Research Article pubs.acs.org/journal/ascecg
Realizing Mussel-Inspired Polydopamine Selective Layer with Strong Solvent Resistance in Nanofiltration toward Sustainable Reclamation Yanchao Xu, Fangjie You, Hongguang Sun, and Lu Shao* MIIT Key Laboratory of Critical Materials Technology for New Energy Conversion and Storage, State Key Laboratory of Urban Water Resource and Environment, School of Chemistry and Chemical Engineering, Harbin Institute of Technology, 92 Xidazhi Street, Harbin 150001, China S Supporting Information *
ABSTRACT: Herein, a mussel-inspired polydopamine (PDA) coating layer has been first explored as a separating layer for organic solvent nanofiltration (OSN). A PDA based separating layer was constructed on polyimide (PI) support via dopamine coating. The subsequent membrane was then treated with 1,6-hexanediamine for cross-linking on both the PDA layer and PI support. Fourier transform infrared (ATR-FTIR) and X-ray photoelectron (XPS) results indicated the deposition of PDA on support and the cross-linked structures of both PDA and PI. Scanning electron microscopy (SEM), atomic force microscopy (AFM), water contact angle, and surface energy measurements were further employed to characterize the morphologies and surface properties of the composite membranes. After an optimized coating time of 4 h, the resultant membrane showed an EtOH permeance of 0.91 L m−2 h−1 bar−1 and a RB rejection of 99%. More importantly, the composite membrane also exhibited good performance for dyes separation from a wide range of solvents including challenging polar aprotic and strongly swelling solvents, such as dimethylformamide and acetone. In addition to demonstrating a facile and effective OSN membrane fabrication approach, this study may stimulate the bioinspired design of a composite membrane for sustainable applications. KEYWORDS: Organic solvent nanofiltration, Polydopamine, Polyimide, Cross-linking, Membrane stability
■
many polymer materials, such as polyimide (PI),4 polybenzimidazole (PBI),5 poly(ether block amide),6 etc., have been cross-linked by various approaches to prepare stable OSN membranes. The most common polymer for OSN application is polyimide (PI) due to the fact that it can be simply crosslinked by diamine via the reaction between amine groups and imide groups.4 Many reported cross-linked PI OSN membranes are integrally skinned asymmetric membranes prepared via a phase inversion technique followed by diamine cross-linking, but these membranes suffer from limitations in terms of permeance for certain solvents. A thin film composite membrane consisting of an ultrathin separating layer on porous cross-linked PI supports is another important kind of OSN membrane. The ultrathin separating layer most commonly
INTRODUCTION Organic solvent nanofiltration (OSN) is an relatively young pressure driven separation that possess the ability to fractionate components with a molecular weight between 200 and 1000 g mol−1 from organic solvents.1−3 As a promising alternative to traditional purification and separation technologies, such as distillation, extraction, and chromatography, OSN has the merits of economy, environmental friendliness, and safety. It has great application prospects in petrochemistry, fine chemicals, and pharmaceutical industries. One of the main challenging issues in OSN is the membrane stability in organic solvent. Ceramic membranes are superior, considering they do not swell or deform in organic solvents; however, their large scale fabrication is complicated, and they are more fragile and costly than polymeric membranes. For these reasons, most of the reported OSN membranes are made out of polymeric materials, and chemical cross-linking of the polymer is usually employed to achieve high stability in harsh solvent. So far, © 2017 American Chemical Society
Received: March 22, 2017 Revised: May 11, 2017 Published: May 16, 2017 5520
DOI: 10.1021/acssuschemeng.7b00871 ACS Sustainable Chem. Eng. 2017, 5, 5520−5528
Research Article
ACS Sustainable Chemistry & Engineering fabricated by interfacial polymerization or dip-coating can be independently designed to maximize the overall membrane performance. 1,6-Hexanediamine (HDA) cross-linked PI ultrafiltration membranes have been widely used as porous supports to prepare composite OSN membranes.7−9 Dopamine (DA), a mussel adhesive protein-inspired molecule, has draw intensive attention due to the fact that it can spontaneously oxidize and self-polymerize in weakly alkaline environment and form a thin polydopamine (PDA) layer onto nearly all types of surfaces, regardless of the material surface.10−13 More importantly, due to the existence of remaining catechol groups, the PDA layer can take the role of a versatile reaction platform for secondary modification via Micheal addition or Schiff base reactions. It has been widely used as a universal and versatile tool in membrane preparation, although the extra DA polymerization mechanism and PDA molecular structure are still unknown so far.14 In most cases, the PDA layer was used as a versatile intermediate layer for further modification to achieve the desirable separation properties. Several trials have been undertaken to explore the possibility of employment of PDA as the separating layer for an aqueous nanofiltration (NF) membrane. Li et al. have fabricated a NF membrane by dipping polysulfone substrate in dopamine solution. The obtained membrane showed high flux while the rejection is unsatisfactory.15 Some other works have further reinforced the PDA separating layer by crosslinking with polyethylenimine, which