Antiorganic Fouling and Low-Protein Adhesion on Reverse-Osmosis

Subscriber access provided by UNIVERSITY OF ADELAIDE LIBRARIES

Article

Anti-Organic Fouling and Low-Protein Adhesion on Reverse-Osmosis Membranes Made of Carbon Nanotubes and Polyamide Nanocomposite Yoshihiro Takizawa, Shigeki Inukai, Takumi Araki, Rodolfo Cruz-Silva, Noriko Uemura, Aaron Morelos-Gomez, Josue Ortiz-Medina, Syogo Tejima, Kenji Takeuchi, Takeyuki Kawaguchi, Toru Noguchi, Takuya Hayashi, Mauricio Terrones, and Morinobu Endo ACS Appl. Mater. Interfaces, Just Accepted Manuscript • DOI: 10.1021/acsami.7b06420 • Publication Date (Web): 25 Aug 2017 Downloaded from http://pubs.acs.org on August 29, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Applied Materials & Interfaces is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Applied Materials & Interfaces

Anti-Organic Fouling and Low-Protein Adhesion on Reverse-Osmosis Membranes Made of Carbon Nanotubes and Polyamide Nanocomposite

Yoshihiro Takizawa1†, Shigeki Inukai1†, Takumi Araki1,2, Rodolfo Cruz-Silva1,3, Noriko Uemura2, Aaron Morelos-Gomez1, Josue Ortiz-Medina1, Syogo Tejima1,2, Kenji Takeuchi1,3, Takeyuki Kawaguchi1,3, Toru Noguchi1,3, Takuya Hayashi1,3, Mauricio Terrones3,4 and Morinobu Endo1,3* 1

Global Aqua Innovation Center, Shinshu University; 4-17-1 Wakasato, Nagano 380-8553,

Japan. 2

Research Organization for Information Science & Technology; 2-32-3, Kitashinagawa,

Shinagawa-ku, Tokyo 140-0001. 3

Institute of Carbon Science and Technology, Shinshu University; 4-17-1 Wakasato,

Nagano 380-8553, Japan. 4

Department of Physics, Department of Materials Science and Engineering, Department of

Chemistry, Center for 2-Dimensional and Layered Materials and Center for Atomically Thin Multifunctional Coatings (ATOMIC). The Pennsylvania State University; University Park, Pennsylvania 16802, USA.

Keywords: reverse-osmosis (RO) membrane, antifouling, nanocomposite, carbon nanotube (CNT), polyamide (PA), molecular dynamics (MD) simulations

Abstract We demonstrate efficient antifouling and low protein adhesion of multi-walled carbon nanotubes-polyamide nanocomposite (MWCNT-PA) reverse-osmosis (RO) membranes by combining experimental and theoretical studies using molecular dynamics (MD) simulations. Fluorescein isothiocyanate (FITC)-labeled bovine serum albumin (FITC-BSA) was used for 1

ACS Paragon Plus Environment


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 37

the fouling studies. The fouling was observed in real time by using a crossflow system coupled to a fluorescence microscope. Notably, it was observed that BSA anchoring on the smooth

MWCNT-PA membrane

was

considerably

weaker

than

that

of

other

commercial/laboratory-made plain PA membranes. The permeate flux reduction of the MWCNT-PA nanocomposite membranes by the addition of FITC-BSA was 15% of its original value, whereas that of laboratory-made plain PA and commercial membranes were much larger at 34%50%. Computational MD simulations indicated that the presence of MWCNT in PA results in weaker interactions between the membrane surface and BSA molecule due to the formation of: (i) a stiffer PA structure resulting in lower conformity of the molecular structure against BSA, (ii) a smoother surface morphology and (iii) an increased hydrophilicity involving the formation of an interfacial water layer. These results are important for the design and development of promising anti-organic fouling RO membranes for water treatment.

Introduction Economic growth and climate change have resulted in a scarcity of clean water.1,2 Reverse osmosis (RO) has become the leading process to purify or desalinate water due to its cost efficiency, high-volume water production, and reasonably simple operation.3,4 Some of the

2


Page 3 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


most successful RO membranes are based on aromatic polyamide (PA) due to their excellent water permeation and salt-rejection performance. Prior to RO processes, the water source is usually pretreated by chlorination,5 ozonation5,6, nano/micro filtration,5,6 and/or other processes5-7 in order to minimize fouling. However, organic matter can still reach the membrane and cause serious fouling.8,9 In this context, the formation of a fouling layer increases the flow resistance, and it reduces the permeate flux and the energetic efficiency of the entire process. Therefore, higher transmembrane pressure and frequent washing with chlorinated water are required to maintain stable water permeation, although frequent washing promotes the degradation of the PA membranes; PA is not chemically resistant against chlorinated water.10-12 It is thus important to develop stable and effective antifouling RO membranes. There are only a few studies on antifouling performance involving fouling-release surfaces that work by using either a biomimetic approach or by releasing an antifouling substance, such as silicon oil or copper ions.13,14 However, water purification technology cannot rely on the latter method due to concerns of water contamination. Several approaches to improve the antifouling performance of RO membranes have been reported, usually by increasing the membrane surface hydrophilicity. Examples include membrane modification by hydrophilic polymers,15-19 or coatings consisting of super-hydrophilic

3



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

nanoparticles.20 An alternative approach is to synthesize a smooth PA membrane surface to prevent the foulant deposition.21,22 Indeed, the typical ridge-and-valley morphology of PA membranes21,22 has some positive effects on permeability but also eases the fouling processes due to their relatively high roughness.21 Regarding nanocomposite membranes, the antifouling performance has been improved by modifications with graphene oxide (GO)23-25 and by the incorporation of multi-walled carbon nanotubes (MWCNTs) to PA.3,26 Antifouling performances have been evaluated by various methods that include surface studies using atomic force microscopy (AFM)20,22,27 and scanning electron microscopy (SEM),21 and by carefully monitoring the water permeation.20,22,27 Molecular dynamics (MD) simulations have become a new way to understand the fouling mechanism and the adhesion forces at the atomic level.28-30 We reported that MWCNT-PA nanocomposite membranes showed higher chlorine resistance, better antifouling properties, and similar desalination performances compared to plain PA membranes.31 In the present study, we investigated in detail the antifouling performance of MWCNT-PA membranes against bovine serum albumin (BSA), a known model for studying organic fouling, by in situ fluorescence microscopy coupled to permeation measurements. Our experimental data were supported by MD simulations that

4


Page 4 of 37

Page 5 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


helped to understand the fouling mechanism. An RO nanocomposite membrane prepared by adding MWCNT to PA is highly attractive because this combination improves the antifouling performance and can also decrease the time needed for membrane maintenance, save energy, and extend the membrane’s lifetime. Our findings also enable the engineering of advanced low-protein adhesion nanocomposites that will be useful in the biomedical, pharmaceutical, food processing, and organic waste fields.

