Poly(meth)acrylate nanocomposite membranes containing in situ exfoliated graphene platelets: Synthesis, characterization and gas barrier properties

European Polymer Journal 94 (2017) 431–445

Contents lists available at ScienceDirect

European Polymer Journal journal homepage: www.elsevier.com/locate/europolj

Macromolecular Nanotechnology

Poly(meth)acrylate nanocomposite membranes containing in situ exfoliated graphene platelets: Synthesis, characterization and gas barrier properties

MARK

⁎

Lenka Polákováa, , Zdeňka Sedlákováa, Petra Ecorchardb, Ewa Pavlovaa, Jakub Petera, Bartosz Paruzela, Hynek Beneša a b

Institute of Macromolecular Chemistry, AS CR, Heyrovsky Sq. 2, 162 06 Prague 6, Czech Republic Institute of Inorganic Chemistry, AS CR, 250 68 Řež, Czech Republic

AR TI CLE I NF O

AB S T R A CT

Keywords: Graphite intercalation compounds Quaternary ammonium salt Exfoliation Polymer composite membrane Gas barrier properties Electrical conductivity

Here we report a novel method for the preparation of polymer nanocomposites containing in situ exfoliated graphene nanoplatelets. First, graphite particles were modified by sodium and ethylenediamine (en), yielding first-stage graphite intercalation compounds (GICs) containing [Na (en)]+ complex between carbon sheets. After subsequent ion-exchange reaction of the intercalated complex with quaternary ammonium salt bearing methacrylamide group, GICs possessing polymerizable double bond were obtained. It was found that the extent of [Na(en)]+ complex intercalation as well as the ion-exchange reaction was significantly dependent on the type of the graphite used (natural vs. synthetic). GICs derived from natural graphite were further employed in the synthesis of polymer nanocomposites. Using NMR and TEM it was confirmed that the methacrylamide group in GICs participated successfully in the process of copolymerization with 2-hydroxyethyl methacrylate and 2-ethylhexyl acrylate, resulting in the in situ GICs exfoliation. The prepared nanocomposite membranes were further characterized using X-ray diffraction, DSC and TGA. Gas permeability measurements on the polymer composites were also carried out, confirming that the initial GIC was present in an exfoliated form.

1. Introduction Graphite is distinguished by its unique multilayered structure, consisting of mutually parallel two-dimensional carbon sheets bonded by weak van der Waals interactions. This bond arrangement allows intercalation of anions or cations via oxidation or reduction reactions without disrupting the layered topology of graphite, resulting in a formation of graphite intercalation compounds (GICs) [1–5]. In a GIC not every layer is necessarily occupied by guests; in so-called first-stage compounds, graphene layer and intercalate layer alternate, whereas in high-stage compounds, intercalate layer randomly separates stacks of graphite multilayers [6]. Obviously, a large variety of possibilities exists between these two extreme cases. In recent years, GICs prepared with the use of low molecular weight quaternary ammonium salts as graphite delaminating agents have been intensively studied by many authors [1,2,7,8]. Several GICs were employed as intermediates for the preparation of exfoliated single or few layered graphene flakes. Zhou et al. obtained high-quality graphene nanoplatelets by mild sonication of graphite intercalated by tetraethylammonium salt [3]. Li et al. inserted tetrabutylammonium hydroxide into previously modified

⁎

Corresponding author. E-mail address: [email protected] (L. Poláková).

http://dx.doi.org/10.1016/j.eurpolymj.2017.07.033 Received 17 May 2017; Received in revised form 17 July 2017; Accepted 24 July 2017 Available online 24 July 2017 0014-3057/ © 2017 Elsevier Ltd. All rights reserved.