can make the layer denser and meanwhile render the membrane surface positively charged.16−18 However, due to the steric hindrance of high molecular polyethylenimine, the cross-linking efficiency is low, and some small molecular compounds were further used to cross-link polyethylenimine so as to make the layer dense and compact enough for good rejection. These make the fabrication process of these membranes tedious and time-consuming and generate amounts of chemical pollutant at different stages. So far, the utilizations of PDA for OSN application are still rare. Mu et al. have prepared thin film composite OSN membranes by dispersing PDA nanoparticles in the aqueous phase prior to the interfacial polymerization. It is proven that the introduction of PDA nanoparticles in the polyamide separating layer can improve the membrane hydrophilicity and solvent resistance.19 In addition, there are some other reports focusing on construction of a specific thin film composite separating layer for OSN using dopamine as monomer in the aqueous phase via interfacial polymerization.20,21 Strictly speaking, the separating layers are polyamide/polyester, which are formed via nucleophilic substitution reaction between amine/hydroxyl groups in dopamine and trimesoyl chloride, not a PDA based layer. Although the insolubility property of PDA in organic solvents potentially allows its application in OSN,22 the direct utilization of PDA as the separating layer for OSN membranes, as far as we know, has not yet been explored. Herein, for the first time, we reveal the potential of PDA to be incorporated as a separating layer on a cross-linked PI support for OSN and further optimize its performance via HDA induced cross-linking. In an improved fabrication route, the simultaneous cross-linking of both the PDA coating and PI support can be completed in a single step, which was much more simple and time-saving, as shown in Figure 1. The resultant composite membrane exhibited desirable solvent permeance with high rejection for dyes in a wide range solvents, including challenging polar aprotic and strongly swelling solvents. The simplicity of the fabrication strategy,
Figure 1. Illustration for the fabrication of a mussel-inspired composited nanofiltration membrane for organic solvent nanofiltration.
good separation performance, and the high stability of this composite membrane predict the potentials in separation and purification in many industrial fields.
■
EXPERIMENTAL SECTION
Materials. P84 polyimide polymer was provided by Granulat SG STD. 1,6-Hexanediamine (HDA), tris(hydroxymethyl)aminomethane (Tris), Methyl Orange (MO), Crystal Violet (CV), Orange G (OG), Acid Fuchsin (AF), Methyl Blue (MB), Rose Bengal (RB), and Solvent Blue II (SB II) were obtained from Aladdin Industrial Co., Ltd. Dopamine hydrochloride (DA) was purchased from SigmaAldrich (USA). All organic solvents were supplied by Xilong Chemical Industrial Co., Ltd. All used water was deionized. Preparation of PI Supports. PI supports were produced via a nonsolvent induced phase separation process. A dope solution was formed by dissolving 16 wt % of P84 polymer in N-methylpyrrolidinone (NMP) and stirring until complete dissolution. The solution was left overnight to disengage air bubbles and then cast on a glass plate using a casting knife set at a thickness of 200 μm. After casting, the membrane was immersed in a water coagulation bath and the solidification of membrane occurred. After 30 min, the membrane was transferred to a fresh water bath and stored for further use. Preparation of Composited Membranes. Two fabrication methodologies were carried out, as shown in Figure 2. Fabrication methodology A was devised in a routine procedure, in which a crosslinked PI support was prepared followed with the construction of active layer on it. This fabrication methodology allows ascertaining the utilization of a pure PDA layer as a separating layer. Fabrication methodology B is an improved methodology with a simplified fabrication process. Fabrication Methodology A. PI supports were immersed with IPA, and then transferred to a 20 g L−1 solution of HDA in IPA for 24 h for cross-linking. The resultant membrane was washed with IPA and denoted as cPI. The cPI membranes were rinsed with DI water, and then immersed in a freshly prepared 2 mg mL−1 dopamine Tris-HCl solution (pH 8.5) and shaken at 30 °C for a certain time. The modified membranes were washed with DI water and denoted as PDA/cPI. Furthermore, these membranes were rinsed with IPA and immersed in a 20 g L−1 solution of HDA in IPA for 24 h for further cross-linking of PDA layers. The obtained membranes were washed with IPA and denoted as cPDA/cPI. Fabrication Methodology B. PI supports were immersed in a freshly prepared 2 mg mL−1 dopamine Tris-HCl solution (pH 8.5) and shaken at 30 °C for a certain time. The modified membranes were rinsed with DI water and denoted as PDA/PI. Furthermore, these membranes were rinsed with IPA and immersed in a 20 g L−1 solution of HDA in IPA for 24 h for cross-linking of the whole membrane. The obtained membranes were washed with IPA and denoted as c(PDA/ PI). Membrane Characterization. The chemical structure of the membranes was characterized by attenuated total reflection Fourier transform infrared spectroscopy (ATR-FTIR, PerkinElmer, USA) and X-ray photoelectron spectra (XPS, AXIS UltraDLD, SHIMADZU, 5521
DOI: 10.1021/acssuschemeng.7b00871 ACS Sustainable Chem. Eng. 2017, 5, 5520−5528
Research Article
ACS Sustainable Chemistry & Engineering
Figure 2. Schematic of membrane fabrication methodologies A and B.
