Accepted Manuscript Covalent surface entanglement of polyvinylidene fluoride membranes with carbon nanotubes Samer Al-Gharabli, Joanna Kujawa, Musthafa O. Mavukkandy, Taofeeqah A. Agbaje, Eyad M. Hamad, Hassan A. Arafat PII: DOI: Reference:
S0014-3057(17)31904-3 https://doi.org/10.1016/j.eurpolymj.2018.01.027 EPJ 8260
To appear in:
European Polymer Journal
Received Date: Revised Date: Accepted Date:
26 October 2017 27 December 2017 26 January 2018
Please cite this article as: Al-Gharabli, S., Kujawa, J., Mavukkandy, M.O., Agbaje, T.A., Hamad, E.M., Arafat, H.A., Covalent surface entanglement of polyvinylidene fluoride membranes with carbon nanotubes, European Polymer Journal (2018), doi: https://doi.org/10.1016/j.eurpolymj.2018.01.027
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Covalent surface entanglement of polyvinylidene fluoride membranes with carbon nanotubes
Samer Al-Gharabli1,2, Joanna Kujawa1,3, Musthafa O. Mavukkandy1, Taofeeqah A. Agbaje1, Eyad M. Hamad4, Hassan A. Arafat 1*
Department of Chemical Engineering, Masdar Institute, Khalifa University of Science and Technology, PO Box 54224, Abu Dhabi, United Arab Emirates
Pharmaceutical and Chemical Engineering Department, German-Jordanian University, Amman, Jordan
Biomedical Engineering Department, German-Jordanian University, Amman, Jordan
* Corresponding author. E-mail: [email protected]
Abstract Polyvinylidene fluoride (PVDF) membranes play a key role in several industrial applications. Tuning the properties of these membranes enables a sharp selectivity for the target application. The goal of this work is to chemically bind carbon nanotubes (CNT) onto the surface of PVDF membranes. Such membranes can open the door for a variety of new applications. To achieve this binding, the PVDF membrane was first functionalized with a labile –OH group. This functionalization was accomplished using a Piranha-reaction approach. The –OH group was then substituted with an amine group. Single walled, highly conductive carbon nanotubes were then covalently attached to the amine functionalized PVDF membrane. The properties of the membrane were evaluated and characterized using a wide range of tools. Extensive physicochemical, structural and surface characterizations of the prepared membranes were conducted utilizing goniometric, microscopic, surface and spectroscopic methods. The conductive CNTs were observed to alter the electrical properties of the polymer creating a conductive polymer membrane.
Keywords: piranha reagent; covalent functionalization; carbon nanotubes; polyvinylidenefluoride; conductive membrane
1. Introduction Covalent binding of versatile nanomaterials, like carbon nanotubes (CNT), with polyvinylidene fluoride (PVDF) membranes enables application-specific functionalization of these membranes. This opens the door for new interesting applications of PVDF membranes, for example in piezoresistive actuators, P-N junctions, and biosensors, to name a few. Being an inert material, a pre-activation of PVDF is necessary to achieve the desired covalent binding with CNT. This is the approach attempted in this study. Membrane-based separation processes are extensively used in industry including food processing, water and wastewater treatment, gas separation and pharmaceutical industries. Polyvinylidenefluoride (PVDF) is a semi-crystalline polymer with exceptional chemical resistance, high mechanical, UV and thermal stability and easy processability. Hence, it is widely used in membrane applications such as micro- and ultrafiltration and membrane distillation (MD) [1-4]. PVDF membranes could potentially be used in various advanced applications such as drug delivery, water softening, bio-separation, charge storage, ferroelectric memories, sensors, and protective coatings [5-7]. Membrane fine-tuning is essential to create highly selective, ultra-permeable and/or anti-fouling smart membranes. Such fine-tuning involves various physical [8, 9] and/or chemical [10, 11] methods. For instance, PVDF membranes suffer from fouling in water/wastewater filtration and wetting in MD. Blending hydrophilic nanomaterials into the membranes has been adopted as an effective way to tackle membrane fouling. On the other hand, blending of various superhydrophobic nanomaterials minimizes the wetting in MD. PVDF materials incorporating nano-filler additives such as metal nanoparticles were also used for a variety of applications.
hydrophilicity/hydrophobicity, permeability, conductivity, mechanical strength and/or fouling resistance . 3
Carbon-based nanomaterials, such as CNT, carbon black, carbon nanofibers (CNF) and graphene oxide have been explored for applications in fouling resistant, ultra-permeable membranes [13-15]. CNT is gaining popularity due to its attributes including good dispersion in polymeric matrices, which stems from its strong affinity with polymers [16-19]. Shaffer and Alan  presented the changes in the properties of poly(vinyl alcohol) (PVA) as a result of CNT loading. A retardation effect at the onset of PVA thermal decomposition and an increase of elastic modulus were observed for multi-walled CNT (MWCNT) loaded PVA. Zonder et al  reported that by blending MWCNT in polyethylene, strong reinforcement effect was developed in polyamide/high density polyethylene. In the case of multi-walled CNT doped PVDF membranes, an increase of Young modulus values as well as elongation at break were noticed . The synergistic effect of CNT and carbon black (CB) in CNT/CB/PVDF blend material was examined by Li et al , who observed a significant improvement of PVDF conductivity. PVDF/CNT fibers made by electrospinning were presented by Baji et al.  where the presence of CNT was found to enhance the dielectric dfb
’ stiffness and tensile strength.
There are limitations to blending-based modifications of polymeric membranes. For instance, in most cases, blending of nanoparticles changes the chemistry of membrane formation. Another issue is nanoparticles encapsulation inside the polymer matrix limiting accessibility to their active surface area. In other cases, those nanoparticles may leach out from the polymer matrix and contaminate the permeate. Therefore, a post-fabrication functionalization of the membrane might be a better choice to avoid the said issues and for improved control of spatial positioning of nanomaterials within the membrane. Still, most of the literature on CNT d
g f PVDF
d ’ f d
reporting a modification involving covalent binding of CNT and PVDF. This is due to the inert nature of PVDF, which necessitates a surface activation prior to covalent binding. One
method for grafting PVDF surface is via hydroxyl groups (-OH) generation through treatment with alkaline media such as NaOH, KOH or LiOH [11, 25-29]. However, it has been reported that PVDF can be degraded during such alkaline treatment [11, 25-29]. Another method used for surface activation involves a piranha treatment [30, 31]. The piranha solution is a mix of 3 volumes of concentrated sulfuric acid and 1 volume of hydrogen peroxide. This mixture is usually dangerous when hot and hence its application is limited. However, the authors have developed an easy, safe and rapid experimental protocol for the activation of PVDF membranes enabling their further surface modification/grafting . In this study, PVDF membranes were activated utilizing the said piranha activation method, where hydroxyl groups were generated onto the PVDF membrane surface allowing other functional groups and nanomaterials to be docked on the membrane. We demonstrated this by covalently attaching CNT onto the PVDF membrane surface. To the best of our knowledge, this is the first study demonstrating the covalent binding of CNTs with PVDF membranes.