Results and Discussion Morphology of reverse-osmosis (RO) membranes Figures 1a-d show SEM images of a nanocomposite RO membrane containing approx. 15.5 wt.% of MWCNT (hereinafter referred to as the ‘MWCNT-PA membrane’), a laboratory-made plain PA membrane, commercial RO membrane-A (CM-A) and commercial RO membrane-B (CM-B). Figures 1e-h show SEM images of the membranes after a 144 hr fouling period. The MWCNT-PA membrane with 15.5 wt.% of MWCNT were used for the substantial salt rejection and permeate flux compared to other MWCNT fractions. This amount of MWCNT was determined by thermogravimetric analysis as reported.31 Different surface features of all membranes originated from the differences in the

5



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

formulation processes (e.g., the presence of MWCNTs), and reaction conditions. The presence of MWCNTs at the organic/inorganic interface during membrane synthesis affects the superficial tension, which is to a great extent responsible for the formation of the characteristic features at the surface of typical PA membranes. Atomic force microscopy (AFM) measurements confirmed the decrease in roughness (Ra) for the MWCNT-PA nanocomposite membrane. The MWCNT-PA nanocomposite (Figure 2a) and laboratory-made plain PA (Figure 2b) membranes show lower Ra values compared to the CM-A and CM-B membranes (Figure 2c,d). As previously mentioned, the presence of MWCNTs within the PA structure contributes to a decrease in the height of the surface features, with a consequent decrease in roughness. In addition, as observed by comparing Figure 2a and 2b, the shape of the features was also considerably affected by the presence of MWCNTs. The ridge-and-valley structures were large and numerous on both the commercial and laboratory-made plain PA membranes with height variations ranging from 160 to 340 nm, whereas those of the MWCNT-PA membrane were smaller with maximum height variations of 140 nm (Figure 2e). From the antifouling performance standpoint, a smooth surface morphology is better,15 even though a ridge-and-valley surface contributes to the high-water permeability of a plain PA membrane. It is important to achieve a balance within the

6


Page 6 of 37

Page 7 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


membrane structure on multiple-scales, such as the surface morphology of the microscopic ridge-and-valley structure and molecular-level smoothness, for salt rejection, permeability, and antifouling nature.

Figure 1. SEM images showing the surface morphology of the RO membrane active layers. (a) MWCNT-PA nanocomposite membrane, (b) laboratory-made plain PA membrane, (c) commercial membrane A (CM-A) and (d) commercial membrane B (CM-B), and (e-h) the same membranes after a 144-hr fouling period where smoother areas in (f-h) correspond to the deposited BSA foulant.

7



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 2. AFM images for (a) the MWCNT-PA nanocomposite membrane and (b) a laboratory-made plain PA membrane, (c) CM-A and (d) CM-B, where the surface morphology and height variations are observed. In (e) the profiles indicated as white dashed lines in the images are plotted with the corresponding roughness parameter from AFM data.

During preconditioning (without BSA, NaCl solution at 0.7 MPa), the water permeate flux and salt permeability gradually decreased, whereas the NaCl rejection increased due to the simultaneous compaction under pressure and hydration of the membranes (see Supporting Figure S-1a,b). The permeate flux and salt rejection were stabilized after 60-150 hr by compaction of the membrane and are shown in Table1. These

8


Page 8 of 37

Page 9 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


values indicate that the MWCNT-PA membrane permeate flux and salt rejection were comparable to other commercial membranes, but remarkably, the change in water permeation during the preconditioning was different from that observed for the PA membranes (see Supporting Figure S-1). This could be related to a decreased proneness to compaction due to the applied pressure, which would be advantageous for maintaining the permeability performance during high pressure operations. In fact, as shown in our previous work,31 at 5.0 MPa of crossflow experiments, the water permeability for MWCNT-PA nanocomposite membranes was higher than that of plain PA membranes, which would be explained by the higher compactability, and thus densification, of the plain PA membrane. In the present study, the fouling process was started by the addition of 100 ppm of fluorescein isothiocyanate-bovine serum albumin (FITC-BSA) to the MWCNT-PA membrane, laboratory-made plain PA membrane, CM-A and CM-B, decreasing the permeate flux to 0.133 m3m2day1 (15% decrease), 0.146 m3m2day1 (42% decrease), 0.161 m3m2day1 (34% decrease) and 0.066 m3m2day1 (50% decrease), respectively (Table 1). CM-A was particularly affected by the decrease in the permeate flux due to the high-flow characteristics of this commercial membrane. The laboratory-made plain PA and the two commercial membranes showed a large decrease in permeability, which is a common property for PA

9



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 37

membranes. It is noteworthy that only the MWCNT-PA membrane differed from the laboratory-made plain PA and commercial membranes, with a slight initial decrease followed by a recovery of the permeate flux within time (see below). Table 1 Salt rejection and permeate flux of the MWCNT-PA nanocomposite membrane, laboratory-made plain PA membrane, CM-A and CM-B before and after initial fouling. Membrane

Salt rejection

3

2

1

Permeate flux (m m d )

(%)

Before fouling

After initial fouling

MWCNT-PA membrane

99.7

0.156

0.133

Laboratory-made plain PA membrane

99.7

0.249

0.146

CM-A

99.6

0.242

0.161

CM-B

99.5

0.132

0.066

Fluorescence microscopy (FM) observation Figures 3a-d show a series of images observed at 0, 72 and 144 hr by FM on each membrane during fouling; all images are available in Supporting Figure S-2. There was no fouling at the beginning of the experiment (0 hr), where the membrane was not exposed to FITC-BSA. This frame captured at 0 hr is also a control to show that there is no fluorescence from the membrane, additives, acrylic cell or the spacer. In the subsequent frames, the areas shown in green are due to the presence of FITC-BSA. For the commercial membrane, CM-A shown in Figure 3a, the BSA foulant preferentially attached to the corner between the membrane and the spacer at first, rather than on the surface of the membrane, although BSA also attached on the surface of the