Figure 3. (A) Possible mechanism of PDA formation and (B) cross-linking reaction of PDA by HDA. Japan). The morphologies of the membranes were characterized by scanning electron microscopy (SEM, Hitachi SU8000). The surface roughness of the membranes was observed by atomic force microscopy (AFM, Nanoscope IIIA) in air atmosphere. The water contact angles of membranes were characterized by a contact angle measuring system (G10 Kruss, Germany). Deionized water and ethylene glycol were used to measure the surface energy (SE). The reported contact angle was calculated by averaging over more than five contact angle values at different sites. SE is the sum of the disperse and polar parts, which were calculated by eq 1: γ1(1 + cos θ) =
4γsdγ1d γsd + γ1d
+
4γspγ1p γsp + γ1p
P=
V A × t × ΔP
(2)
where P is permeance (L m−2 h−1 bar−1), V (L) is the volume of the permeate, A is the membrane active surface (m2), t is the operation time (h), and ΔP is the applied pressure across the membrane (bar). The rejections of dyes were calculated from eq 3: ⎛ Cp ⎞ ⎟ × 100% R = ⎜⎜1 − Cf ⎟⎠ ⎝
(3)
where Cp and Cf were the dye concentrations in permeate and feed, respectively. The concentrations of dyes in solvents were measured by a UV−vis Cintra20-GBC apparatus.
(1)
■
where γ refers to surface energy, the subscripts l and s refer to liquid and solid, and the superscripts d and p refer to dispersive and polar components, respectively. θ refers to contact angles between ordinary liquids (H2O or ethylene glycol) and the membranes. Nanofiltration Experimental Procedure. The nanofiltration experiments were carried out at 5 bar and 25 °C with a self-made filtration apparatus (Figure S1). The active surface area of each membrane was 21.1 cm2, and at least three independently prepared membranes were tested. The membranes were compacted for 2 h at 5 bar to reach a steady value, and then the permeance and rejection were measured. The feed containing 35 μM dyes in a variety of solvents was poured into the cell and stirred at 11.66 Hz (700 rpm) to minimize the possible concentration polarization. The solvent permeance was measured as given in eq 2:
RESULTS AND DISCUSSION Possible Mechanisms of PDA Formation and Its Cross-linking by HDA. The most recognizable reaction pathway leading to PDA, as shown in Figure 3A, is that DA is first oxidized to dopaminequinone under alkaline conditions, followed by cyclization to yield leukodopaminechrome.10,23 After that, leukodopaminechrome further suffers from oxidization and rearrangement to form 5,6-dihydroxyindole. The branching reaction between 5,6-dihydroxyindole structures leads to the formation of covalent binding among aryl rings. These covalent bindings, together with other noncovalent bindings, such as π−π stacking, hydrogen bonding, and charge 5522
DOI: 10.1021/acssuschemeng.7b00871 ACS Sustainable Chem. Eng. 2017, 5, 5520−5528
Research Article
ACS Sustainable Chemistry & Engineering
structure, respectively, significantly increased.31 A broad peak at around 3300 cm−1 related to the −NH− and −OH groups in PDA can be observed. These proved the successful coating of PDA on the PI surface. After cross-linking, the imide peaks attenuated and the amide peaks appeared in the spectra of c(PDA/PI). The broad peak at 3300 cm−1 shifted to 3270 cm−1, and its intensity dramatically enhanced, which can be related to the −NH− in cross-linked PI and PDA structures. Most importantly, the spectrum of c(PDA/PI) is consist with that of cPDA/cPI. A more detailed study of the membrane chemistry was investigated by XPS analysis. The element contents of the membrane surface are presented in Table 1, and deconvolution
transfer interactions, lead to the deposition of PDA on substrates.24 Moreover, in an alkaline solution, the catechol groups in PDA are ready to be oxidized to the corresponding quinone form. The quinone can spontaneously react with amines via Michael addition or Schiff base reaction to form a covalent −C−NH or −CN bond.10 This enables the cross-linking reaction of the PDA coating by HDA, as shown in Figure 3B. Chemical Properties of Membranes. Figure 4A shows the ATR-FTIR spectra of membranes in fabrication method-
Table 1. Elemental Composition of Membranes as Determined by XPS membrane
C (%)
N (%)
O (%)
PI cPI PDA/cPI cPDA/cPI PI PDA/PI c(PDA/PI)
78.05 77.69 70 73.55 78.05 69.79 73.27
6.11 9.07 6.78 8.55 6.11 5.27 8.6
15.84 13.25 23.22 17.91 15.84 24.94 18.12
of the O 1s spectrum was shown in Figure 5. In fabrication methodology A, cPI showed a decreased O content compared to PI due to the introduction of HDA during cross-linking (Table 1). Both these membranes exhibited a signal O 1s peak which is attributed to −CO in imide and amide groups, respectively. After dopamine coating, a significant increase in O content could be found in PDA/cPI due to the fact that the support has been covered by PDA and the O content in PDA is richer than that in the support. Meanwhile, a peak at 532.5 eV assigned to catechol C−OH was observed. With the crosslinking of HDA, the intensity of this peak decreased and it could be related to the oxidation of catechol to quinone, which was further subjected to the following Michael addition or Schiff base reaction, as shown in Figure 3B. Meanwhile, the O content in cPDA/cPI decreased as a result of the substitution of O by N during Schiff base reaction. In fabrication methodology B, the O content also increased after dopamine coating and decreased after a further crosslinking. For the O 1s spectrum, the peak related to the catechol C−OH peak appeared after dopamine coating and its intensity decreased after cross-linking, showing a similar trend to that in fabrication methodology A. The elemental composition and O 1s spectra of c(PDA/PI) are very close to that of cPDA/cPI. Considering their similar chemical properties, as indicated by ATR-FTIR and XPS results, as well as their identical performance separation, which will be further discussed below, we focused the following characterizations on membranes in fabrication methodology B. Morphologies of Membranes. Figure 6 shows the morphologies of the PI support and the PDA/PI and c(PDA/PI) membranes. Top surface SEM images showed that the PI support had a relatively clear surface, while the PDA/PI composite membrane displayed some nanoaggregates, indicating the formation of PDA on the membrane surface. During the self-polymerization process of dopamine, PDA generated from covalent cross-linking and subsequent noncovalent aggregation would deposit onto the membrane surface
Figure 4. ATR-FTIR spectra of membranes in fabrication methodologies A (A) and B (B).