2. Experimental 2.1 Materials C
0.2 μ )
was supplied by Thermo Scientific (USA). The chemicals: hydrogen peroxide (30%), sulfuric acid (98%), and (3-aminopropyl)triethoxysilane, methanol, toluene, glycerol, were purchased from Sigma-Aldrich (USA) and were used as received without further purification. Single walled carbon nanotubes (SWCNT) (outer diameter: 1 – 2 nm, length: 5 – 30 µm), functionalized with carboxyl group (2.75wt% carboxylic groups) was purchased form Mknano, Canada. The activation solution of Piranha was prepared as per our earlier report  using deionized (DI) water (resistivity = 15MΩ·
2.2. Membrane modification The modification process leading to the CNT functionalized PVDF membrane is schematically illustrated in Figure 1. It involves the activation of a commercial PVDF b
d “PVDF”) w
w d b
functionalization with (3-aminopropyl)triethoxysilane to form amine groups on the membrane surface. These amine groups were utilized in the covalent attachment of CNT onto the membrane surface. The effectiveness of the Piranha method in activating the PVDF membrane and the subsequent functionalization of this surface with amine groups was thoroughly assessed during the initial phase of this research and is comprehensively discussed elsewhere .
Fig. 1. Schematic presentation of the synthetic steps leads to CNT functionalized membrane
In order to prepare the piranha solution, 7.5 mL of H 2SO4 (98%) were first added to 40 ml of DI water. A 2.5 mL of 30% aqueous hydrogen peroxide solution was then carefully added to the above solution. A commercial PVDF membrane with 0.2 µm nominal pore size (Thermo Scientific, USA) was cut into 2 x 8 cm2 and pre-wetted with methanol until it became transparent. The wetted 6
membrane was carefully transferred to freshly prepared 20% Piranha solution and stirred for 10 minutes at room temperature. To quench the reaction, the membrane was immersed in a DI water bath for 5 minutes then washed thoroughly with DI water to remove any traces of the piranha solution. Afterwards, the membrane was washed 5 times with methanol, and dried at 70 °C for 12 hours. This will henceforth be referred to as the hydroxyl functionalized membrane (P-OH). The P-OH membrane was placed in 50 mL of 0.1 M (3-aminopropyl)triethoxysilane (T-NH2) in toluene at room temperature under inert atmosphere to produce amine functionalized membrane (referred to as P-NH2). The membrane was stirred for 3 hours, washed 5 times with toluene, methanol, and dichloromethane, respectively, to remove any unreacted silane or other reagents. The membrane was then dried at 70 °C for 12 hours to obtain P-NH2. The next step was to covalently furnish the P-NH2 membrane with CNT. A suspension of CNT (10 mg) in 50 ml of dichloromethane (DCM) was sonicated for 10 minutes at room ′ ′-tetramethyluronium tetrafluoroborate
temperature. 2 mg of O-(Benzotriazol-1-yl)-
(TBTU) and 1 µL of N,N-diisopropylethylamineamine (DIPEA) were added to the above suspension to form the active ester. P-NH2 was then immersed in the reaction mixture and sonicated for 3 hours. The membrane was then removed from the reaction mixture and washed under sonication with DCM, methanol, DI water and methanol, respectively, for 5 times each. The membrane was then dried at 70 °C for 3 hours to produce the CNTfunctionalized PVDF membrane (referred to as P-CNT).
2.3. Membrane characterization Pristine and modified membranes were characterized for their apparent contact angle (CA), f
(γ cr) and surface free energy (SFE) using the static and dynamic
contact angle measurements. During the CA measurements, 3 μl of liquids were used.
Measurements were done at room temperature using a goniometer (Krüss Easy Drop Analyzer, Germany) equipped with Drop Analyzer software. The standard error in measurements was about ±0.5o. The following solvents were utilized as the wetting liquids: w
d f x
d (γ=37.1 -1
), n-d d
). From the CA measurements using these liquids, the critical
surface tension values of the pristine and modified samples were calculated based on the Zisman method . For the SFE determination, the Owens-Wendt theory was implemented . Microscopic imaging of all investigated samples was conducted using an FE-SEM device (FEI Nova NanoSEM 650). Energy-dispersive X-ray spectroscopy (EDAX) mode was used for assessment of samples composition. Prior to SEM analysis, all PVDF membranes were sputtered with a 100 Å layer of metals (gold–palladium) in order to improve the conductivity and quality of the collected data. The roughness parameters and adhesion force of the samples were measured via atomic force microscopy (AFM), using a Nano-Observer AFM Microscope (CSInstruments, France). The root mean square (RMS) roughness was chosen as a representative parameter and was measured using AFM tip scanning mode. The RMS values were estimated using the built-in mathematical algorithm in NanoScope Analysis Software (1.40, Build R3Sr5.96909, 2013 Bruker Corporation). The scan size of the sample was 5 x 5 µm2. All membranes were measured at least five times, and an average value was calculated and presented (accuracy ± 4%). Adhesive force values were achieved from 25 measurements applying silicon nitride (Si3N4) probes NP-1 in the contact-mode (probes with a spring constant of 0.58 Nm-1, sensitive deflection of 36.1 nm V -1, tip radius of 40 nm, and tip half angle of 35° were
provided by Bruker). The ultimate loading force was 50–70 nN and the tip velocity was 7.88 μ
. All tests were done at ambient temperature.
High resolution transmission electron microscopy (HR-TEM) (Tecnai TEM, G2 F20 X-Twin, FEI Europe) analysis was performed to assess the presence of CNT attached to the PVDF membrane structure, applying an accelerating voltage of 200 kV. Samples were first cut using an ultramicrotome (EM UC7, Leica Microsystems, USA) and subsequently placed on a copper grid. Raman spectra were collected for all the membranes using a Witec Alpha 300 RAS apparatus, Germany. The following laser lines were chosen: 532, and 633 nm. All spectra were gathered with integration time of 2-30 sec and accumulation up to 250. Membrane pore size distribution (PSD) and air permeability were characterized using a capillary flow porometer (CFP, Porous Materials Inc., USA). For PSD, GalwickTM (surface tension 15.9 mN m-1) was used as the wetting liquid. Details of the tests can be found in the literature [35, 36]. The electrical conductivity of the modified membrane was assessed by generating an I-V curve of a membrane sheet using a Lakeshore HMS-7607 Hall Effect Measurement System employing four-point probe method. Sheet resistivity was calculated by placing four tungsten probes on the corners of a 1 cm2 membrane sample (van Der Pauw method). The samples were placed in a 10 KG magnetic field, followed by applying current (I) from -100 nA to +100 nA with 20 nA current step across the two opposite probes, while measuring the voltage on the other probes sequentially and rotationally.