10


Page 11 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


spacer to a small extent. It is noteworthy that the foulant deposited behind the spacer made of polypropylene was still observable due to its transparency, and the fluorescence was scattered and distorted by the spacer (Supporting Figure S-3a,b). For the CM-B shown in Figure 3b, BSA begins to attach in the middle of the objective area as protein aggregates after 24 hr. When the exposure to BSA foulant solution continued, homogeneous fouling occurred as indicated by the fluorescent green signal from the entire membrane. On the laboratory-made plain PA membrane (Figure 3c), underneath the spacer, the fluorescence intensity (green) increased after 24 hr, similar to the two commercial membranes, thus indicating that the occurrence of organic fouling phenomena is equivalent between the commercial membranes and the laboratory-made plain PA membrane, and we concluded that the fouling observed here is a characteristic of the PA membrane. In contrast, as shown in Figure 3d, the images of the MWCNT-PA membrane are very different from those described above. It is known that MWCNTs act as a fluorescence quencher, but since they are embedded within the PA matrix, the FITC-BSA deposited on the surface could still be observed. The results of the additional experiments shown in Supporting Figure S-4 demonstrate that the quenching effect by MWCNTs was negligible, and the fluorescence intensity of each membrane was comparable. Compared to the plain PA membrane, the fouling on the MWCNT-PA membrane

11



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

area without the spacer was much lower, and the presence of BSA was confirmed only underneath or on the spacer, the fluorescence intensity of which increased over the BSA solution exposure duration. The spacer on the MWCNT-PA membrane was still not visible after 72 hr; however, on the other RO membranes the spacer was already green after 24 hr. These results clearly show that the MWCNT-PA membrane possesses a high level of antifouling properties against BSA. Figure 3e shows the increase in the average fluorescence intensity from the membrane surface over time for the other three RO membranes, whereas the average fluorescence intensity from the MWCNT-PA membrane stayed the same over time. Permeate flux after the addition of BSA The permeate flux was measured during the fouling study. The normalized permeate flux as a function of time (Jr(t)) is given by eq (1):

() = ()/ (1) where J0 is the permeate flux before BSA addition and J(t) is the permeate flux after the BSA addition at time t. As the BSA fouling phenomenon progresses, the value of J(t) decreases. Interestingly, the BSA addition decreased the value of Jr(t) within 10 hr for the MWCNT-PA membrane, laboratory-made plain PA membrane, CM-A and CM-B to 0.85, 0.58, 0.66, and 0.50, respectively (Table 1). A high Jr(t) value (close to the original Jr=1) indicates that the

12


Page 12 of 37

Page 13 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


permeate flux is not greatly affected by the addition of BSA. The plot in Figure 3f shows that the MWCNT-PA membrane had the best initial antifouling behavior among all membranes. The BSA thin layer covered the surface of the membrane at the microscopic level, and this resulted in the closure of water-permeable channels and a large decrease in the Jr(t) values. Since the formation of the thin BSA layer occurs at the molecular level, the initial fouling stage is not observable by FM, but the FM can depict the further progression of foulant deposition. It is also interesting to note that the Jr(t) values of both the laboratory-made plain PA membrane and the commercial membranes barely changed or gradually decreased over time after the BSA addition. The Jr(t) value of the MWCNT-PA membrane increased up to 1 after the initial reduction, which means that the permeate flux recovered its initial value due to the self-cleaning of the membrane. These results thus revealed that the MWCNT-PA membrane exhibits low protein-adhesion properties against BSA deposition. This low protein-adhesion property of the MWCNT-PA membrane suggests that the BSA detachment is easier than that of the laboratory-made plain PA membrane and commercial membranes. A possible explanation is that when a critical amount of BSA is deposited on the surface of the MWCNT-PA membrane, it detaches under the shear stress of the flow turbulence. This is also supported by Figures

13



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

3a-d, where the fouling of the MWCNT-PA membrane is apparently lower, and the fluorescence intensity is less than those of the other RO membranes. The reason for the reduced fouling on the MWCNT-PA membrane is most likely due to the lower interaction between the membrane and the foulants. This phenomenon will be discussed later using MD simulations.

14


Page 14 of 37

Page 15 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 3. Snapshots of the membrane and the spacer observed under a fluorescence microscope as a function of time in a crossflow experiment at 0-144 hr on the (a) CM-A, (b) CM-B, (c) laboratory-made plain PA and (d) MWCNT-PA nanocomposite membranes. Arrows indicate representative BSA foulant adhesion. The images were captured using the same light intensity and exposure time (196 msec). (e) Increase in fluorescence intensity from the RO membranes (the spacer was not included in the measurement) over time. (f) The normalized permeate flux, Jr(t), after the addition of BSA to the four membranes as a function of time.

15



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Effects of MWCNTs in nanocomposite on fouling behavior Figure S-5 shows the images of a MWCNT-PA membrane taken by reflected white light mode and the corresponding FM image. A glass rod was used to spread the dispersion of MWCNT and m-phenylenediamine (MPD) and in some cases, a line with a relatively low concentration of MWCNT was formed due to the stick-slip behavior of the solution (Supporting Figure S-5a). It is also noteworthy that by contrast enhancing the image (Supporting Figure S-5b), the higher fouling regions overlap with the white line regions in the optical microscope image. These results indicate that a higher MWCNT concentration in the membrane prevents the adhesion of BSA; that is, BSA preferentially attaches to the PA-rich areas. These observations would explain why both the laboratory-made plain membrane and the commercial membranes were more prone to fouling compared to the MWCNT-PA membrane.

Computer simulations of membrane structures and the interfacial water layer Supporting Figure S-6 shows (a) an MD snapshot and (b) the charge density map subtracting the PA partial charge and the graphene neutral charge from three layers of graphene-PA (GPA) composite models calculated by the charge equilibration method QEq. A significant charge transfer occurred from the PA molecule around Z = 1.00 nm to the

16


Page 16 of 37

Page 17 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


graphene molecule. A similar phenomenon was confirmed for large dia. MWCNTs (Supporting Figure S-6c,d). We therefore assumed that GPA can be used for the MD simulation as an equivalent model for simulating the MWCNT-PA membrane. The diffusion coefficients of solvated GPA and PA were studied in terms of the PA, the water molecule in the membrane, and the hydrogen bond established between water and PA (see Figure 4a,b). A comparison of Figures 4c and 4d shows that the addition of graphene to PA decreased the diffusion coefficient of PA from 2.88 × 107 to 1.44 × 107 cm2/sec. This means that the PA structures of GPA become stiffer. Especially, this stiffness is higher at the outer most surface region (45 nm). Therefore, in GPA, the average diffusion coefficients of water (2.20 × 105 cm2/sec, Figure 4e) and the hydrogen bond between water and PA (1.43 × 105 cm2/sec, Figure 4g) are slightly lower than those values (2.29 × 105 cm2/sec, Figure 4f and 1.47 × 105 cm2/sec, Figure 4h) of PA. In terms of antifouling performance, a stiffer PA structure at the surface of GPA is advantageous because it lowers the structural conformity with BSA, therefore reducing their interaction.32,33 In addition to a stiffer PA structure of GPA, the important finding is that the hydrogen bonds established between water molecules and PA are localized at the surface region in GPA (45 nm; Figure 4g). This result suggests the presence of a water layer bound to the GPA surface. This water layer on the GPA surface is known as the interfacial

