n-butanol in hermetically closed bottles for several days at 30 °C to avoid any safety issues related to handling hot samples. The weight of each swollen membrane at sorption equilibrium, wS, was estimated by gravimetry following our formerly reported procedure.64 The total swelling S (wt %) was calculated from eq 1:
S=
wS − wD × 100 wD
permeate flux =
wp (2)
Δt × A
Normalized fluxes, J50μm, were reported for a reference thickness of 50 μm corresponding to the average thicknesses for all the membranes except for the highest grafting rate (eq 3). The error was less than 5% for the total fluxes. J50μm =
wp Δt × A
×
actual membrane thickness 50
(3)
The permeate butanol weight fraction, C′BuOH, was determined by gas chromatography using our formerly reported procedure.16 The error made for the determination of the butanol weight fractions was ±0.005. The membrane separation factor, βBuOH/water, was calculated from eq 465
βBuOH/water =
′ CBuOH ′ 1 − CBuOH C 1−C
(4)
where C′BuOH and C are the butanol weight fraction in the permeate and feed, respectively. The partial permeability, Pi, were also calculated to compare intrinsic membrane properties irrespective of the membrane operating conditions (see Supporting Information for the detailed calculation procedure).
(1)
2.5. Pervaporation Measurements. Pervaporation was carried out with a model dilute aqueous solution containing 2 wt % of nbutanol at 50 °C with a constant stirring of the feed at 100 rpm. The membranes with an active area A of 1.66 × 10−3 m2 were operated under low pressure (