3. Results and discussion 3.1. Functionalization and CNT binding of PVDF membrane
Raman spectroscopy is a useful technique in determining the chirality, diameter distribution, architecture and purity of CNT [16, 17, 23]. The most distinguished Raman attributes of CNT spectrum are: RBM – radial breathing mode (in the range of 300-75 cm-1); D (disordered – in the range of 1360 – 1330 cm-1), G (graphite – in the range of 1630 – 1470 cm-1)
(second-order Raman scattering from D-band variations – in the range of 2900 – 2500 cm-1) modes. Raman spectra of PVDF and P-CNT are presented in Figure 2. It can be seen that characteristic peaks for PVDF (3027 cm-1 strong peak of CH2 asymmetric stretching mode; 2985 cm-1 very strong peak of CH2 symmetric stretching mode; 1431 cm-1 medium peak of deformation CH2; 1200 cm-1 weak peak of CF2 asymmetric stretching mode; 1056 cm-1 weak peak of CF2 symmetric stretching mode; 875 cm-1 medium peak of CC symmetric stretching mode; 800 cm-1 strong peak of CH2 rocking deformation mode and 410 cm-1 weak peak of CF2 wagging mode) disappeared in the case of P-CNT while new peaks were observed (Figure 2). A very intense peak appeared at 1588 cm-1, which is associated with the C-C stretching bonds between the two dissimilar carbon atoms in the graphite plane (Figure 2. C1). G-band [37, 38] with its two components, G+ and G-, are related to the vibrations along the length of the nanotube axis (longitudinal optical phonon mode) and vibrations along the circumferential direction of the nanotube (transverse optical phonon mode), respectively . Two intense peaks related to the RBM (Figure 2. C2) were recorded at 219 and 152 cm-1 which are attributed to CNT and cannot be noticed in the graphite plate (Figure 2. B). The RBM are associated with coherent movement of carbon atoms in the radial direction . The RBM possessed two components: one shoulder shift in the direction of higher frequencies, corresponding to CNTs in a bundle environment, and another intense peak associated with isolated nanotubes. In the case of SWCNT, frequency of Raman shift of RBM is directly proportional to the CNT dimension. The dimension of isolated (Eq.1) and
bundled (Eq. 2) SWCNT can be determined from the RBM Raman shift according to the following formulae:
ω 10 (234/d)
Calculated diameters of isolated and bundled CNT were equal to 1.63 and 1.12 nm, respectively. The G band for pure CNT as well as for functionalized membranes were characterized by sharp Lorenz profiles (FWHMCNT = 37cm-1 and FWHMP-CNT = 30cm-1). The achieved value of FWHM for G-bands reflects that SWCNT were mostly in bundled environment. The FWHM for isolated form of SWCNT are reported to be in the range of 6 – 15 cm-1 [37, 38, 40].
Fig. 2. Raman spectra of (A): neat PVDF, (B): PVDF with covalently bonded CNTs (PCNT), (C): pure CNT, (C.1): G bands consisting G+ (vibrations along the nanotube axis -
longitudinal optical phonon mode) and G - (vibrations along the circumferential direction of the nanotube - transverse optical phonon mode), (C.2): RBM band.
On the other hand, D-band is a longitudinal optical phonon related to the disordered mode due to the defect, where energy is needed to elastically scatter the CNT. This mode is in linear correlation with excitation energy of the laser. It is present in all types of carbon allotropes, comprising sp2 and sp3 amorphous carbon. In the case of CNT, this mode is initiated from the 1st order scattering process of sp2 carbons by the presence of in-plane substitutional heteroatoms, grain boundaries, vacancies or other defects, and by finite-size effects. The aforementioned features can contribute to decreasing the symmetry of the crystal of the quasi-infinite lattice. D-band is generally applied for the analysis of defects in CNT structure of the walls as well as for the detection of the presence of amorphous carbon. The intensities of D and G bands may be compared for the qualitative analysis of the samples. Usually, high-quality CNT with less defects and amorphous carbon are characterized by a D/G ratio lower than 2 . In the present work, D/G ratio increased from 0.22 for pure CNT to 1.4 for the P-CNT membrane. The increased intensity of D-band (Raman shift = 1325 cm-1) for modified membrane (P-CNT), as compared to pure CNT (Raman shift = 1320 cm-1), may be due to the generation of additional active sites, consumption of the graphene sidewall around the vacancy, generation of vacancies in the graphite plane as well as the transportation of carbon atom hybridization from sp2 to sp3 . The mentioned changes indicate the formation of stable covalent bonds between the PVDF membrane and the CNT (Figure 2. C1). Moreover, intense peaks are located at higher frequency region, 2626 cm-1 for P-CNT and 2622 cm-1, for pure CNT (Figure 2. B&C). This b
graphite and CNT and can be observed in absolutely defect-free CNT sample in which the D band will be absent [39, 41]. The efficiency of PVDF surface functionalization with CNT was also assessed using HRTEM analysis (Figure 3). TEM image of pure PVDF is presented in Figure 3. A1. The presence of only amorphous phase of polymer was proven by the concentric rings on the selected area of electron diffraction pattern (SAED) (Figure 3. A.2). In the case of P-CNT membrane, higher degree of crystal orientation of the structure was observed (Figure 3. B.1). This is due to the introduction of CNT and its partial alignment on the surface of the polymer membrane. This proves that CNT was efficiently anchored to the PVDF polymeric membrane. Changes in SAED pattern indicate that there was an increase of crystallinity in the P-CNT sample (Figure 3. B.2&B.3). Thus, different orientation and directions were found (Figure 3. B.1a). However, the distance between parallel fringes was in the range of 0.26 nm to 0.31 nm. This small size of the SWCNT, make it impossible to see the isolated CNT (outer diameter of CNT: 1 – 2 nm). A similar case was reported by Huang et al. , indicating a difficulty in observing SWCNT via TEM due to the small sizes of CNT. However, they highlighted that higher degree of crystal orientation of SWCNT was observed in comparison to MWCNT in the polymeric matrix of PVDF nanofibers. This implies that SWCNT could be better aligned than MWCNT in the nanofibers structure .
Fig. 3. TEM images of pure PVDF (A.1) and SAED pattern for PVDF membrane (A.2). PCNT (B.1 and B.1a – the magnified selected area) and SAED pattern for P-CNT sample (B.2 and B.3).