17



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

water layer,34 and a previous report has pointed out the importance of this water layer in the mechanism of the reversible adsorption of BSA on surfaces.35 Indeed, interfacial water keeps the BSA and the membrane surface apart, thus avoiding their interaction. The average number of water molecules per unit area (nm2) was calculated on the GPA model (XY plane = 123.45 Å, 126.64 Å) and the PA model (XY plane = 150.67 Å, 152.83 Å); those were 18.2 nm2 on the GPA surface and 12.1 nm2 on the PA surface. Although the water layer is present in both membrane models, the water layer on the GPA model is well-ordered and more compact than on the plain PA model (Supporting Figure S-7a,b), and the water layer thus hinders the amide bonds from the surface, preventing their possible interaction with the protein. Other research groups reported that this water layer formed on modified hydrophilic surfaces,15-19 which agrees with the interfacial water layer on GPA. This result obtained by simulations is consistent with the experimental results showing that the MWCNT-PA membrane had less BSA fouling compared to the plain PA membrane.

18


Page 18 of 37

Page 19 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 4. Snapshots of water molecules in (a) graphene-PA composite (GPA) and (b) plain PA. Comparison of the diffusion coefficients of the PA (c,d), water (e,f) and hydrogen bond between the water and PA molecules (g,h) in the GPA and plain PA, respectively.

Simulations of foulant deposition Figure 5a and 5b show the height map of the simulated GPA and plain PA models, respectively. GPA exhibited a smoother surface compared to plain PA as shown by a more homogeneous height map at the molecular level. In particular, the plain PA surface has typical ridge-and-valley structures (Figure 5b). We simulated the effect of water flow on the surface of a fouled model membrane by molecular dynamics. In the case of the GPA, BSA is relatively easy to remove with a water flow (1.0 × 106 nm/psec) from the surface (Figure 5c,d) compared to plain PA (Figure 5e,f). The average of total energies (interaction strength) of BSA adsorption from 2 to 12 nsec are 13,840 kcal/mol for the GPA case and 19



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

14,620 kcal/mol for the plain PA case (Figure 5g). The higher interaction strength of plain PA means that BSA attached more strongly on the uneven surface by interacting with polyamide chain branches by hydrogen bonds (Figure 5h), because of the higher conformity of the PA structure. The water permeation hindrance by the BSA attachment on the membrane is weaker in GPA due to the smaller number of hydrogen bonds established between PA in GPA and BSA (Figure 6a,b).

Figure 5. Two-dimension plots of the membrane surface based on aromatic rings distribution to the Z direction of (a) GPA and (b) plain PA. BSA attached on the GPA membrane (c) without water flow and (d) with water flow, and BSA on the plain PA 20


Page 20 of 37

Page 21 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


membrane (e) without water flow and (f) with water flow. 1.0 × 106 nm/psec of water flow is applied for 500 psec. The white line and arrow indicate the displacement of BSA. (g) The total energy (interaction strength) between the attached BSA and membrane for the simulation time. (h) The number of hydrogen bonds between the BSA and membrane for the MD time.

Figure 6. Plots of water density mapping after 12 nsec of molecular dynamics with BSA on the membrane, (a) GPA and (b) plain PA.

The antifouling mechanism and the low protein-adhesion performance of MWCNT-PA membranes In light of the results of our experimental studies and theoretical simulations, it was confirmed that the MWCNT-PA has (i) a stiffer structure, (ii) a smoother surface at the microscopic and molecular levels, and (iii) the formation of an interfacial water layer preventing BSA adhesion on the membrane, thus providing antifouling and low protein-adhesion features. Figure 7a,b shows the schematic models of the MWCNT-PA membrane surface with a small ridge-and-valley structure, and the plain PA membrane surface with a large

21



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ridge-and-valley structure. The addition of BSA to the water source causes the initial fouling, which forms a thin layer of BSA. It causes a sudden drop in the Jr(t) value (Figure 7, steps i and ii). However, the initial drop with the MWCNT-PA membrane was lower due to the reduced interaction between PA and BSA, the smoother surface, and the formation of an interfacial water layer. Therefore, the foulant is easily detached from the surface of the membrane by the shear stress of the water flow compared with a plain PA membrane (Figure 7, step iii).

Figure 7. Schematic model showing the antifouling and low protein-adhesion process of (a) the MWCNT-PA nanocomposite membrane and (b) a plain PA membrane based on the surface morphology and MD simulation. Step i: Formation of the interfacial water layer (MWCNT-PA membrane). Step ii: Formation of a thin layer of BSA. Step iii: The attachment (both the MWCNT-PA membrane and plain PA membrane) and detachment (the MWCNT-PA membrane) of BSA.

22


Page 22 of 37

Page 23 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Conclusion We successfully prepared an MWCNT-PA nanocomposite RO membrane containing a high carbon nanotube content (approx. 15.5 wt.%) with the permeate flux and NaCl rejection properties of 0.16 m3m 2day 1 and 99.7%, respectively for a 10 mmol/l-NaCl feed concentration at 0.7 MPa, using a crossflow experimental setup. The MWCNT-PA membrane exhibited outstanding antifouling performance for the BSA model organic-foulant. MD simulations indicated that the antifouling and outstanding low protein-adhesion properties of the MWCNT-PA membrane against BSA foulant are induced by the stiff PA structure and the interfacial water layer. The smoother surface of the MWCNT-PA membrane compared to a plain PA membrane, from the microscopic to molecular levels, also contributes to the antifouling and low protein-adhesion performance. The present MWCNT-PA nanocomposite RO membrane is promising and could be used for a wide range of applications. In addition to the desalination process, the reported technology has important potential to advance low-protein adhesion on nanocomposite membranes for the biomedical, pharmaceutical, organic waste and food processing fields.