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

thermally expanded graphite, and then sonicated the product in a solution of a surfactant in DMF, yielding a homogenous suspension of graphene monolayers [9]. Process of GICs exfoliation is generally irreversible, and leads to a spongy, low-density and high-surfacearea carbon material. The resulting materials exhibit, besides others, excellent electrical [10,11] and thermal conductivity [12,13], high thermal stability [10,11], or gas sorption properties convenient for example as gas sensors [14] or membranes for gases separation [15–17]. The application of GICs for fabrication of graphene/polymer nanocomposites presents an alternative route to mainstream methods utilizing graphene oxide (GO) or its reduced analog (RGO), beneficially avoiding large formation of oxidative defects in the graphene structure. Generally, there are many synthetic ways leading to preparation of graphene-containing polymer nanocomposites using GICs [18]. The first approach involving direct intercalation of polymer chain between graphene sheets is, however, limited to oligomers or low molecular weight polymers [19]. Thermal shock treatment of GICs followed by ultrasonication was applied for preparation of semi-interpenetrating network of epoxy resin and unsaturated polyester containing graphite nanosheets [20]. An in situ exfoliation of GICs by melt blending [21–24] or in polymer solution [25,26] were also mentioned. An in situ polymerization is another way of nanocomposite preparation, which has been proved to be more efficient overcoming common drawbacks of the above mentioned methods such as too high polymer viscosity, the necessity of a solvent usage, low GICs exfoliation degree, etc. During the in situ polymerization, a monomer is first introduced into GIC galleries, and subsequently (co)polymerized. The GICs expansion took place only when the reactivity of monomer is high enough, for example in the case of vinyl monomers (isoprene, styrene, 1,3butadiene) in the presence of donor-type GICs (e.g. KC24) [27]. Xu et al. performed the in situ polymerization of methyl methacrylate (MMA) in the presence of graphite nanoplatelets with the assistance of sonication and heating, which helped MMA and growing polymer chains to penetrate into the pores of graphite enabling its exfoliation [28]. These examples demonstrate that an initiator or a monomer located in the GIC galleries induce inter-gallery polymerization and subsequently might promote delamination of graphene layers. In our previous work, we have synthesized GICs intercalated by a quaternary ammonium salt bearing polymerizable double bond, and by its subsequent free radical copolymerization with n-butyl methacrylate we have obtained polymer composite containing graphite/graphene structures [29]. However, the process of the intercalate in situ exfoliation in the composite proceeded to a certain extent only, probably due to more hydrophobic nature of the polymer matrix when compared to the modified GICs. The current study is therefore focused on (i) optimization of reaction conditions for the synthesis of various GICs possessing polymerizable methacrylamide group (using three different types of graphite as a starting material), (ii) preparation of polymer composite membranes by free radical copolymerization of 2-hydroxyethyl methacrylate and 2-ethylhexyl acrylate in the presence of GICs, and (iii) characterization of the prepared membranes using XRD, NMR and electron microscopy. Gas permeability measurements were carried out to determine the influence of the various filler types on gas transport properties of the related nanocomposites. Thermal stability of the prepared nanocomposites, which is also an important indicator for the membranes applicability, was evaluated by DSC and TGA. 2. Materials and methods 2.1. Materials Natural graphite (NG; PMM11 Very fine crystalline graphite powder, Koh-i-noor Grafit, Ltd., Czech Republic), synthetic graphite (SG; Synthetic graphite Grade 8028B, Graphite Týn, Ltd., Czech Republic) and TIMREX synthetic graphite (TG; TIMREX® KS75, Songhan Plastic Technology Co., Ltd., China) were used as received. All chemicals were purchased from Sigma-Aldrich Ltd. and used as received. N-[3-(dimethylamino)propyl] methacrylamide (DMAPMA), 2-hydroxyethyl methacrylate (HEMA) and 2-ethylhexyl acrylate (2-EHA) were distilled prior to use. 2.2. Synthesis Monomeric quaternary salt (MQS). MQS was prepared by the reaction of DMAPMA with n-hexadecyl iodide [29]. The G1 and G2 intercalates were synthesized using a slightly modified procedure described therein. G1 intercalates. Natural graphite (500 mg) and freshly cut sodium (100 mg) were dispersed in freshly distilled ethylenediamine (en; 60 mL). The reaction proceeded overnight in nitrogen atmosphere under stirring at 60 °C. Subsequently, en (20 mL) was added to dissolve unreacted sodium, and the reaction mixture was stirred at room temperature for additional 4 h. The black solid was then separated from the supernatant solution by centrifugation, dried under vacuum to a constant weight and designated as NG1. Yield: 750 mg. Intercalates of the synthetic graphite and the TIMREX graphite (SG1 and TG1, respectively) were prepared analogically. G2 intercalates. First, reaction conditions for the G2 synthesis were optimized for NG1 as a starting material by the following procedure. The prepared NG1 intercalate was dispersed in a solution of MQS in DMF (2 mL). The mixture was bubbled with nitrogen and sealed. The ion-exchange reaction was carried out under stirring at 25 °C. Related moieties of starting materials and reaction periods are summarized in Table 1. The product was then separated by centrifugation, washed three times with acetone and methanol to remove salts and unreacted monomer, and dried under vacuum to a constant weight. The extent of the ion-exchange process in dependence on reaction conditions was then carefully assessed using XRD method. Subsequently, the SG2 and TG2 intercalates were synthesized under conditions that have been evaluated as optimal. 2-EHA/HEMA polymer composites. In a typical procedure, NG2_A (22.0 mg, 1.0 wt.%), 2-EHA (0.8 g, 4.3 mmol), HEMA (1.31 g, 10.1 mmol) and acetone (6.3 mL) were placed into an ampoule. After addition of AIBN (35 mg, 213 µmol) the mixture was 432