3.2. Effect of CNT docking on membrane structure Membrane activation and CNT-functionalization processes influenced the structural and physicochemical properties of the membranes. SEM micrographs of PVDF, P-NH2 and PCNT are presented in Figure 4. They illustrate that the P-NH2 membranes (Figure 4. B, D) possessed more open structure and highly heterogeneous morphology compared to the PVDF membrane (Figure 4. A). This open structure of P-NH2 membrane is ascribed to the surface activation by piranha treatment. Small spheres with dimension of ca. 1 µm, (Figure 4. B) were noticed on the surface of P-NH2. The micrograph of pure SWCNT is presented in Figure 4. C. While the diameter of individual CNT fibers is in the range of 1 – 2 nm, the majority of them were found to be in bundle form (i.e., agglomerated CNT). The covalently attached CNT onto the PVDF membrane (P-CNT) is shown in Figure 4. D. Significant changes in membrane morphology were observed as a result of CNT anchoring. An abundance of bonded CNT can be seen on top of the membrane with a highly heterogeneous net-like structure observed. EDAX analyses (Figure 4. A and B) reveal a presence of nitrogen and silicon in the P-NH2, as a result of surface silanization with (3aminopropyl)triethoxysilane. The EDAX analysis of the top surface of P-CNT show a reduction in its fluorine content, when compared to P-NH2 (Figure 4. A and D), which is due to the surface coverage of membrane with CNT as well as partial defluorination due to the use of piranha solution in the activation process . Furthermore, significant increase in carbon, nitrogen and oxygen elements were recorded in P-NH2, due to the functionalization process.
Fig. 4. SEM and EDAX analyses of non-modified PVDF (A), PVDF modified with (3aminopropyl)triethoxysilane (B), pure carbon nanotubes (C) and PVDF with chemically attached CNT (view from the top – D and cross section – D.1 and D.2).
Surface roughness of the membrane samples was estimated using Gwyddion 2.45 Software (freeware version) and is presented as a root mean square (RMS) of height deviations from the mean data plane. Although the RMS roughness value is highly sensitive to great deviations with respect to the mean line , it is still considered as appropriate in analyzing the influence of modification on the spatial features of the membranes. The effects of activation and functionalization of the membrane on surface roughness are reflected in Figure 5. The RMS values were found to be 145 ± 7 nm, 183 ± 9 nm and 260 ± 13 nm for 16
PVDF, P-NH2 and P-CNT, respectively. The amplitude and height profiles of the surface morphology were also measured using the AFM technique (Figure 5). The measured size range of the CNT, as obtained from the AFM amplitude profiles, was 1-2 nm, which is comparable to the size supplied by the manufacturer. It can be seen that the heterogeneity of membrane
aminopropyl)triethoxysilane (P-NH2) and then further upon CNT attachment to the P-NH2 surface.
Fig. 5. (A): AFM amplitude profiles of P-CNT sample showing a CNT diameter comparable to the range supplied by the manufacturer. (B): AFM topography of pristine membrane (PVDF), functionalized membrane (P-NH2) and membrane covalently-entangled with CNT (P-CNT). 17
3.3 Effect of treatment on membrane hydrophilicity The physicochemical properties of the PVDF, P-NH2 and P-CNT membranes were evaluated based on the following parameters: contact angle (CA), overall surface free energy (SFE) and g
(γ cr). All membranes possessed
a hydrophobic character, with CA > 110o. P-NH2 and P-CNT were characterized by an 8% and 20% increase, respectively, in their CA, compared to the PVDF membrane (Figure 6).
Fig. 6. Physicochemical properties (SFE, CA and roughness) of PVDF, P-NH2 and P-CNT
SFE of the three membranes is inversely proportional to their roughness and CA (Figure 6). The Owens−Wendt method were implemented during the determination of SFE [34, 44]. This method assumes that the overall SFE includes two components: polar and dispersive [34, 45, 46]. Polar interaction can be found in molecules having a dipole moment. This component of SFE consists of hydrogen, polar, induction (Debye) as well as acid-base forces [34, 44-46]. The common feature of molecules in the aforementioned interactions is having 18
stable inequity in their electron density due to the adequate electronegativities of the bonding partners, having concurrently asymmetrical geometry (e.g., water). Molecules with a dipole moment can form polar interactions among each other. The second component of SFE, the dispersive one, is generated due to the temporary differences in electron density where electrons are not perfectly localized within the molecule . As an effect of this delocalization, temporary dipoles are generated, which further induce momentary dipoles in nearby molecules. Dispersive interactions are generally weak interactions but often are more dominant [34, 45, 46]. A dominant characteristic of hydrophobic and super-hydrophobic materials is a lower contribution of polar interactions towards their total SFE [34, 45, 46]. According to the collected data (Figure 6), both a reduction of total SFE and a smaller contribution of the polar component thereof were observed upon the functionalization and CNT binding of the PVDF membrane, along with the formation of more hydrophobic surface. It can also be seen that the modification process has an important influence on the polar component of SFE. A reduction of 65% and 56% of polar component was detected for P-NH2 and P-CNT, respectively, when compared to the pristine membrane (PVDF). One of the most important features of a newly developed material is its physical interaction with water. This interaction can be best evaluated using the Zisman plot , based on which the critical surface tension of the surface can be calculated. The Zisman plot for pristine and modified membranes is depicted in Figure 7. It can be seen that both amine functionalization and the modification with CNTs caused a d
(γ cr). Thus, P-
NH2 and P-CNTs samples are much more resistant to wetting when compared to the pristine PVDF. The liquid should have a surface tension lower than 31.7 mN m-1 and 29.0 mN m-1 in order to wet P-NH2 and P-CNT membranes, respectively. This observation is in agreement with the observed higher CA as well as higher heterogeneity of both modified membranes.
Fig. 7. Zisman plot for pristine (PVDF) sample, functionalized membrane with (3aminopropyl)triethoxysilane (P-NH2) and chemically modified with CNTs (P-CNT).
The Kao diagram [48-51] combines the hydrophobicity/hydrophilicity of a surface with homo/heterogeneity as well as wettability properties (Figure 8.A). The surface hydrophobicity has two major components, chemical and geometrical. Shibuichi et al [48-51] prepared a set of various samples possessing well-developed fractal surfaces. The samples were made from alkylketene dimer and were characterized by contact angle value of around to 174°. These highly heterogeneous, rough surfaces were then compared with perfectly flat samples of the same chemistry. The highest CA value measured on the flat surfaces was below 109o. The collected data were presented in one plot, called Kao diagram. It depicts the correlation between the calculated cosine of apparent CA on a rough surface (θr) and the cosine of CA measured on perfectly smooth surface (so-called Young surface) (θs). CA measurements for a broad spectrum of liquids, with different liquid surface tension values, 20
were conducted using the PVDF, P-NH2 and P-CNT, and a Kao plot was generated as shown in Figure 8. In the first quadrant of the coordinate system, the data for hydrophilic and superhydrophilic surfaces are gathered, regardless of surface roughness. Solvents with low liquid surface tension (such as N,N-dimethyl-formamide, xylene, toluene, tetrahydrofuran, and ndodecane) were found to readily wet and soak the membrane (Figure 8). More detailed analysis of Kao diagram was done using Wenzel and/or Cassie-Baxter models [52-55]. The Wenzel model  is used for the description of wettability behavior on homogeneous surfaces, whereas the Cassie-Baxter theory  is for interpretation of the heterogeneous surface possessing hydrophobic or superhydrophobic character. Considering the testing liquids used during our measurements, membrane roughness values and their critical surface tension, it is possible to assess the wettability of the evaluated PVDF membranes. For instance, when applying water as a testing liquid on a highly hydrophobic rough surface, the micro air pockets trapped within the rough surface will lead to a composite solid–liquid–air interface formation and increase of CA value. Pristine PVDF surface was placed in the Wenzel region in Figure 8.B. The PVDF membrane was characterized by the lowest RMS value and possessed less hydrophobic character (Figure 5). The P-NH2 was located somewhere between Wenzel and Cassie-Baxter region due to its high hydrophobicity (CA = 120o) and rough surface. Whereas, the P-CNT membrane with chemically attached CNT was better described by the Cassie-Baxter model . P-CNT possessed both high roughness and hydrophobic character (Figure 5). Because of the introduction of CNT nanostructure, the apparent CA changed significantly (third quadrant of the coordinate system). Kao diagram proved the above-mentioned relation between roughness, hydrophobicity and wettingresistance observed in the current study. Additionally, the established results are in good accordance with the literature and prediction from Kao theory [48-54]. Thus, it can be seen (Figure 8.B) that grafting with (3-aminopropyl)triethoxysilane (P-NH2) and subsequent
chemical anchoring of CNT changed the physicochemistry and morphology of the PVDF membranes.