Experimental Section BSA functionalization with fluorescein The protein foulant, BSA (IgG-Free and Protease-Free) purchased from Jackson 23



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Immuno-Research Laboratories (West Grove, PA, USA) was used as an organic foulant model because of its well-known fouling behavior and relatively simple chemistry. BSA was modified by covalently bonding fluorescein to the free amino groups of the lysine residues of the protein backbone in order to make the protein visible by fluorescence microscopy. Typically, 100 mg of FITC was dissolved in 30 ml of dimethyl sulfoxide (DMSO). This FITC solution was then slowly added to the protein solution (1.0 g of BSA was dispersed in 500 ml of a buffer solution containing sodium bicarbonate and sodium carbonate, pH 10.0). Preparation of the RO membranes A MWCNT-PA nanocomposite RO membrane containing approx. 15.5 wt.% of MWCNT, a laboratory-made plain PA membrane and two commercial membranes, i.e., CM-A and CM-B were carefully compared. The MWCNT-PA membrane and laboratory-made plain PA membrane were prepared by interfacial polymerization between m-phenylenediamine (MPD) and trimesoyl chloride (TMC) on the porous polysulfone (PSf) substrate as described.31 For the preparation of the MWCNT-PA membrane, MPD was dissolved in the MWCNT pre-dispersed aqueous solution purchased from KJ Specialty Paper Co., Ltd. (Shizuoka, Japan). Characterization of the RO membranes SEM images of each membrane’s surface were recorded in an SU8000 ultra-high

24


Page 24 of 37

Page 25 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


performance scanning electron microscope (Hitachi, Tokyo) operating at 1.0 kV. For the SEM observations, all of the samples were coated with platinum (approx. 1.0 nm thickness) to avoid surface charging during the observation. Atomic force microscopy (Agilent Technologies AFM 5500, Santa Clara, CA, USA) was performed directly on each membrane, using tapping mode for topography imaging. Crossflow test Each membrane sample (25.0 mm dia.) was loaded in a specially made transparent acrylic crossflow cell (each with 25.0 mm dia. and 0.356 mm height in the cell chamber; Figure 8a), with an effective membrane surface area of 3.46 cm2. Due to the transparency of the experimental cell, the membrane surface was observable during the fouling experiments. A 1.2-mm-thick spacer (GE Water & Process Technologies, Trevose, PA, USA) was placed on the membrane’s surface (Figure 8b) to replicate a typical large-scale RO module. Feed spacers are generally used to keep membranes apart to allow the water flow and to create turbulence that reduces the foulant attachment and salt concentration gradient. The SEM images of the top view (Figure 8c) and side view (Figure 8d) of the spacers observed using a TM3030Plus Miniscope (Hitachi, Tokyo) show that threads in one direction (i) can touch the experimental RO membrane surface, while the threads in the other direction (ii) are fixed on direction-(i) threads without touching the membrane. The crossflow system FTU-1 (Membrane Solutions Technology, Tokyo) was used at 0.7 MPa. A 25



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

schematic illustration of the experimental system is shown in Figure 8e. The water flow rate was 500 ml/min along the surface of the membrane. The water source temperature was kept at 21±1°C. A 10 mmol/l-NaCl aqueous solution was used for the initial compaction process. The NaCl rejection, permeate flux, and salt permeability were measured on the four different membranes during compaction. During the steady-state of the fouling experiments, the permeate flux was also measured. The NaCl rejection was calculated by eq (2):

(%) =

− × 100 (2)

where Cf is the feed concentration and Cp is the permeate concentration of the solution. Both the Cf and Cp values were obtained from the electrical conductivity measurements of the solutions by a portable electrical conductivity meter, ES-71 (Horiba, Kyoto, Japan) with conductivity electrodes #9382-10D and #3574-10C (Horiba, Kyoto, Japan) for source water and permeate water, respectively. Permeate flux (J) was calculated based on eq (3):

=

∆ (3) ∆

where ∆ is the volume of permeated water collected for the permeation time ∆ and A is the effective surface area of the membrane.

26


Page 26 of 37

Page 27 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figure 8. (a) The acrylic cell used for the BSA fouling in situ observation during crossflow. (b) Visible light image of the mesh-like spacer and the RO membrane surface. SEM images of the mesh-like spacer (c) top view and (d) side view of the spacer used for the crossflow test (the numbers on the thread are correlated with each other). Gaps between the membrane and the spacer are visible. (e) The crossflow system used for the RO membrane evaluation and fouling observation. The BSA fouling test First, 100 ppm of FITC-BSA was dissolved in 10 mmol/l-NaCl aqueous solution after the membrane compaction was completed in the crossflow test. The fouling tests were conducted at pH 10. The concentration of BSA and that of the NaCl aqueous solution were similar to that reported.16

27



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The surface of the membrane was observed through the openings of the net-like spacer (Figure 8b) at the center of the acrylic crossflow cell every 24 hr over the 144 hr fouling period. In all of the fluorescent images of the membrane, the water was fed from the left to the right. By fluorescence microscopy, FITC-BSA deposited on the surface of the membrane was observable using blue light (λ=490 nm). The camera used was a CMOS microscope digital eyepiece camera (MC500, Ostec, Guanzhou, China), with the exposure time of 196 msec, equipped with an epifluorescence mode using a stereomicroscope (SMZ18, Nikon, Tokyo), equipped with a Nikon P2-EFLC green filter. The FITC-BSA solution flowing above the membrane inside the acrylic cell prevented the continuous observation of the membrane surface due to the high fluorescence of the FITC-BSA solution that surpasses the fluorescence from the attached FITC-BSA on the membrane surface. Therefore, the water feed was replaced with a 100-ppm non-fluorescent BSA/10 mmol/l-NaCl aqueous solution during image acquisition to avoid BSA detachment. After imaging the membrane surface, the feed water was restored to the 100-ppm FITC-BSA/10 mmol-NaCl aqueous solution for the subsequent fouling experiments. The molecular dynamics (MD) simulation conditions and membrane construction Computer simulations were performed by LAMMPS code using the classical MD simulation

28


Page 28 of 37

Page 29 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


package.36 Our target was to simulate the BSA adsorption on plain PA and three layers of graphene-PA (GPA) composite models. We assumed that a GPA model can behave equivalently to the MWCNT-PA system. This assumption is based on the similar charge-transfer taking place from PA molecules to graphene compared to the MWCNT-PA model. The BSA structure was adopted from the report by Bujacz.37 For the BSA model, a single chain was extracted from the PDB file of the protein crystal downloaded from the Protein Data Bank (PDB ID: 4F5S). The method used to construct the membrane model is similar to that reported in our previous study of MWCNT-PA composite membranes.30 Briefly, a few TMC molecules were used as seeds to create the PA membrane topology. The GPA model was built by placing the TMC molecules scattered over the three-layer graphene surfaces, followed by the polymerization of these seeds. After the polymerization, isothermal-isobaric (NPT) relaxation was achieved by maintaining the total pressure at 1 atm, while the structure was relaxed during 200 psec using 0.5-fsec steps. For the GPA model, we studied the charge transfer from the PA molecules to the three-layer graphene by using a charge equilibration method (QEq).38,39 During the simulations, the SPC/E model40 was adopted for the water molecules. For the plain PA and GPA, we used the General Amber Force Field (GAFF)41 for the