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

Table 1 Conditions for the ion-exchange reaction between the NG1 intercalate and MQS. Sample

Molar ratio MQS/[Na(en)]+

NG1 (mg)

MQS (mg)

Reaction time (h)

NG2_A NG2_B NG2_C NG2_D

10:1

40

915

5:1

80

24 96 24 96

bubbled thoroughly with nitrogen and sealed. The polymerization was carried out at 50 °C for 24 h under stirring. Subsequently, the reaction mixture was precipitated in diethylether, the polymer was filtered off, dried under vacuum to a constant weight and designated as PNG2_1.0. Yield: 1.9 g (89%). Neat polymer matrix, polymer composites with different NG2_A loading (0.1, 0.25, 0.5, 2.5 and 5.0 wt.%) and polymer composites with NG and NG1 were prepared under similar conditions. Membrane preparation. Polymer composite membranes were prepared by the following procedure. In a 10 mL flask equipped with magnetic stirrer, PNG2_1.0 (500 mg) was dissolved in N,N–dimethylformamide (DMF; 4.5 g). The obtained clear solution was degased on water pump to remove air bubbles, and 2 mL were then poured carefully onto a round flat Teflon mold with a diameter of 5 cm and covered with a thin perforated aluminum foil. The solvent was allowed to evaporate slowly at 25 °C, and the formed membrane was subsequently removed from the mold and dried in vacuum at room temperature to a constant weight. Polymer composite membranes casted from tetrahydrofuran (THF) and acetone solutions were prepared under similar conditions. 2.3. Characterization 1

H nuclear magnetic resonance (NMR) spectra were recorded on the Bruker Avance DPX 300 spectrometer (300.1 MHz; THF-d4 as a solvent, 25 °C). X-ray diffraction patterns (XRD) of the samples were collected with diffractometer Bruker D2 equipped with conventional X-ray tube (Cu Kα radiation, 30 kV, 10 mA). The primary divergence slit module width 0.6 mm, Soller Module 2.5, Air scatter screen module 2 mm, Ni Kβ-filter 0.5 mm, step 0.00809° and time per step 0.5 s and the LYNXEYE one-dimensional detector were used. Transmission electron microscopy (TEM) was performed using ultramicrotomy (Ultrotome III LKB, Sweden) at room temperature. The ultrathin sections were transferred onto copper grids, transferred to a transmission electron microscope (Tecnai G2 Spirit Twin 12, FEI) and observed in bright field mode at accelerating voltage 120 kV. Scanning electron microscopy (SEM) of the neat graphites and GICs was performed using Vega Plus TS 5135 (Tescan, Czech Republic) with secondary electron imaging at 30 kV. The powder samples were fixed on a metallic support using conductive doubleadhesive carbon tape (Christine Groepl, Austria). Thermogravimetric analysis (TGA) was performed on a Pyris 1 TGA (PerkinElmer). The content of intercalated compounds in GICs was determined using the following TGA procedure: powder sample (5–10 mg) was heated at constant heating rate 10 °C/min under nitrogen flow of 30 mL/min from 35 °C to 800 °C and kept few minutes at 800 °C till wt.% plateau (no mass change) appeared. The content of intercalated compounds was determined as the sample mass loss. Thermal degradation of polymer composites was also studied using TGA. In this case, bulk sample (5–10 mg) was heated at a constant heating rate 20 °C/min under nitrogen flow of 30 mL/min from 35 °C to 650 °C. Standard deviation of TGA measurements was under 5%. Differential scanning calorimetry (DSC) measurements were carried out on a Q 2000 calorimeter (TA Instruments) with nitrogen as a purge gas (50 mL/min). The instrument was calibrated using indium as a standard. Samples of approximately 10 mg were encapsulated in aluminum pans. The DSC runs were performed in a cycle of heating – cooling – heating from −60 °C to 150 °C at 10 °C/min. Glass transition temperature (Tg) was determined from the second heating run as the midpoint between the glassy and rubbery branches of the DSC trace. The typical precision of DSC measurements for Tg determination was ± 1 °C. Gas permeability through the membranes was determined according the time-lag method [30] using a high vacuum laboratory apparatus with a static permeation cell. An effective membrane area was 1.24 cm2, measurements were carried out at 30 °C. Prior the measurement, each membrane was placed and sealed in a membrane cell and evacuated in the apparatus at 30 °C for 24 h to degas the sample. Feed pressure pi was 1.5 bar. The permeability coefficient P was determined from the increase of permeate pressure Δpp per time interval Δt in a calibrated volume Vp of the product part of the membrane cell during the steady state of permeation. For calculation of permeability coefficient, the following formula was used:

P=

Δpp Δt

×

Vp l Api

×

1 RT

(1)

where l is the membrane thickness, pi feed pressure, A the area, T the temperature and R the gas constant. Permeabilities are reported in units of Barrer (1 Barrer = 10−10 × cm3STP × cm/(cm2 ×s × cm Hg)). Relative standard deviations of Δpp and Δt were lower than 0.3% (given by the MKS Barratron pressure transducer precision). Relative standard deviation of membrane thickness measurement was 1%, relative standard deviation of calibrated volume was lower than 0.5% and relative standard deviation of feed pressure was 0.3%. For very low permeability coefficients (below 0.01 Barrers), error of the measurement can be up to 10%, and for P values < 0.005 Barrers error rises up to 30%, due to lower precission of the MKS Baratron at very low pressures. All measurements were carried out at 30 °C. In our experiments the following gases were studied: H2, O2, N2, CH4, and CO2. Each gas exhibited a purity 433

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

of 99.99% and was used as received from Messer Technogas s.r.o. (Czech Republic). Gas diffusivities were estimated from the timelag data, using the relation:

D=

l2 6θ

(2)

where l is the film thickness and θ is the time-lag [30]. A precision of 0.1 s for the time-lag determination allowed the determination of the diffusion coefficients of helium and hydrogen. Apparent solubility coefficients were calculated using the following equation: (3)

S = P /D

The overall selectivity of a polymer membrane for a pair of gases i and j is commonly expressed in terms of an ideal separation factor, αij, defined by the following relation:

αij =

Pi S D = i· i Pj Sj Dj

(4)

where Pi and Pj are pure gas permeabilities, Di/Dj is the mobility (or diffusion) selectivity and Si/Sj the solubility selectivity. All timelag measurements were performed at 30 °C. Electrical conductivity measurements. Volume electrical conductivity was measured with the use of Broadband Dielectric Spectroscopy method (BDS) at frequency range of 10−2–106 Hz using Novocontrol Alpha-A Analyzer at applied A.C. voltage Vrms = 1.0 V at ambient temperature. To ensure a proper electrical contact, thin gold electrodes were deposited on both sides of membrane by physical vapor deposition (PVD) (pressure 2 × 10−5 torr, deposition rate 1 nm/s). Effective thickness of the layers (typically 50 nm) as well as the deposition rate were determined using a crystal balance monitor. 3. Results and discussion Due to the layered structure of graphite, small molecules can be introduced (intercalated) between the carbon layers, giving graphite intercalation compounds (GICs). Here, the individual graphite samples were heated in the presence of sodium and ethylenediamine, yielding GICs designated as G1. During the reaction, the [Na(en)]+ complex is forming, acting as a delaminating agent of the layered carbon nanosheets [1,4]. The subsequent ion-exchange reaction between the G1 with intercalated [Na(en)]+ complex and the previously synthesized methacrylamide bearing quaternary ammonium group (MQS) was carried out to obtain the G2 GICs having ion-bonded polymerizable group in the structure [29]. General scheme of the GICs preparation as well as the synthesis of their related polymer composites is shown in Fig. 1. 3.1. Graphite samples Table 2 summarizes main characteristics of the three types of commercially available powder graphites which have been employed within this study – natural graphite (NG) and two synthetic ones (SG, TG). Generally, the graphites composed of particles formed of mutually parallel graphene layers with basal repeat distance of 0.337 nm (NG and SG) or 0.338 nm (TG). This distance is commonly regarded as the interplanar spacing corresponding to the van der Waals diameter of carbon [31]. The multilayered structure with densely packed nanosheets is clearly visible in SEM photographs of all the graphite samples (Fig. 2). In average, the synthetic graphites show almost double particle size than the natural graphite.

Fig. 1. General scheme of the preparation of the natural graphite (NG) intercalation compounds (NG1, NG2) and the related polymer composites (PNG, PNG1, PNG2).