Fig. 8. (A): Theoretical Kao diagram. (B): Kao diagram for pristine (PVDF) sample, functionalized membrane with (3-aminopropyl)triethoxysilane (P-NH2) and chemically modified with CNTs (P-CNT).
The AFM technique was used for the determination of adhesion force (Fadh), which describes the tribological properties of the membranes. The Fadh value of PVDF membrane increased from 16.2 to 26.1 nN after the attachment of CNT. On the other hand, it has increased fourfold (upto 60 nN) in the case of P-NH2 sample. From the results, it can be stated that the functionalization process has a noticeable impact on the value of adhesion force. This change was related to the silanization process by (3-aminopropyl)triethoxysilane grafting agent and amino furnishing of the surface. The smaller Fadh value of P-CNT in comparison to P-NH2 is due to the higher contact angle, more heterogeneous surface (surface classified to CassieBaxter model – Figure 8B) and lower surface free energy (Figure 6) of the former. Additionally, this phenomenon is related to the fact that the friction forces are controlled by 22
the surface interactions under low contact pressures [43, 56]. In this certain case, adhesion plays essential role. The presence of adhesive forces stems from chemical covalent bonding, electrostatic, van der Waals forces, as well as capillary forces between two contacting surfaces [43, 56]. In this particular case, adhesion plays essential role. The presence of adhesive forces stems from chemical covalent bonding, electrostatic, van der Waals forces, as well as capillary forces between two contacting surfaces . Nano-hardness (H) and Young modulus were determined during nanotribological analysis of the samples. The results from nano-indentation are presented in Figure 9. It was observed that introduction of CNT onto the PVDF membrane improved its mechanical and tribological properties significantly. As a result of covalent attachment of CNT, Young modulus of P-CNT increased from 2.1 to 2.63 GPa. Furthermore, the nano-hardness changed from 0.11 to 0.31 GPa. These substantial changes in nano-hardness of the membranes can be attributed to the attached carbon nanotubes . Wang et al.  highlighted that the addition of CNT to polymeric matrix (in epoxy composites) could enhance the mechanical properties as well as the friction performance.
Fig. 9. Nanoindentation test for PVDF and P-CNT samples – AFM images and profile analysis.
3.4 Membrane resistance to wetting Membrane wettability is crucial in deciding certain target membrane applications, such as membrane distillation, membrane bioreactors or pervaporation. Therefore, it is essential to determine the impact of membrane functionalization on the water resistance of PVDF membranes. Dynamic contact angles as well as spreading pressure (S) were measured for this purpose. During dynamic CA measurements, the following parameters were tracked: height of wetting liquid drop, drop base diameter, drop volume, and work of adhesion, all as functions of time (Figure 10). Generally, two major wetting behaviors are possible to observe during dynamic CA measurements: i) surface soaking and ii) evaporation of
contacting liquid from the examined surface [46, 52]. This is related to the physicochemistry of the surface, hydrophobicity/hydrophilicity level and the roughness of the membranes.
Fig. 10. Dynamic contact angle measurements: apparent contact angle (A), volume of wetting liquid drop (B), drop base diameter (C), height of the drop (D), and work of adhesion (E), all as functions of time.
Rapid wetting with water will occur in the case of hydrophilic samples, manifested in a reduction of drop height and volume accompanied with an increase of drop base diameter. On the other hand, for highly hydrophobic surfaces, no wetting but only evaporation of water from the surface will be observed. During this phenomenon, reduction of droplet volume and height will be observed, with the drop base diameter remaining constant until the evaporation of the liquid is complete. For transitional hydrophilic/hydrophobic surfaces, both surface wetting and evaporation of the contacting liquid can be observed. Significant influence of activation and functionalization of PVDF on the hydrophobicity of the membranes and on their wetting resistance is observed. A hydrophobic character was observed for PVDF, P-NH2 and P-CNT (Figure 10). The apparent contact angle for PVDF, P-NH2 and P-CNT were 110o, 1200 and 1310, respectively. It can also be noted that P-NH2 and P-CNT surfaces showed higher resistance to wetting compared to the PVDF membrane (Figure 10. A). The observed trend is in good agreement with the literature, where Prasad and co-workers [58, 59] reported a significant increase of water CA after blending CNT into PVDF membranes. An interesting behavior was seen here after 30 sec of the dynamic measurements for P-NH2 and P-CNT (Figure 10. C). After drop deposition on the tested surface, partial soaking was observed. Then, the typical curve characteristic of liquid evaporation from the surface was seen. A slight increase of drop base was also noticed. These changes can be attributed to the ~20 % increase of work of adhesion of the membrane/testing liquid interfaces (Figure 10. E). According to the data presented in Figure 10 A, it is shown that the P-NH2 has a higher water resistance than P-CNT. This observation can be explained by the presence of carbon in the polymeric matrix. Although, the P-CNT sample is characterized by higher contact angle value, the heterogeneity (roughness) played an important role in the mechanism of surface wettability.