ology A. The PI support showed typical imide peaks at 1780, 1718, and 1352 cm−1. The above imide peaks significantly attenuated; meanwhile, the amide peaks at 1648 and 1534 cm−1 emerged in the spectra of cPI. In addition, a broad peak at 3270 cm−1 related to −NH− in amide, and two peaks at 2936 and 2862 cm−1 originated from −CH2− in HDA also proved the successful cross-linking of PI support.25−27 After dopamine coating, the peak at 3270 cm−1 became stronger in intensity, which can be ascribed to both catechol −OH and −NH from the PDA layer.28,29 This peak slightly decreased after a further cross-linking of the PDA layer, which might be due to the conversion of catechol −OH to quinone.30 In addition, membrane relative mass gain suggests the increased density of the PDA layer after cross-linking (Figure S2). Figure 4B shows the ATR-FTIR spectra of membranes in fabrication methodology B. Compared with the pristine PI support, the PDA modified membrane PDA/PI exhibited enhanced adsorption ranging from 1660 to 1050 cm−1. Typically, two peaks at 1605 and 1509 cm−1, which were attributed to the −CC− in the aromatic ring and the −NH− 5523
DOI: 10.1021/acssuschemeng.7b00871 ACS Sustainable Chem. Eng. 2017, 5, 5520−5528
Research Article
ACS Sustainable Chemistry & Engineering
Figure 5. O 1s spectra of membranes.
greatly increased to 14.9 after dopamine coating, which could be ascribed to the deposition of PDA nanoaggregates on the membrane surface. After cross-linking, a decrease down to 9.95 was observed. The changes of roughness were extremely consistent with the changes observed from SEM images, which further confirm the deposition of the PDA layer and the subsequent cross-linking. Membrane Surface Properties. The static and timedependent water contact angles of membranes are shown in Figure 7. The pristine PI support showed the highest water
Figure 6. Top surface (top), cross-sectional (middle), and AFM (bottom) images of membranes.
to form nanoaggregates.32 Some aggregates with larger size could be due to the conjugation and accumulation of PDA nanoaggregates. After cross-linking of HDA, the heterogeneity of the membrane top surface was decreased and only a few nanoaggregates can be observed, which could be due to the introduction of flexible HDA chains. Similar observations have also been reported for the polyethylenimine cross-linked PDA surface.17 Cross-sectional images revealed that there is no significant boundary between support and coating layers, indicating the good compatibility between PI and PDA. The blurry boundary between the support and the PDA layer has also been reported in other works,32−34 and the generally accepted reasons for this phenomenon are (1) the small DA molecule penetrated into the nanopores of the support and polymerized in the pores, obscuring the boundaries, and (2) the PDA layer is too thin (usually less than 50 nm) to be detected.10 AFM images showed that the top surface of PI support was relatively smooth, with a roughness of 5.79, while the roughness
Figure 7. Water contact angle of membranes versus drop age.