434

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

Table 2 Main characteristics of the pristine graphites used in this study. Graphite code

Graphite type

Particle size (µm)

Content of carbon (%)

Interlayer spacing (nm)

NG SG TG

Natural (Koh-i-noor) Synthetic graphite Grade 8028B TIMREX® KS75

30.0 (d90) 53.0 (d95) 55.8 (d90)

≥96.0 ≥99.5 ≥99.3

0.337 0.337 0.338

Fig. 2. SEM images of the neat graphites (NG, SG, TG; top row) and their related G2-intercalates (NG2_A, SG2, TG2; bottom row).

3.2. Graphite intercalation compounds (GICs) Changes in the structural arrangement of the pristine graphites after their intercalation and subsequent ion-exchange reaction were monitored by XRD, which is a widely used method for the characterization of graphite-based samples. As can be clearly seen from Fig. 3, the extent of modification differed significantly for the three original graphites. The (0 0 2) high angle reflection of the neat NG (2θ = 26.4°, d ∼ 0.337 nm) split in two diffraction maxima after reaction with sodium and ethylenediamine, and a new sharp (0 0 1) reflection appeared at low angles, suggesting that the intercalation of the [Na(en)]+ complex yielded ordered first-stage NG1 GIC with remarkably larger basal spacing (2θ = 12.6°, d ∼ 0.701 nm). This is in agreement with Sirisaksoontorn et al., who suggested direct intercalation of the [Na(en)]+ complex within the graphene sheets [1]. The first modification step of the both synthetic graphites led to the shifting and broadening of the (0 0 2) reflection (2θ = 25.8° for both SG1 and TG1, d ∼ 0.345 nm). A low angle (0 0 1) reflection appeared in both samples. However, whereas the maximum of this reflection in TG1 was relatively sharp (2θ = 12.6°, d ∼ 0.701 nm), in the case of SG1 this reflection exhibited a strongly diffuse character, corresponding to a broad distribution of interlayer spacing, which indicates the presence of non-uniformed, only partially intercalated particles [2]. Simultaneously, NG1 GIC contained ca 20 wt.% of the [Na(en)]+ complex as determined by TGA, whereas in both synthetic SG1 and TG1 the [Na(en)]+ content was found to be only 14 and 15 wt.%, respectively (Table S1). These data correlate well with the XRD results, indicating that the process of the complex intercalation proceeded to a slightly larger extent when natural graphite was used as a starting material. This observation could be explained by a generally known approach, according to which the extent of modification reactions of the graphites based on redox processes is dependent on the graphite type, in particular on the size of the graphite particles and their active surface area, defined as a cumulated surface area of the different types of defects present on the carbon surface (stacking faults, single or multiple vacancies, dislocations) [32,33]. Since these sites were related to be responsible for the interactions with the species adsorbed on the carbon surface, different reactivity of the used natural and synthetic graphites during the sodium-en modification can be expected. Prior the preparation of NG2, SG2 and TG2 intercalates, reaction conditions for the ion-exchange between [Na(en)]+ and MQS were optimized for NG1 as a starting material. Within these preliminary experiments, molar ratio of MQS: [Na(en)]+ in the reaction feed as well as reaction time were varied (see Table 1). The obtained products were characterized with the use of XRD technique; the influence of the reaction conditions on the XRD patterns of the resulting intercalates is presented in Fig. 4. As follows from the figure, molar ratio 5:1 of MQS to the complex and reaction time 24 h yielded product (NG2_C) having a bimodal distribution with two distinct maxima (2θ = 25.9°, d ∼ 0.344 nm; and 2θ = 25.4°, d ∼ 0.351 nm), suggesting that the ion-exchange reaction proceeded to a certain extent only. Prolongation of reaction time (NG2_D) contributed positively to the yield of the reaction, however, the sample 435

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

Fig. 3. XRD patterns of the pristine graphites (NG, SG, TG) and their related G1 and G2 intercalation compounds.

Fig. 4. XRD patterns of the NG2 intercalates prepared under different conditions.