Another factor used for the assessment of wetting behavior is spreading pressure (S) . It is calculated based on the work of adhesion and work of cohesion . The S value was found to be negative for all investigated membranes, with a value of -97. 7, -109.2 and 120.6 mNm-1 for PVDF, P-NH2 and P-CNT, respectively. The reduction of S value was related to the observed changes in surface physicochemistry and morphology after functionalization. The diminution of S indicates the highly wetting-resistant surfaces (P-NH2 and P-CNT) (Figure 10). The negative value is theoretically impossible from a thermodynamic point of view. Nevertheless, it is a common phenomenon in the case of coated and functionalized hydrophobic surfaces. Erbin  explained this phenomenon in detail, describing the influence of hydrophobicity on the S value. It was explained that the modification process contributing to the formation of highly hydrophobic surfaces can negatively affect the S factor. This phenomenon is associated with higher substitution of hydroxyl groups and smaller basicity of the modified/functionalized sample. Consequently, smaller value of total surface tension associated with higher CA value will be observed on this type of surfaces. The negative values of S are characteristic of surfaces on which the wetting was incomplete or where there is lack of wetting . Another parameter evaluated using the dynamic contact angle was work of adhesion (Wadh). Wadh is key in determining the mechanism of wetting of the modified surface. Work of adhesion corresponds to the energy that is released in the wetting process . A diminution of Wadh after functionalization process, first by (3aminopropyl)triethoxysilane (22% reduction) and subsequently by chemical modification with CNT (47% reduction) was observed (Figure 10. E). The observed changes were related to a more hydrophobic character of the samples. 3.5 Effect of treatment on membrane structure The membrane modification protocols followed in this study to react CNT with the PVDF membrane had an impact on the porous structure of the membrane. Mean flow pore size and
maximum pore size of the PVDF and P-CNT membranes are presented in Figure 11.A. When compared to pristine PVDF, a slightly lower mean flow pore size was observed for the P-CNT. Pristine PVDF had a mean flow pore size of 0.33 ± 0.04 μ , whereas P-CNT had a 0.27 ± 0.01 μ
mean flow pore size. This reduction is due to the embellishment of membrane
surface and pore walls with CNTs. A similar trend is observed for the maximum pore size as well (Figure 11.A). Air permeability results, presented in Figure 11.B also showed a similar trend as that of the pore sizes. The functionalized membrane (P-CNT) showed slightly lower air permeability than that of pristine PVDF membrane.
Fig. 11: Mean pore size and maximum pore size (A) and air permeability (B) of PVDF and PCNT membranes. showing lower mean pore size but higher maximum pore size of membranes with nanoparticles.
3.5 Effect of modification on membrane’s electrical conductivity Membrane fouling remains an obstacle to a wide range of applications for microfiltration and ultrafiltration. Recently, some studies showed that fouling could be effectively controlled by applying an external electric field, as long as the membrane can be rendered electrically
conductive [63-65]. For this purpose, efforts have been invested to alter the electrical properties of this material by adding several nano-composites . In this regard, CNT has remarkable electrochemical, electric, magnetic, and mechanical properties. These properties are tunable by changing few characteristics of the CNT such as their diameter, length and chirality. The current-voltage (I-V) curve generated for the P-CNT membrane is presented in Figure 12 and it illustrates a perfect correlation between current through the membrane and the voltage applied, indicating membrane conductance. Zero-field sheet resistivity, sheet resistivity, sheet carrier density, and hall mobility values were found to be 3.89 x 10 4 ohm/sqr, 3.53 x 104 ohm/sqr, 4.25 x 1011 cm-1 and 3.79 x 102 cm/VS, respectively. Moreover, P-CNT membranes could also be considered in various other applications, such as stimuli-responsive materials and sensors. In order to achieve a high membrane conductivity in such applications, higher concentration of CNTs is required in general. Nevertheless, in the traditional way of adding CNT to polymeric membranes via blending, it is challenging to blend higher concentration of CNT with the polymeric solution as that affect the membrane formation thermodynamics. Therefore, the herein presented process of surface grafting, via covalent bonding, of CNT onto the PVDF membranes offers a potential solution in achieving a highly conducting polymer membrane.
Fig. 12. I – V curve of P-CNT
4. Conclusion PVDF membranes were activated by utilizing a novel piranha reaction method where hydroxyl groups were generated onto the PVDF membrane surface. The membrane was then functionalized with amine groups, which could be used to anchor a range of very interesting functional groups and nanomaterials on the membrane surface for a variety of applications. We demonstrated this by covalently attaching CNT onto the PVDF membrane surface. It was found that piranha activation and consequent functionalization processes impacted the structural and physicochemical properties of PVDF membranes. A direct correlation of surface chemistry and surface roughness of the membranes was demonstrated using Kao diagram. The highly conductive CNT, upon being covalently bound to the PVDF membrane surface, significantly altered the electrical properties of the polymer yielding a conductive polymeric membrane. Membrane wetting resistance data and the Zisman plot showed that amine-functionalized as well as CNT-bound PVDF membranes possessed higher water resistance compared to the pristine PVDF membrane. The modified polymer membrane has promising applications in piezoresistive actuators, P-N junctions, and biosensors, to name a few. Thus, the developed post-fabrication protocol could enable the target-specific finetuning of PVDF membrane.
References  B. Onal-Ulusoy, Effects of Plasma-Modified Polyvinylidenefluoride and Polyethersulfone Ultrafiltration (UF) Membrane Treatments on Quality of Soybean Oil, Journal of Food Quality 38(4) (2015) 285-296.
 W.-T. Xu, Z.-P. Zhao, M. Liu, K.-C. Chen, Morphological and hydrophobic modifications of PVDF flat membrane with silane coupling agent grafting via plasma flow for VMD of ethanol–water mixture, Journal of Membrane Science 491(Supplement C) (2015) 110-120.  J. Gao, J. Yu, C. Li, Chemical Modification of Polyvinylidene Fluoride (PVDF) Membrane and Its Application to Milk Purification, International Journal of Nonlinear Sciences and Numerical Simulation, 2010, p. 37.  M.O. Mavukkandy, M.R. Bilad, A. Giwa, S.W. Hasan, H.A. Arafat, Leaching of PVP from PVDF/PVP blend membranes: impacts on membrane structure and fouling in membrane bioreactors, Journal of Materials Science 51(9) (2016) 4328-4341.  K. Ishihara, M. Kobayashi, N. Ishimaru, I. Shinohara, Glucose Induced Permeation Control of Insulin through a Complex Membrane Consisting of Immobilized Glucose Oxidase and a Poly(amine), Polym J 16(8) (1984) 625-631.  A.M. Mika, R.F. Childs, J.M. Dickson, Chemical valves based on poly(4-vinylpyridine)filled microporous membranes, Journal of Membrane Science 153(1) (1999) 45-56.  L. Xiao, D.M. Davenport, L. Ormsbee, D. Bhattacharyya, Polymerization and Functionalization of Membrane Pores for Water Related Applications, Ind. Eng. Chem. Res. 54(16) (2015) 4174-4182.  J. Fu, S. Das, G. Xing, T. Ben, V. Valtchev, S. Qiu, Fabrication of COF-MOF Composite Membranes and Their Highly Selective Separation of H2/CO2, Journal of the American Chemical Society 138(24) (2016) 7673-7680.  E. Ahmadi Feijani, A. Tavasoli, H. Mahdavi, Improving Gas Separation Performance of Poly(vinylidene fluoride) Based Mixed Matrix Membranes Containing Metal–Organic
Frameworks by Chemical Modification, Ind. Eng. Chem. Res. 54(48) (2015) 1212412134.  Y.H. Teow, A.A. Latif, J.K. Lim, H.P. Ngang, L.Y. Susan, B.S. Ooi, Hydroxyl functionalized PVDF–TiO2 ultrafiltration membrane and its antifouling properties, Journal of Applied Polymer Science 132(21) (2015) n/a-n/a.  X. Shen, X. Yin, Y. Zhao, L. Chen, Improved protein fouling resistance of PVDF membrane grafted with the polyampholyte layers, Colloid. Polym. Sci. 293(4) (2015) 1205-1213.  A. Ferrreira, J.G. Rocha, A. Ansón-Casaos, M.T. Martínez, F. Vaz, S. Lanceros-Mendez, Electromechanical performance of poly(vinylidene fluoride)/carbon nanotube composites for strain sensor applications, Sensors and Actuators A: Physical 178 (2012) 10-16.  P.D. Petrov, G.L. Georgiev, Fabrication of super-macroporous nanocomposites by deposition of carbon nanotubes onto polymer cryogels, European Polymer Journal 48(8) (2012) 1366-1373.  J. Martín, M. Hernández-Vélez, O. de Abril, C. Luna, A. Munoz-Martin, M. Vázquez, C. Mijangos, Fabrication and characterization of polymer-based magnetic composite nanotubes and nanorods, European Polymer Journal 48(4) (2012) 712-719.  C.M. Wu, M.H. Chou, Sound absorption of electrospun polyvinylidene fluoride/graphene membranes, European Polymer Journal 82(Supplement C) (2016) 3545.  N. Karousis, N. Tagmatarchis, D. Tasis, Current Progress on the Chemical Modification of Carbon Nanotubes, Chemical Reviews 110(9) (2010) 5366-5397.