29



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

interaction of the molecules. Plain PA is usually described by GAFF.29,42,43 The atom charges for the plain PA was set with ANTECHAMBER 1.27 and AM1-BCC partial charges.44,45 For the GPA case, the atom charges of the PA and graphene parts were obtained by the QEq method.38 The BSA description was achieved using the CHARMM Force Field.46 The interactions between molecules were calculated by Lenard-Jones (LJ) and Coulomb interactions with particle-particle mesh solver (PPPM).47 For all MD simulations, the time step was set to 1.0 fsec and the trajectory data were saved at 10,000-step intervals for analysis. For the calculations of the charge density maps, the mesh was set to 0.2 Å. The distance criterion for hydrogen bonds was set to ≤4.0 Å. All calculations were performed inside a unit cell with periodic conditions. In order to study the state of the interfacial water layer on the GPA and plain PA, we set 5,200 water molecules on the membrane surface to perform MD simulations with water interactions in a NPT ensemble for 5 nsec. For the fouling MD simulations, the initial states for the membrane models were set to 200 TMC molecules. After polymerization and relaxation, the plain PA models had a total of 65,165 atoms in the XY plane = (150.67 Å, 152.83 Å) and the GPA model had a total of 52,073 atoms in the XY plane = (123.45 Å, 126.64 Å). BSA was constructed with 9,253 atoms and the number of water molecules was set at 95,915 in the plain PA and 56,425 in GPA. The MD simulations with the NPT ensemble were maintained at 1 atm for 12 nsec.

30


Page 30 of 37

Page 31 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Associated content: Supporting information. This material is available free of charge on the ACS Publications website. Permeate flux, salt rejection and salt permeability of RO membranes during compaction process; snapshots of the membrane under a fluorescence microscope as a function of time in a crossflow experiment at 0-144 hr; fluorescent patterns through the transparent spacer; fluorescent patterns on the MWCNT-PA membrane and CM-A; fluorescence-microscope image of the MWCNT-PA membrane; MD snapshots and charge transfer mapping of GPA and MWCNT-PA; MD snapshots of the interfacial water formed on the GPA and plain PA. Author Information Corresponding Author *E-mail:

[email protected].

Tel:

(+81)-26-269-5588.

Fax.

(+81)-26-269-5667. †

These authors contributed equally to this work.

Author Contributions Y.T. performed fouling experiments, fluorescence microscope observation and SEM, and wrote the manuscript. S.I. and T.N. prepared the membranes. T.A. and S.T. performed MD simulations and analysis, and wrote the manuscript. R.C.S. carried out FITC-BSA preparation, image analysis, MD analysis, and advised and discussed the results during project. N.U. performed BSA model construction. A.M.G, J.O.M., K.T., T.K., T.H. and M.T. advised and discussed the results during project. M.E. proposed and supervised the project, and wrote the manuscript. All authors reviewed the approved the manuscript. Acknowledgement This research was funded by the Center of Innovation Program “Global Aqua Innovation Center for Improving Living Standards and Water-sustainability” from Japan Science and Technology Agency, JST. The numerical calculations were carried out on the TSUBAME2.5 supercomputer in the Tokyo Institute of Technology and Earth Simulator in the Japan Marine Science and Technology Center (JAMSTEC). List of Figure captions 31



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 1. SEM images showing the surface morphology of the RO membrane active layers. (a) MWCNT-PA nanocomposite membrane, (b) laboratory-made plain PA membrane, (c) commercial membrane A (CM-A) and (d) commercial membrane B (CM-B), and (e-h) the same membranes after a 144-hr fouling period where smoother areas in (f-h) correspond to the deposited BSA foulant. Figure 2. AFM images for (a) the MWCNT-PA nanocomposite membrane and (b) a laboratory-made plain PA membrane, (c) CM-A and (d) CM-B, where the surface morphology and height variations are observed. In (e) the profiles indicated as white dashed lines in the images are plotted with the corresponding roughness parameter from AFM data. Figure 3. Snapshots of the membrane and the spacer observed under a fluorescence microscope as a function of time in a crossflow experiment at 0-144 hr on the (a) CM-A, (b) CM-B, (c) laboratory-made plain PA and (d) MWCNT-PA nanocomposite membranes. Arrows indicate representative BSA foulant adhesion. The images were captured using the same light intensity and exposure time (196 msec). (e) Increase in fluorescence intensity from the RO membranes (the spacer was not included in the measurement) over time. (f) The normalized permeate flux, Jr(t), after the addition of BSA to the four membranes as a function of time. Figure 4. Snapshots of water molecules in (a) graphene-PA composite (GPA) and (b) plain PA. Comparison of the diffusion coefficients of the PA (c,d), water (e,f) and hydrogen bond between the water and PA molecules (g,h) in the GPA and plain PA, respectively. Figure 5. Two-dimension plots of the membrane surface based on aromatic rings distribution to the Z direction of (a) GPA and (b) plain PA. BSA attached on the GPA membrane (c) without water flow and (d) with water flow, and BSA on the plain PA membrane (e) without water flow and (f) with water flow. 1.0 × 106 nm/psec of water flow is applied for 500 psec. The white line and arrow indicate the displacement of BSA. (g) The total energy between the attached BSA and membrane for the simulation time. (h) The number of hydrogen bonds between the BSA and membrane for the MD time. Figure 6. Plots of water density mapping after 12 nsec of molecular dynamics with BSA on the membrane, (a) GPA and (b) plain PA. Figure 7. Schematic model showing the antifouling and low protein-adhesion process of (a) 32


Page 32 of 37

Page 33 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the MWCNT-PA nanocomposite membrane and (b) a plain PA membrane based on the surface morphology and MD simulation. Step i: Formation of the interfacial water layer (MWCNT-PA membrane). Step ii: Formation of a thin layer of BSA. Step iii: The attachment (both the MWCNT-PA membrane and plain PA membrane) and detachment (the MWCNT-PA membrane) of BSA. Figure 8. (a) The acrylic cell used for the BSA fouling in situ observation during crossflow. (b) Visible light image of the mesh-like spacer and the RO membrane surface. SEM images of the mesh-like spacer (c) top view and (d) side view of the spacer used for the crossflow test (the numbers on the thread are correlated with each other). Gaps between the membrane and the spacer are visible. (e) The crossflow system used for the RO membrane evaluation and fouling observation. List of Table caption Table 1 Salt rejection and permeate flux of the MWCNT-PA nanocomposite membrane, laboratory-made plain PA membrane, CM-A and CM-B before and after initial fouling.

References 1.

Ying, W.; Siebdrath, N.; Uhl, W.; Gitis, V.; Herzberg, M., New Insights on Early

Stages of RO Membranes Fouling During Tertiary Wastewater Desalination. J. Membr. Sci. 2014, 466, 26-35. 2.

Elimelech, M.; Phillip, W. A., The Future of Seawater Desalination: Energy,

Technology, and the Environment. Science 2011, 333 (6043), 712-717. 3.