436

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

still exhibited a bimodal distribution. Increase in molar ratio MQS/[Na(en)]+ to 10:1 (NG2_A) lead to a significant narrowing of the reflection; the XRD curve remained almost unchanged after reaction time prolongation (NG2_B). These results were further supported by calculation of the intercalated compounds content using TGA (Table S1). As follows, the highest content (28 wt.%) was found in the NG2_A sample, while NG2_C and NG2_D intercalates exhibited a remarkably lower content of the intercalated species (18 and 20 wt.%, respectively). Therefore, SG2 and TG2 GICs were subsequently synthesized under conditions corresponding to the preparation of NG2_A sample. XRD curves for the G2 intercalates are shown in Fig. 3. As previously mentioned, visible narrowing of the (0 0 2) reflection can be observed in the NG2_A curve after the ion-exchange reaction with MQS; the (0 0 1) maximum was slightly shifted to low angles (2θ = 12.5°, d ∼ 0.708 nm), which is an agreement with our previous study [29]. On the other hand, WAXS curves of SG1 as well as TG1 remained almost unchanged after the reaction with MQS monomer, indicating that the ion-exchange proceeded to a negligible extent only. This finding is in a good agreement with TGA calculations (Table S1), giving the amount of intercalated compounds in SG2 and TG2 almost the same as in initial SG1 and TG1. SEM images confirmed that the layered morphology was maintained after modification reactions of the graphites (Fig. 2). Sirisaksoontorn et al., who synthesized GICs successfully intercalated by tetraalkylammonium salts via ion-exchange reaction, observed curving of the edges and an increase in the particles delamination when compared to the pristine graphite [1,4]. Cooper et al. described a distortion of the graphite planes by micron-sized pores, created by the electrochemical intercalation of both tetrabutylammonium and tetraethylammonium salts into the graphite galleries [34]. With respect to these studies, fan-shaped intercalated stacks, single or multiple rippled sheets and curving of the sheet edges, which were observed in all three G2 GICs, gave us an evidence that a chemical modification of the initial graphites took place. However, the above mentioned results based on the XRD and TGA data clearly show that these individual G2 GICs differ significantly in their structure and chemical composition, proposing NG2_A intercalate as the best candidate for the preparation of intended polymer composite membranes due to the highest content of intercalated MQS. 3.3. Polymer composite membranes First, neat copolymer samples were synthesized by free radical copolymerization of HEMA and 2-EHA at different ratios in the feed of polymerization reaction. We have observed that content of 2-EHA units in the feed lower than 30 mol.% resulted in a filmforming, but too brittle copolymer, whereas content of 2-EHA higher than 30 mol.% in the feed led to a non-film-forming and sticky copolymer. The optimal molar ratio 2-EHA/HEMA in the feed was therefore found to be 30/70 with respect to consistency and mechanical stability of the resulting copolymer sample being solid, film-forming and flexible at room temperature, thus suitable for the intended polymer membrane fabrication. Considering the results obtained from XRD, TGA and SEM techniques, NG2_A intercalate was used for the preparation of 2-EHA/ HEMA polymer composites. This intercalate bearing ion-bonded polymerizable methacrylamide group was copolymerized with 2EHA and HEMA monomers at various weight ratios, yielding PNG2 polymer composites with different content of the filler. Within the comparative study, polymer composites containing natural graphite and its related NG1 GIC were also synthesized. Membranes from the obtained polymer composites were prepared by a solution casting method. Since the properties of polymer nanocomposites depend strongly on how well they are dispersed, the selection of the used solvent was first optimized to meet the requirements for both polymer solubility and homogenous dispersion of the filler (NG2_A) in the matrix. Due to the polar nature of the polymer matrix mostly containing hydrophilic HEMA units, polar solvents were tested. It was found that the homogeneity of the filler dispersion increases accordingly to increasing dielectric constant of the solvent used for the membrane fabrication. Photographs of the polymer composite membranes are shown in Fig. S1. Whereas the filler in the membranes casted from THF (a) and acetone (b) solution exhibited certain heterogeneities in dispersion (visible as stripes or thin lines), a desired homogenous dispersion was observed when the membrane was casted from DMF (c), as a result of interaction between hydrophilic matrix and ionic groups in the GIC with this highly polar solvent. DMF was used also for the preparation of membranes from PNG and PNG1 polymer composites. The uniformity in the dispersion of natural graphite (d) as well as NG1 GIC (e) in the 2-EHA/HEMA polymer matrix was, however, significantly worse in comparison with NG2_A GIC, indicating that the ion-exchange step in the intercalate synthesis is necessary to ensure a completely homogeneous filler dispersion. A more detailed insight into the structure of the prepared polymer composites membranes was provided by XRD technique (Fig. 5). All the presented WAXS curves exhibited two distinct maxima at low angles (2θ = 10.4° and 18.0°), corresponding to the van der Waals contact of non-bonded atoms and to more complex atomic interactions within the 2-EHA/HEMA polymer matrix [35]. Significant differences in XRD patterns of the composites were observed in dependence on the filler used for their preparation. When 2-EHA and HEMA were copolymerized in the presence of natural graphite, the resulting PNG_1.0 and PNG_2.5 composites showed only one sharp maximum at high angles (2θ = 26.4°), corresponding to the same basal spacing observed for the pristine graphite (d ∼ 0.337 nm). This confirms that the filler still consists of multilayer graphite sheets, indicating that neither intercalation nor exfoliation occurred during the process of polymerization. Similar observations were recently reported by Yasmin et al. and Bhowmick et al. [36–38]. In accordance with this conclusion, TEM of the PNG composite showed only the presence of large aggregates of graphite particles barely compatible with the polymer matrix (Fig. 6a). Composites prepared by 2-EHA and HEMA copolymerization in the presence of NG1 GIC (PNG1_1.0 and PNG1_2.5) exhibited three distinct maxima, whose positions (2θ = 12.6°, 25.5° and 26.0°) corresponded exactly to the maxima of the initial NG1 intercalate (see Fig. 3). It should be noted that the low-angle reflection is present even in the sample with the lowest filler loading (1 wt. %). A closer examination revealed that the relative intensity of the maximum at 2θ = 25.5° increased significantly after the process of 437