contact angle of 67°, and it declined only 3° in 120 s. In comparison, the water contact angle of PDA/PI decreased down to 52°, and it declined 14° in 120 s, demonstrating the improved hydrophilicity and wettability. These can be explained by the synergic effect of the hydrophilic hydroxyl groups in PDA and the relatively higher membrane surface roughness. Moreover, after cross-linking by HDA, the hydrophilicity and wettability of the membrane surface were decreased as the water contact angle was elevated to 60° and it declined 10° in 120 s. These could be related to the conversion of hydrophilic −OH groups to hydrophobic −C N− groups after cross-linking, which also concurred with FTIR and XPS results. In addition, the decreased membrane surface 5524
DOI: 10.1021/acssuschemeng.7b00871 ACS Sustainable Chem. Eng. 2017, 5, 5520−5528
Research Article
ACS Sustainable Chemistry & Engineering
permeation resistance. Meanwhile, the intrinsic cross-linked and stacked PDA structure could provide a sieve effect for dye separation to some degree and give rise to an increased RB rejection. However, due to the fact that the pure PDA layer was not dense enough, the RB rejection was only 74% even when EtOH permeance was compromised to 0.2 L m−2 h−1 bar−1, showing an unsatisfactory performance. The effect of cross-linking of the PDA layer on membrane performance was further investigated. As shown in Figure 8B, after a further cross-linking of the PDA layer, the RB rejection was considerably elevated at any tested coating time. This could be due to the fact that the loose PDA layer became relatively denser for RB rejection after the introduction of a flexible HDA chain via covalent bonds. The EtOH permeance was correspondingly decreased because of the increased resistance from the active layer. At an optimized coating time of 4 h, the cPDA/cPI membrane exhibited an EtOH permeance of 0.86 L m−2 h−1 bar−1 and a RB rejection of 98%. Membrane performances in DMF and toluene also exhibited increased dye rejection of cPDA/cPI compared to PDA/cPI, further indicating the improvement of PDA cross-linking on separation performance (Figure S3). For the comparison of the separation performance between cPDA/cPI and c(PDA/PI), the dopamine coating times in both fabrication methodologies were fixed at 4 h. As shown in Figure 8C, c(PDA/PI) showed a EtOH permeance of 0.91 L m−2 h−1 bar−1 and an RB rejection of 99%, which is a nearly equivalent separation performance to cPDA/cPI. But narrower variations in both EtOH permeance and RB rejection were observed for c(PDA/PI). These narrower variations may be expected, as the fabrication methodology B was more simplified and brought in less system errors. In addition, fabrication methodology B avoids a repeating cross-linking operation; thus, time and effort are saved during the synthesis process. Moreover, it requires less reagent and creates less waste, which are of great interest from an environmental point of view.
roughness could be also responsible for the decreased hydrophilicity of c(PDA/PI). Membrane surface energy characterizations were further carried out to elucidate the membrane surface property, and the results are shown in Table 2. Although the membranes showed Table 2. Contact Angles and Surface Energies of Membranes Contact Angle (deg)a
a
Surface-energy components (mJ m−2)
Membrane
Water
Ethylene Glycol
γLd
γLp
γL
PI PDA/PI c(PDA/PI)
67 53 60
30 23 31
30.00 17.16 19.68
12.30 29.77 22.52
42.30 46.93 42.20
The error in the contact angle measurement is no more than 2°.
close values in total surface energy, significant differences in dispersive and polar components can be observed. Comparing to the pristine PI support, the polar component of PDA/PI dramatically increased, while its dispersive component decreased, which can be rational considering the abundant − OH groups at PDA/PI surface. After cross-linking, due to the hydrophilic −OH groups have partly conversed to hydrophobic − CN- groups, the polar component of c(PDA/PI) has decreased and its dispersive component slightly increased. Membrane Performance. The effect of a pure PDA layer on membrane separation performance was shown in Figure 8A. It can be seen that the pristine cPI support showed an EtOH permeance of 142 L m−2 h−1 bar−1 and a RB rejection less than 10%, demonstrating its ultrafiltration nature. After PDA coating, the EtOH permeance decreased and RB rejection increased significantly. The EtOH permeance further decreased from 24.2 to 0.2 L m−2 h−1 bar−1 while the RB rejection improved from 14% to 74% with prolonging coating time from 1 to 6 h. The decrease of EtOH permeance could be rationalized as due to a continuous PDA layer gradually forming on the support surface, which would enhance solvent
Figure 8. (A) Separation performances of PDA/cPI in terms of coating time, (B) separation performances of cPDA/cPI in terms of coating time, (C) comparison of separation performance between cPDA/cPI and c(PDA/PI) (coating time of 4 h), (D) separation performances of c(PDA/PI) toward various dyes in EtOH, (E) separation performances of c(PDA/PI) in various solvents, and (F) long-term separation performance of c(PDA/ PI) in EtOH and DMF. 5525
DOI: 10.1021/acssuschemeng.7b00871 ACS Sustainable Chem. Eng. 2017, 5, 5520−5528
Research Article
ACS Sustainable Chemistry & Engineering Table 3. Comparison of Some Reported OSN Membranes to c(PDA/PI) in This Studya
a
Membrane material
Stability in polar aprotic solvent
Solvent
Permeance (L m−2 h−1 bar−1)
Solute (MW)
Rejection (%)
(Catechol/PEI)/PAN
dissolve
(DA/TMC)/PES (DA/TMC)/cPAN c(PDA/PI)
dissolve stable stable
IPA EtOH EtOH DMF IPA EtOH DMF
0.26 0.6 3.2 3.1 0.3 0.91 0.63
RB (1017) RB (1017) MB (800) RB (1017) RB (1017) MB (800) RB (1017)
99 84 81 93 99 99 99
ref 30 16 17 This work
PEI, PAN, and PES are the abbreviations of polyethylenimine, polyacrylonitrile, and poly(ether sulfone).
equivalent or higher dye rejections. Although the EtOH and DMF permeance of c(PDA/PI) are lower than those of the two DA based polyamide membranes, it shows much higher dye rejections.