Zhao, H.; Qiu, S.; Wu, L.; Zhang, L.; Chen, H.; Gao, C., Improving the Performance

of Polyamide Reverse Osmosis Membrane by Incorporation of Modified Multi-Walled Carbon Nanotubes. J. Membr. Sci. 2014, 450, 249-256. 4.

Lee, K. P.; Arnot, T. C.; Mattia, D., A Review of Reverse Osmosis Membrane

Materials for Desalination—Development to Date and Future Potential. J. Membr. Sci. 2011, 370 (1–2), 1-22. 5.

Jamaly, S.; Darwish, N. N.; Ahmed, I.; Hasan, S. W., A Short Review on Reverse

Osmosis Pretreatment Technologies. Desalination 2014, 354, 30-38. 6.

Zhang, J. H.; Northcott, K.; Duke, M.; Scales, P.; Gray, S. R., Influence of

Pre-Treatment Combinations on RO Membrane Fouling. Desalination 2016, 393, 120-126. 7.

Yang, Q.; Liu, Y.; Li, Y., Control of Protein (BSA) Fouling in RO System by

Antiscalants. J. Membr. Sci. 2010, 364 (1-2), 372-379. 33



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

8.

Page 34 of 37

Karabelas, A. J.; Sioutopoulos, D. C., New Insights into Organic Gel Fouling of

Reverse Osmosis Desalination Membranes. Desalination 2015, 368, 114-126. 9.

Khan, M. T.; Hong, P.-Y.; Nada, N.; Croue, J. P., Does Chlorination of Seawater

Reverse Osmosis Membranes Control Biofouling? Water Res. 2015, 78, 84-97. 10.

Glater, J.; Hong, S.-k.; Elimelech, M., The Search for a Chlorine-Resistant Reverse

Osmosis Membrane. Desalination 1994, 95 (3), 325-345. 11.

Powell, J.; Luh, J.; Coronell, O., Bulk Chlorine Uptake by Polyamide Active Layers

of Thin-Film Composite Membranes upon Exposure to Free Chlorine—Kinetics, Mechanisms, and Modeling. J. Environ. Sci. Technol. 2014, 48 (5), 2741-2749. 12.

Avlonitis, S.; Hanbury, W. T.; Hodgkiess, T., Chlorine Degradation of Aromatic

Polyamides. Desalination 1992, 85 (3), 321-334. 13.

Galhenage, T. P.; Hoffman, D.; Silbert, S. D.; Stafslien, S. J.; Daniels, J.; Miljkovic,

T.; Finlay, J. A.; Franco, S. C.; Clare, A. S.; Nedved, B. T.; Hadfield, M. G.; Wendt, D. E.; Waltz, G.; Brewer, L.; Teo, S. L. M.; Lim, C.-S.; Webster, D. C., Fouling-Release Performance of Silicone Oil-Modified Siloxane-Polyurethane Coatings. ACS Appl. Mater. Interfaces 2016, 8 (42), 29025-29036. 14.

Howell, C.; Vu, T. L.; Lin, J. J.; Kolle, S.; Juthani, N.; Watson, E.; Weaver, J. C.;

Alvarenga, J.; Aizenberg, J., Self-Replenishing Vascularized Fouling-Release Surfaces. ACS Appl. Mater. Interfaces 2014, 6 (15), 13299-13307. 15.

Toru Ishigami, K. A., Akihiro Fujii, Yoshikage Ohmukai, Eiji Kamio, Tatsuo

Maruyama, Hideto Matsuyama, Fouling Reduction of Reverse Osmosis Membrane by Surface Modification via Layer-by-Layer Assembly. Sep. Purif. Technol. 2012, 99, 1-7. 16.

Azari, S.; Zou, L.; Cornelissen, E., Assessing the Effect of Surface Modification of

Polyamide RO Membrane by l-DOPA on the Short Range Physiochemical Interactions with Biopolymer Fouling on the Membrane. Colloids and Surf., B 2014, 120, 222-228. 17.

Kang, G.; Liu, M.; Lin, B.; Cao, Y.; Yuan, Q., A Novel Method of Surface

Modification

on Thin-Film

Composite Reverse Osmosis

Membrane by Grafting

Poly(ethylene glycol). Polymer 2007, 48 (5), 1165-1170. 18.

Dong, B.; Jiang, H.; Manolache, S.; Wong, A. C. L.; Denes, F. S., Plasma-Mediated

Grafting of Poly(ethylene glycol) on Polyamide and Polyester Surfaces and Evaluation of Antifouling Ability of Modified Substrates. Langmuir 2007, 23 (13), 7306-7313. 19.

Liu, M.; Chen, Q.; Wang, L.; Yu, S.; Gao, C., Improving Fouling Resistance and

Chlorine Stability of Aromatic Polyamide Thin-Film Composite RO Membrane by Surface Grafting of Polyvinyl alcohol (PVA). Desalination 2015, 367, 11-20. 20.

Tiraferri, A.; Kang, Y.; Giannelis, E. P.; Elimelech, M., Superhydrophilic Thin-Film

Composite Forward Osmosis Membranes for Organic Fouling Control: Fouling Behavior and 34


Page 35 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Antifouling Mechanisms. J. Environ. Sci. Technol. 2012, 46 (20), 11135-11144. 21.

Elimelech, M.; Zhu, X.; Childress, A. E.; Hong, S., Role of Membrane Surface

Morphology in Colloidal Fouling of Cellulose Acetate and Composite Aromatic Polyamide Reverse Osmosis Membranes. J. Membr. Sci. 1997, 127 (1), 101-109. 22.

Lu, X.; Arias Chavez, L. H.; Romero-Vargas Castrillón, S.; Ma, J.; Elimelech, M.,

Influence of Active Layer and Support Layer Surface Structures on Organic Fouling Propensity of Thin-Film Composite Forward Osmosis Membranes. J. Environ. Sci. Technol. 2015, 49 (3), 1436-1444. 23.

Ray, J. R.; Tadepalli, S.; Nergiz, S. Z.; Liu, K.-K.; You, L.; Tang, Y.; Singamaneni,

S.; Jun, Y.-S., Hydrophilic, Bactericidal Nanoheater-Enabled Reverse Osmosis Membranes to Improve Fouling Resistance. ACS Appl. Mater. Interfaces 2015, 7 (21), 11117-11126. 24.

Choi, W.; Choi, J.; Bang, J.; Lee, J.-H., Layer-by-Layer Assembly of Graphene

Oxide Nanosheets on Polyamide Membranes for Durable Reverse-Osmosis Applications. ACS Appl. Mater. Interfaces 2013, 5 (23), 12510-12519. 25.