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

Fig. 5. XRD patterns of the neat polymer matrix and related polymer composites with different filler loading. Intensities in the low-angle area were multiplied by ten for better visibility.

Fig. 6. TEM images of the polymer composites prepared by 2-EHA/HEMA copolymerization in the presence of natural graphite (a. PNG_1.0); NG1 intercalate (b. PNG1_1.0); and NG2_A intercalate (c.+d. PNG2_1.0; e. PNG2_5.0, which was obtained by filtration of the PNG2_5.0 solution in THF through 0.2 µm filter). Thickness of the membranes was 100 ± 20 µm.

polymerization. This observation together with the persisting low-angle maximum at 2θ = 12.6° in PNG1 composites curves suggests that a certain portion of methacrylate monomers or oligomers was intercalated between the modified graphite sheets during the polymerization, giving predominantly un-exfoliated structures dispersed in the polymer matrix. Similar behavior was described previously by Xiao et al., who have obtained polymer intercalated nanocomposites by the polymerization reaction of styrene in the presence of potassium metal modified GIC [39]. Xiao has revealed that most of the monomer polymerized on the surface of the GIC, while a small part of the monomer was intercalated into the interlayer space of graphite and polymerized therein, causing an increase in interlayer spacing when compared to the graphite. The representative TEM micrograph of the PNG1_1.0 composite shows unexfoliated GIC microparticles which are poorly compatible with the polymer matrix, causing clearly visible tearing of the matrix (Fig. 6b). In contrast, different behavior was found for the composites prepared with the NG2_A intercalate containing ion-bonded polymerizable methacrylamide group. Disappearing of the low-angle maximum (2θ = 12.5°) in XRD patterns of the PNG2 composites having 0.5–2.5 wt.% of the filler indicates that the methacrylamide double bonds participated successfully in a formation of copolymer chains, which subsequently led to the exfoliation of the layered structure of the filler. The presence of the single graphene nanosheets in the PNG2_1.0 composite was confirmed by TEM, displaying also well-dispersed graphene nanoparticles coated with polymer matrix (Fig. 6c and d) [40,41]. However, a certain amount of the high-stage intercalate remained un-exfoliated in all PNG2 composites, which is visible as a broad reflection in the high-angle area (2θ = 25.5°). Interestingly, a sharp low-angle reflection (2θ = 12.5°) occurred in the PNG2_5.0 composite with the highest filler loading (5.0 wt.%), suggesting that besides the exfoliated structures the initial first-stage intercalate is also present in the composite sample. This phenomenon could suggest that there is a certain limitation in the process of the filler exfoliation with respect to the filler content, in accordance with Ruoff et al., who prepared composites of poly(methyl methacrylate) by in situ radical polymerization in the presence of GO or RGO at various filler loadings. Both types of the filler underwent exfoliation during the process of polymerization, however, the extent of exfoliation and dispersion was provably dependent on the initial filler content. At higher filler loadings (4 wt.%), the composites appeared to show an increased number of multi-layered platelets with a larger average thickness [42]. Based on the XRD results we have assumed that the modification of the graphite by MQS yielded probably exfoliated structures when copolymerized with 2-EHA and HEMA, suggesting that the methacrylamide group in MQS participated successfully in the copolymerization process, resulting subsequently in the exfoliation of the graphite structures. To support this assumption, NMR study 438