 T. Lin, V. Bajpai, T. Ji, L. Dai, Chemistry of Carbon Nanotubes, Australian Journal of Chemistry 56(7) (2003) 635-651.  H. Wang, R. Wang, L. Sun, Z. Liu, Y. Zhu, Y. Zhu, Mechanical and tribological characteristics of carbon nanotube-reinforced polyvinylidene fluoride (PVDF)/epoxy composites, RSC Advances 6(51) (2016) 45636-45644.  L. Wang, Z.-M. Dang, Carbon nanotube composites with high dielectric constant at low percolation threshold, Applied Physics Letters 87(4) (2005) 042903.  M.S.P. Shaffer, A.H. Windle, Fabrication and Characterization of Carbon Nanotube/Poly(vinyl alcohol) Composites, Advanced Materials 11(11) (1999) 937-941.  L. Zonder, A. Ophir, S. Kenig, S. McCarthy, The effect of carbon nanotubes on the rheology and electrical resistivity of polyamide 12/high density polyethylene blends, Polymer 52(22) (2011) 5085-5091.  X.-G. Tang, M. Hou, J. Zou, R. Truss, M. Yang, Z. Zhu, Toughening and reinforcement f
Composites Science and Technology 72(2) (2012) 263-268.  L. Li, M. Zhang, W. Ruan, Studies on synergistic effect of CNT and CB nanoparticles on PVDF, Polymer Composites 36(12) (2015) 2248-2254.  A. Baji, Y.-W. Mai, M. Abtahi, S.-C. Wong, Y. Liu, Q. Li, Microstructure development in electrospun carbon nanotube reinforced polyvinylidene fluoride fibers and its influence on tensile strength and dielectric permittivity, Composites Science and Technology 88 (2013) 1-8.  H. Dong, K. Xiao, X. Li, Z. Wang, S. Guo, Preparation of PVDF/Al2O3 hybrid membrane via alkaline modification and chemical coupling process, Desalin. Water Treat. 51(19-21) (2013) 3800-3809. 33
 D. Sinirlioglu, A.E. Muftuoglu, Investigation of proton conductivity of PVDF based d
x g ”
g J. P
(PEM ) b
. R . 22(12) (2015) 1-16.
 D. Sinirlioglu, A.E. Muftuoglu, A. Bozkurt, Investigation of perfluorinated proton exchange membranes prepared via a facile strategy of chemically combining poly(vinylphosphonic acid) with PVDF by means of poly(glycidyl methacrylate) grafts, J. Polym. Res. 22(8) (2015) 1-17.  M.F. Rabuni, N.M. Nik Sulaiman, M.K. Aroua, N.A. Hashim, Effects of Alkaline Environments at Mild Conditions on the Stability of PVDF Membrane: An Experimental Study, Ind. Eng. Chem. Res. 52(45) (2013) 15874-15882.  W.-Z. Lang, X. Zhang, J.-P. Shen, H.-P. Xu, Z.-L. Xu, Y.-J. Guo, The contrastive study of chemical treatment on the properties of PVDF/PFSA and PVDF/PVP ultrafiltration membranes, Desalination 341 (2014) 72-82.  W.M. de Vos, B. Cattoz, M.P. Avery, T. Cosgrove, S.W. Prescott, Adsorption and Surfactant-Mediated Desorption of Poly(vinylpyrrolidone) on Plasma- and PiranhaCleaned Silica Surfaces, Langmuir 30(28) (2014) 8425-8431.  K.-S. Koh, J. Chin, J. Chia, C.-L. Chiang, Quantitative Studies on PDMS-PDMS Interface Bonding with Piranha Solution and its Swelling Effect, Micromachines 3(2) (2012) 427.  S. Al-Gharabli, J. Kujawa, M.O. Mavukkandy, H.A. Arafat, Functional groups docking on PVDF membranes: Novel Piranha approach, European Polymer Journal 96(Supplement C) (2017) 414-428.
 W.A. Zisman, Relation of the Equilibrium Contact Angle to Liquid and Solid Constitution, Contact Angle, Wettability, and Adhesion, AMERICAN CHEMICAL SOCIETY1964, pp. 1-51.  D.K. Owens, R.C. Wendt, Estimation of the surface free energy of polymers, Journal of Applied Polymer Science 13(8) (1969) 1741-1747.  E. Guillen-Burrieza, A. Servi, B.S. Lalia, H.A. Arafat, Membrane structure and surface morphology impact on the wetting of MD membranes, Journal of Membrane Science 483(Supplement C) (2015) 94-103.  R. Thomas, E. Guillen-Burrieza, H.A. Arafat, Pore structure control of PVDF membranes using a 2-stage coagulation bath phase inversion process for application in membrane distillation (MD), Journal of Membrane Science 452(Supplement C) (2014) 470-480.  M.S. Dresselhaus, G. Dresselhaus, A. Jorio, A.G. Souza Filho, R. Saito, Raman spectroscopy on isolated single wall carbon nanotubes, Carbon 40(12) (2002) 20432061.  M.S. Dresselhaus, G. Dresselhaus, R. Saito, A. Jorio, Raman spectroscopy of carbon nanotubes, Physics Reports 409(2) (2005) 47-99.  J.R. Sanchez-Valencia, T. Dienel, O. Groning, I. Shorubalko, A. Mueller, M. Jansen, K. Amsharov, P. Ruffieux, R. Fasel, Controlled synthesis of single-chirality carbon nanotubes, Nature 512(7512) (2014) 61-64.  Z. M k M
ć D. K ć
ć I. H ć B.