Perreault, F.; Tousley, M. E.; Elimelech, M., Thin-Film Composite Polyamide

Membranes Functionalized with Biocidal Graphene Oxide Nanosheets. J. Environ. Sci. Technol. Lett. 2014, 1 (1), 71-76. 26.

Chan, W.-F.; Marand, E.; Martin, S. M., Novel Zwitterion Functionalized Carbon

Nanotube Nanocomposite Membranes for Improved RO Performance and Surface Anti-Biofouling Resistance. J. Membr. Sci. 2016, 509, 125-137. 27.

Mo, Y.; Tiraferri, A.; Yip, N. Y.; Adout, A.; Huang, X.; Elimelech, M., Improved

Antifouling Properties of Polyamide Nanofiltration Membranes by Reducing the Density of Surface Carboxyl Groups. J. Environ. Sci. Technol. 2012, 46 (24), 13253-13261. 28.

Xiang, Y.; Liu, Y.; Mi, B.; Leng, Y., Molecular Dynamics Simulations of Polyamide

Membrane, Calcium Alginate Gel, and Their Interactions in Aqueous Solution. Langmuir 2014, 30 (30), 9098-9106. 29.

Harder, E.; Walters, D. E.; Bodnar, Y. D.; Faibish, R. S.; Roux, B., Molecular

Dynamics Study of a Polymeric Reverse Osmosis Membrane. J. Phys. Chem. B 2009, 113 (30), 10177-10182. 30.

Araki, T.; Cruz-Silva, R.; Tejima, S.; Takeuchi, K.; Hayashi, T.; Inukai, S.; Noguchi,

T.; Tanioka, A.; Kawaguchi, T.; Terrones, M.; Endo, M., Molecular Dynamics Study of Carbon Nanotubes/Polyamide Reverse Osmosis Membranes: Polymerization, Structure, and Hydration. ACS Appl. Mater. Interfaces 2015, 7 (44), 24566-24575. 31.

Inukai, S.; Cruz-Silva, R.; Ortiz-Medina, J.; Morelos-Gomez, A.; Takeuchi, K.;

Hayashi, T.; Tanioka, A.; Araki, T.; Tejima, S.; Noguchi, T.; Terrones, M.; Endo, M., High-Performance Multi-Functional Reverse Osmosis Membranes Obtained by Carbon 35



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

NanotubePolyamide Nanocomposite. Sci. Rep. 2015, 5, 13562. 32.

Berglin, M.; Pinori, E.; Sellborn, A.; Andersson, M.; Hulander, M.; Elwing, H.,

Fibrinogen Adsorption and Conformational Change on Model Polymers: Novel Aspects of Mutual Molecular Rearrangement. Langmuir 2009, 25 (10), 5602-5608. 33.

Vyner, M. C.; Liu, L.; Sheardown, H. D.; Amsden, B. G., The Effect of Elastomer

Chain Flexibility on Protein Adsorption. Biomaterials 2013, 34 (37), 9287-9294. 34.

Sendner, C.; Horinek, D.; Bocquet, L.; Netz, R. R., Interfacial Water at Hydrophobic

and Hydrophilic Surfaces: Slip, Viscosity, and Diffusion. Langmuir 2009, 25 (18), 10768-10781. 35.

Kubiak-Ossowska, K.; Tokarczyk, K.; Jachimska, B.; Mulheran, P. A., Bovine

Serum Albumin Adsorption at a Silica Surface Explored by Simulation and Experiment. J. Phys. Chem. B 2017, 121 (16), 3975-3986. 36.

Valiev, M.; Bylaska, E. J.; Govind, N.; Kowalski, K.; Straatsma, T. P.; Van Dam, H. J.

J.; Wang, D.; Nieplocha, J.; Apra, E.; Windus, T. L.; de Jong, W. A., NWChem: A Comprehensive and Scalable Open-Source Solution for Large Scale Molecular Simulations. Comput. Phys. Commun. 2010, 181 (9), 1477-1489. 37.

Bujacz, A., Structures of Bovine, Equine and Leporine Serum Albumin. Acta

Crystallogr., Sect. D 2012, 68 (10), 1278-1289. 38.

Rappe, A. K.; Goddard, W. A., Charge Equilibration for Molecular Dynamics

Simulations. J. Phys. Chem. 1991, 95 (8), 3358-3363. 39.

Aktulga, H. M.; Fogarty, J. C.; Pandit, S. A.; Grama, A. Y., Parallel Reactive

Molecular Dynamics: Numerical Methods and Algorithmic Techniques. Parallel Comput. 2012, 38 (4–5), 245-259. 40.

Berendsen, H. J. C.; Grigera, J. R.; Straatsma, T. P., The Missing Term in Effective

Pair Potentials. J. Physic. Chem. 1987, 91 (24), 6269-6271. 41.

Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K. M.; Ferguson, D. M.;

Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.; Kollman, P. A., A Second Generation Force Field for the Simulation of Proteins, Nucleic Acids, and Organic Molecules J. Am. Chem. Soc. 1995, 117, 5179−5197. J. Am. Chem. Soc. 1996, 118 (9), 2309-2309. 42.

Ding, M.; Szymczyk, A.; Ghoufi, A., On the Structure and Rejection of Ions by a

Polyamide Membrane in Pressure-Driven Molecular Dynamics Simulations. Desalination 2015, 368, 76-80. 43.

Shen, M.; Keten, S.; Lueptow, R. M., Dynamics of Water and Solute Transport in

Polymeric Reverse Osmosis Membranes via Molecular Dynamics Simulations. J. Membr. Sci. 2016, 506, 95-108. 44.

Wang, J.; Wang, W.; Kollman, P. A.; Case, D. A., Automatic Atom Type and Bond 36


Page 36 of 37

Page 37 of 37

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Type Perception in Molecular Mechanical Calculations. J. Mol. Graph. Model. 2006, 25 (2), 247-260. 45.

Jakalian, A.; Jack, D. B.; Bayly, C. I., Fast, Efficient Generation of High-Quality

Atomic Charges. AM1-BCC Model: II. Parameterization and Validation. J. Comput. Chem. 2002, 23 (16), 1623-1641. 46.

MacKerell, A. D.; Banavali, N.; Foloppe, N., Development and Current Status of the

CHARMM Force Field for Nucleic Acids. Biopolymers 2000, 56 (4), 257-265. 47.

E.L. Pollock, J. G., Comments on P3M, FMM, and the Ewald Method for Large

Periodic Coulombic Systems. Comput. Phys. Commun. 1996, 95 (2-3), 93-110.

For Table of Contents Only

37


Antiorganic Fouling and Low-Protein Adhesion on Reverse-Osmosis

Recommend Documents