The separation performances of c(PDA/PI) toward various dyes in EtOH were investigated and shown in Figure 8D. The chemical structures, molar weights, and charges of these dyes are summarized in Figure S4. It can be seen that the EtOH permeance changed little for different dyes, indicating the type of dyes has no significant influence on solvent permeance. In addition, the dye rejections depend mainly upon their molecular weights. Dyes with relatively higher molecular weight, such as RB and MB, were nearly completely rejected. Although the rejections to MO, CV, and OG were not so satisfactory, they can be further improved by prolonging the PDA coating time to 5 h (Figure S5). These results indicate the obvious sieving effect in this system. The separation performances of c(PDA/PI) in various solvents were shown in Figure 8E. Due to the fact that RB is insoluble in toluene, an oil-soluble dye SB 35 was used to replace RB in a toluene solution. It can be seen that the membrane showed good separation performances in a wide range of solvents, including alcohols, polar aprotic solvents, and nonpolar solvents, which can be ascribed to the cross-linked structures in both the selective layer and the support. No significant trend has been observed for filtrations in different solvents. The solvent/solute transport mechanism in OSN is much more complex and less discovered than aqueous nanofiltration. Many factors, such as specific properties of used solvents (molar volume, viscosity, surface tension, etc.) and solutes (structure, charge, etc.), as well as membrane− solvent−solute interactions, make the experimental observations hardly comparable.35,36 However, for the alcohol homologous series, some regulations can be observed. The permeance values of alcohol solvents are in the order MeOH > EtOH > IPA. The highest MeOH permeance could be due to its smallest molar volume and viscosity, while the lowest IPA permeance could be related to its highest molar volume and viscosity. Besides that, the RB rejection displayed an opposite trend with permeance, which could be due to the fact that the different solute−solvent interactions lead to different RB conformations in these solvents.37 In addition, the highest MeOH permeance could lead to a strong dragging effect, which was responsible for the lowest RB rejection.38 A weight loss less than 1% was observed for c(PDA/PI) after an immersion test in DMF for 2 weeks. The long-term membrane separation performance was especially tested in around 48 h using solutions of RB in EtOH and DMF as feeds. It can be seen from Figure 8F that both solvent permeance and RB rejection did not changed after an initial stage of membrane compaction, further demonstrating the stability of the composited membrane. Table 3 listed the separation performance of c(PDA/PI) and some reported OSN membranes. Compared with (catechol/ PEI)/PAN, c(PDA/PI) shows higher solvent permeance with
■
CONCLUSION In this study, composite OSN membranes were prepared via mussel-inspired PDA coating at cPI supports. The pure PDA coating layer can act as a separating layer for OSN, while crosslinking of PDA by HDA could lead to a denser compact layer in the composite membrane, remarkably increasing dye rejection. In an improved fabrication methodology, the cross-linking of both the PDA separating layer and PI support can be achieved in a single step, which significantly simplified the membrane fabrication route. Moreover, the resultant composite membrane exhibits good separation performance toward dyes in a wide range of organic solvents, including challenging polar aprotic and strongly swelling solvents. Furthermore, the composite membrane shows good structural stability in long-term performance testing.
■
ASSOCIATED CONTENT
S Supporting Information *
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acssuschemeng.7b00871. Schematic of NF dead-end stirred-cell filtration system, relative mass gain of membranes, membrane performances of PDA/cPI and cPDA/cPI in DMF and toluene, characteristics of the dyes used for membrane performance testing, and separation performances toward CV, OG, and MO in EtOH of c(PDA/PI) prepared at a PDA coating time of 5 h (PDF)
■
AUTHOR INFORMATION
Corresponding Author
*Tel: +86-451-86413711. Fax: +86-451-86418270. E-mail:
[email protected]. ORCID