ć M. D f
nanotube thin films treated by laser irradiation and dynamic and isothermal oxidation, Journal of Raman Spectroscopy 43(10) (2012) 1413-1422. 35
 K.J. Ziegler, Z. Gu, H. Peng, E.L. Flor, R.H. Hauge, R.E. Smalley, Controlled Oxidative Cutting of Single-Walled Carbon Nanotubes, Journal of the American Chemical Society 127(5) (2005) 1541-1547.  S. Huang, W.A. Yee, W.C. Tjiu, Y. Liu, M. Kotaki, Y.C.F. Boey, J. Ma, T. Liu, X. Lu, Electrospinning of Polyvinylidene Difluoride with Carbon Nanotubes: Synergistic Effects of Extensional Force and Interfacial Interaction on Crystalline Structures, Langmuir 24(23) (2008) 13621-13626.  J. Kujawa, W. Kujawski, Functionalization of Ceramic Metal Oxide Powders and Ceramic Membranes by Perfluoroalkylsilanes and Alkylsilanes Possessing Different Reactive Groups: Physicochemical and Tribological Properties, ACS Applied Materials & Interfaces 8(11) (2016) 7509-21.  G. Soliveri, V. Pifferi, R. Annunziata, L. Rimoldi, V. Aina, G. Cerrato, L. Falciola, G. Cappelletti, D. Meroni, Alkylsilane–SiO2 Hybrids. A Concerted Picture of Temperature Effects in Vapor Phase Functionalization, J. Phys. Chem. C 119(27) (2015) 1539015400.  E. Chibowski, Surface free energy of a solid from contact angle hysteresis, Adv. Colloid Interf. Sci. 103(2) (2003) 149-172.  E. Chibowski, M. Jurak, L. Holysz, A. Szczes, Wetting properties of model biological membranes, Curr. Opin. Colloid In. 19(4) (2014) 368-380.  R. Walsh, Bond dissociation energy values in silicon-containing compounds and some of their implications, Accounts Chem. Res. 14(8) (1981) 246-252.  T. Onda, S. Shibuichi, N. Satoh, K. Tsujii, Super-Water-Repellent Fractal Surfaces, Langmuir 12(9) (1996) 2125-2127.
 S. Shibuichi, T. Onda, N. Satoh, K. Tsujii, Super Water-Repellent Surfaces Resulting from Fractal Structure, J. Phys. Chem-US 100(50) (1996) 19512-19517.  S. Shibuichi, T. Yamamoto, T. Onda, K. Tsujii, Super Water- and Oil-Repellent Surfaces Resulting from Fractal Structure, J. Colloid Interf. Sci. 208(1) (1998) 287-294.  K. Tsujii, T. Yamamoto, T. Onda, S. Shibuichi, Super Oil-Repellent Surfaces, Angew. Chem. Int. Edit. 36(9) (1997) 1011-1012.  J. Kujawa, S. Cerneaux, S. Koter, W. Kujawski, Highly Efficient Hydrophobic Titania Ceramic Membranes for Water Desalination, ACS Applied Materials & Interfaces 6(16) (2014) 14223-14230.  J. Kujawa, S. Cerneaux, W. Kujawski, M. Bryjak, J. Kujawski, How To Functionalize Ceramics by Perfluoroalkylsilanes for Membrane Separation Process? Properties and Application of Hydrophobized Ceramic Membranes, ACS Applied Materials & Interfaces 8(11) (2016) 7564-7577.  R.N. Wenzel, Resistance of solid surfaces to wetting by water, Ind. Eng. Chem. Res. 28(8) (1936) 988-994.  A.B.D. Cassie, S. Baxter, Wettability of porous surfaces, T. Faraday Soc. 40(0) (1944) 546-551.  B. Bhushan, M. Cichomski, E. Hoque, A. DeRose, P. Hoffmann, J. Mathieu, Nanotribological characterization of perfluoroalkylphosphonate self-assembled monolayers deposited on aluminum-coated silicon substrates, Microsyst. Technol. 12(6) (2006) 588-596.  J. Kujawa, W. Kujawski, Functionalization of Ceramic Metal Oxide Powders and Ceramic Membranes by Perfluoroalkylsilanes and Alkylsilanes Possessing Different
Reactive Groups: Physicochemical and Tribological Properties, ACS Applied Materials & Interfaces 8(11) (2016) 7509-7521.  R.P.S. Chakradhar, G. Prasad, P. Bera, C. Anandan, Stable superhydrophobic coatings using PVDF–MWCNT nanocomposite, Applied Surface Science 301 (2014) 208-215.  G. Prasad, A.A. Prabu, Effect of Stirring on Hydrophobicity of PVDF/CNT Nanocomposite Coatings, Adv. Mat. Res. 938 (2014) 199-203.  G.M. Dorris, D.G. Gray, Adsorption, spreading pressure, and london force interactions of hydrocarbons on cellulose and wood fiber surfaces, J. Colloid Interf. Sci. 71(1) (1979) 93-106.  G. Heydari, E. Thormann, M. Järn, E. Tyrode, P.M. Claesson, Hydrophobic Surfaces: Topography Effects on Wetting by Supercooled Water and Freezing Delay, Journal of Physical Chemistry C 117(42) (2013) 21752-21762.  H.Y. Erbil, Calculation of Spreading Pressure of Water on Callulosic Films from Contact Angle Data, Tr. J. of Chemistry 21 (1997) 332-345.  J. Huang, Z. Wang, J. Zhang, X. Zhang, J. Ma, Z. Wu, A novel composite conductive microfiltration membrane and its anti-fouling performance with an external electric field in membrane bioreactors, Scientific reports 5 (2015) 9268.  C.-F. de Lannoy, D. Jassby, K. Gloe, A.D. Gordon, M.R. Wiesner, Aquatic Biofouling Prevention by Electrically Charged Nanocomposite Polymer Thin Film Membranes, Environmental Science & Technology 47(6) (2013) 2760-2768.  A.V. Dudchenko, J. Rolf, K. Russell, W. Duan, D. Jassby, Organic fouling inhibition on electrically conducting carbon nanotube–polyvinyl alcohol composite ultrafiltration membranes, Journal of Membrane Science 468(Supplement C) (2014) 1-10.
Novel approach of piranha regent for PVDF activation at mild conditions Efficient functionalization of activated PVDF membranes with CNT Correlation between physicochemistry and wetting properties of the functionalized PVDF Membranes became electrically conductive as a result of CNT docking