Lu Shao: 0000-0002-4161-3861 Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by National Natural Science Foundation of China (21676063, U1462103), State Key Laboratory of Urban Water Resource and Environment (Harbin Institute Technology) (No. 2017DX07), and HIT 5526
DOI: 10.1021/acssuschemeng.7b00871 ACS Sustainable Chem. Eng. 2017, 5, 5520−5528
Research Article
ACS Sustainable Chemistry & Engineering
(19) Mu, W.; Liu, J.; Wang, J.; Mao, H.; Wu, X.; Li, Z.; Li, Y. Bioadhesion-inspired fabrication of robust thin-film composite membranes with tunable solvent permeation properties. RSC Adv. 2016, 6 (106), 103981−103992. (20) Zhao, J.; Su, Y.; He, X.; Zhao, X.; Li, Y.; Zhang, R.; Jiang, Z. Dopamine composite nanofiltration membranes prepared by selfpolymerization and interfacial polymerization. J. Membr. Sci. 2014, 465, 41−48. (21) Pérez-Manríquez, L.; Behzad, A. R.; Peinemann, K.-V. Sub-6 nm Thin Cross-Linked Dopamine Films with High Pressure Stability for Organic Solvent Nanofiltration. Macromol. Mater. Eng. 2016, 301, 1437−1442. (22) Dreyer, D. R.; Miller, D. J.; Freeman, B. D.; Paul, D. R.; Bielawski, C. W. Elucidating the structure of poly(dopamine). Langmuir 2012, 28 (15), 6428−35. (23) Ding, Y.; Weng, L. T.; Yang, M.; Yang, Z.; Lu, X.; Huang, N.; Leng, Y. Insights into the aggregation/deposition and structure of a polydopamine film. Langmuir 2014, 30 (41), 12258−69. (24) Hong, S.; Na, Y. S.; Choi, S.; Song, I. T.; Kim, W. Y.; Lee, H. Non-Covalent Self-Assembly and Covalent Polymerization CoContribute to Polydopamine Formation. Adv. Funct. Mater. 2012, 22 (22), 4711−4717. (25) Xu, Y. C.; Cheng, X. Q.; Long, J.; Shao, L. A novel monoamine modification strategy toward high-performance organic solvent nanofiltration (OSN) membrane for sustainable molecular separations. J. Membr. Sci. 2016, 497, 77−89. (26) Jiang, X.; Li, S.; Shao, L. Pushing CO2-philic membrane performance to the limit by designing semi-interpenetrating networks (SIPN) for sustainable CO2 separations. Energy Environ. Sci. 2017, DOI: 10.1039/C6EE03566C. (27) Yang, X.; Jiang, X.; Huang, Y.; Guo, Z.; Shao, L. Building Nanoporous Metal-Organic Frameworks ″Armor″ on Fibers for HighPerformance Composite Materials. ACS Appl. Mater. Interfaces 2017, 9 (6), 5590−5599. (28) Chen, S.; Cao, Y.; Feng, J. Polydopamine as an efficient and robust platform to functionalize carbon fiber for high-performance polymer composites. ACS Appl. Mater. Interfaces 2014, 6 (1), 349−56. (29) Zhang, Y. Q.; Yang, X. B.; Wang, Z. X.; Long, J.; Shao, L. Designing multifunctional 3D magnetic foam for effective insoluble oil separation and rapid selective dye removal for use in wastewater remediation. J. Mater. Chem. A 2017, 5, 7316−7325. (30) Xu, Y. C.; Wang, Z. X.; Cheng, X. Q.; Xiao, Y. C.; Shao, L. Positively charged nanofiltration membranes via economically musselsubstance-simulated co-deposition for textile wastewater treatment. Chem. Eng. J. 2016, 303, 555−564. (31) Luo, R.; Tang, L.; Zhong, S.; Yang, Z.; Wang, J.; Weng, Y.; Tu, Q.; Jiang, C.; Huang, N. In vitro investigation of enhanced hemocompatibility and endothelial cell proliferation associated with quinone-rich polydopamine coating. ACS Appl. Mater. Interfaces 2013, 5 (5), 1704−14. (32) Zhao, J.; Fang, C.; Zhu, Y.; He, G.; Pan, F.; Jiang, Z.; Zhang, P.; Cao, X.; Wang, B. Manipulating the interfacial interactions of composite membranes via a mussel-inspired approach for enhanced separation selectivity. J. Mater. Chem. A 2015, 3 (39), 19980−19988. (33) Li, Y.; Su, Y.; Li, J.; Zhao, X.; Zhang, R.; Fan, X.; Zhu, J.; Ma, Y.; Liu, Y.; Jiang, Z. Preparation of thin film composite nanofiltration membrane with improved structural stability through the mediation of polydopamine. J. Membr. Sci. 2015, 476, 10−19. (34) Xu, Y. C.; Tang, Y. P.; Liu, L. F.; Guo, Z. H.; Shao, L. Nanocomposite organic solvent nanofiltration membranes by a highlyefficient mussel-inspired co-deposition strategy. J. Membr. Sci. 2017, 526, 32−42. (35) Bhanushali, D.; Kloos, S.; Kurth, C.; Bhattacharyya, D. Performance of solvent resistant membranes for non-aqueous systems: solvent permeation results and modeling. J. Membr. Sci. 2001, 189, 1− 21. (36) Yang, X. J.; Livingston, A. G.; dos Santos, L. F. Experimental observation of nanofiltration with organic solvent. J. Membr. Sci. 2001, 190, 45−55.
Environment and Ecology Innovation Special Funds (HSCJ201619).
■
REFERENCES
(1) Marchetti, P.; Jimenez Solomon, M. F.; Szekely, G.; Livingston, A. G. Molecular separation with organic solvent nanofiltration: a critical review. Chem. Rev. 2014, 114 (21), 10735−806. (2) Sun, S.-P.; Chan, S.-Y.; Xing, W.; Wang, Y.; Chung, T.-S. Facile synthesis of dual-layer organic solvent nanofiltration (OSN) Hollow Fiber Membranes. ACS Sustainable Chem. Eng. 2015, 3, 3019−3023. (3) Lim, S. K.; Setiawan, L.; Bae, T.-H.; Wang, R. Polyamide-imide hollow fiber membranes crosslinked with amine-appended inorganic networks for application in solvent-resistant nanofiltration under low operating pressure. J. Membr. Sci. 2016, 501, 152−160. (4) See Toh, Y. H.; Lim, F. W.; Livingston, A. G. Polymeric membranes for nanofiltration in polar aprotic solvents. J. Membr. Sci. 2007, 301 (1−2), 3−10. (5) Valtcheva, I. B.; Marchetti, P.; Livingston, A. G. Crosslinked polybenzimidazole membranes for organic solvent nanofiltration (OSN): Analysis of crosslinking reaction mechanism and effects of reaction parameters. J. Membr. Sci. 2015, 493, 568−579. (6) Aburabie, J.; Peinemann, K. V. Crosslinked poly(ether block amide) composite membranes for organic solvent nanofiltration applications. J. Membr. Sci. 2017, 523, 264−272. (7) Karan, S.; Jiang, Z.; Livingston, A. G. Sub−10 nm polyamide nanofilms with ultrafast solvent transport for molecular separation. Science 2015, 348, 1347−1351. (8) Jimenez-Solomon, M. F.; Song, Q.; Jelfs, K. E.; Munoz-Ibanez, M.; Livingston, A. G. Polymer nanofilms with enhanced microporosity by interfacial polymerization. Nat. Mater. 2016, 15 (7), 760−7. (9) Dobrak-Van Berlo, A.; Vankelecom, I. F. J.; Van der Bruggen, B. Parameters determining transport mechanisms through unfilled and silicalite filled PDMS-based membranes and dense PI membranes in solvent resistant nanofiltration: Comparison with pervaporation. J. Membr. Sci. 2011, 374 (1−2), 138−149. (10) Lee, H.; Dellatore, S. M.; Miller, W. M.; Messersmith, P. B. Mussel-inspired surface chemistry for multifunctional coatings. Science 2007, 318, 426−430. (11) Liu, Y.; Ai, K.; Lu, L. Polydopamine and its derivative materials: synthesis and promising applications in energy, environmental, and biomedical fields. Chem. Rev. 2014, 114 (9), 5057−115. (12) Gao, H.; Sun, Y.; Zhou, J.; Xu, R.; Duan, H. Mussel-Inspired Synthesis of polydopamine-functionalized graphene hydrogel as reusable adsorbents for water purification. ACS Appl. Mater. Interfaces 2013, 5, 425−432. (13) Zhou, J.; Xiong, Q.; Ma, J.; Ren, J.; Messersmith, P. B.; Chen, P.; Duan, H. Polydopamine-enabled approach toward tailored plasmonic nanogapped nanoparticles: From nanogap engineering to multifunctionality. ACS Nano 2016, 10 (12), 11066−11075. (14) Yang, H.-C.; Luo, J.; Lv, Y.; Shen, P.; Xu, Z.-K. Surface engineering of polymer membranes via mussel-inspired chemistry. J. Membr. Sci. 2015, 483, 42−59. (15) Li, X.-L.; Zhu, L.-P.; Jiang, J.-H.; Yi, Z.; Zhu, B.-K.; Xu, Y.-Y. Hydrophilic nanofiltration membranes with self-polymerized and strongly-adhered polydopamine as separating layer. Chin. J. Polym. Sci. 2012, 30 (2), 152−163. (16) Li, M.; Xu, J.; Chang, C.-Y.; Feng, C.; Zhang, L.; Tang, Y.; Gao, C. Bioinspired fabrication of composite nanofiltration membrane based on the formation of DA/PEI layer followed by cross-linking. J. Membr. Sci. 2014, 459, 62−71. (17) Zhang, R.; Su, Y.; Zhao, X.; Li, Y.; Zhao, J.; Jiang, Z. A novel positively charged composite nanofiltration membrane prepared by bio-inspired adhesion of polydopamine and surface grafting of poly(ethylene imine). J. Membr. Sci. 2014, 470, 9−17. (18) Lv, Y.; Yang, H.-C.; Liang, H.-Q.; Wan, L.-S.; Xu, Z.-K. Nanofiltration membranes via co-deposition of polydopamine/ polyethylenimine followed by cross-linking. J. Membr. Sci. 2015, 476, 50−58. 5527
DOI: 10.1021/acssuschemeng.7b00871 ACS Sustainable Chem. Eng. 2017, 5, 5520−5528
Research Article
ACS Sustainable Chemistry & Engineering (37) Bhanushali, D.; Kloos, S.; Kurth, C.; Bhattacharyya, D. Solute transport in solvent-resistant nanofiltration membranes for nonaqueous systems: experimental results and the role of solute−solvent coupling. J. Membr. Sci. 2002, 208, 343−359. (38) Li, X.; Feyter, S. D.; Vankelecom, I. F. J. Poly(sulfone)/ sulfonated poly(ether ether ketone) blend membranes: Morphology study and application in the filtration of alcohol based feeds. J. Membr. Sci. 2008, 324, 67−75.
5528
DOI: 10.1021/acssuschemeng.7b00871 ACS Sustainable Chem. Eng. 2017, 5, 5520−5528