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

Fig. 7. 1H NMR spectra (THF-d4, 300.1 MHz) of the neat 2-EHA/HEMA matrix and the nanocomposite prepared by a filtration of PNG2_5.0 solution in THF through 0.2 µm filter.

of the PNG2 composites was further carried out. The composites were first “dissolved” in THF, high-stage and first-stage intercalates were removed by filtration through a 0.2 µm filter, and THF was completely evaporated. The presence of homogenously dispersed graphene sheets covered by polymer chains in the PNG2_5.0 composite after filtration was confirmed by TEM (Fig. 6e). 1H NMR spectra of the filtered composites as well as the neat 2-EHA/HEMA matrix were recorded (Fig. 7). It was found that the molar ratio of 2-EHA and HEMA units incorporated in the neat copolymer matrix was similar to that in the composites (35/65), i.e. the 2-EHA/ HEMA ratio was not influenced by the presence of the filler. Moreover, new signals at 5.31, 4.02, 2.71, 2.18, 1.38, 1.27 and 0.86 ppm appeared in the composite spectrum when compared to the neat matrix, corresponding to the presence of MQS in the composite. Considering the absence of two singlets at 5.75 and 5.40 ppm in the spectrum of the composite, which correspond to the signals of protons from CH2] bond in the unreacted methacrylamide monomer [29], we can conclude the monomer MQS was incorporated into the structure of the copolymer chains. Taking into account that the MQS monomer in the initial intercalate was ion-bonded to the graphite interlayers, we can further speculate the exfoliated graphene structures in the composite were attached to the copolymer chains. This conclusion is in accordance with Marand et al., who has prepared polymer nanocomposites via free radical copolymerization of n-butyl methacrylate in the presence of montmorillonite previously modified by monomer [2-(acryloyloxy)ethyl]trimethylammonium chloride [43]. During the copolymerization process, the in situ exfoliation of the filler took place, yielding nanocomposites containing dispersed nanoclay platelets attached by ionic bond to the copolymer chains.

3.4. Properties of polymer membranes 3.4.1. Gas transport properties Gas transport properties of the polymer membranes containing various initial NG2_A filler loading (from 0 wt.% to 2.5 wt.%) were measured on the time-lag apparatus [31]. Gases and their individual pairs were selected with respect to the most common gas separation applications, such as natural gas treatment (CO2/CH4), nitrogen production from air (O2/N2), flue gas treatment (CO2/N2), hydrogen recovery (H2/N2) and biohydrogen production (H2/CO2) [44–49]. The obtained permeability, diffusion and solubility coefficients as well as their corresponding selectivities of selected gas pairs are presented in Fig. 8. For better clarity, the accurate values for the individual quantities are also listed in Tables S2–S4. The composite having 5.0 wt.% of NG2_A in the feed was found to be unsuitable for the measurement of gas transport properties since this high filler content contributed significantly to the worsening 439

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

Fig. 8. Gas transport properties for the 2-EHA/HEMA membrane and corresponding PNG2 composite membranes with various initial NG2_A loading.

of compatibility of the matrix and the dispersed filler; as a result, microscopic defects (microcracks, pinholes) occurred in the membrane, and the composite was thus excluded from the permeation measurements. The neat 2-EHA/HEMA matrix exhibited relatively low gas permeability coefficient values, comparable to other polymer materials based on (meth)acrylates [50]. Gas permeability coefficients for the matrix as well as for all the measured PNG2 composites showed the following increasing trend for individual gases: N2 < CH4 < O2 < CO2 < H2. This order indicates that the gas transport is driven preferably by diffusion than by solubility, especially for non-polar gases. This observation is in accordance with DSC results (see below), demonstrating that the permeability measurements were performed at the temperature significantly below the Tg values of the polymer samples. As can be also seen from Fig. 8, a dramatic decrease in permeabilities for all measured gases occurred in composites with increasing amount of NG2_A. Clearly, even very low content of the filler (1.0 wt.%) caused a decrease in permeability coefficients in tens of percent for all gases, for example 45% decrease in the H2 permeability coefficient and 58% decrease in CH4 permeability coefficient, when compared to the neat polymer matrix. This significant reduction in permeabilities caused by very low filler loading can be explained by the enlargement of the surface area of the initial multi-layered filler after its exfoliation to single or few-layered graphene sheets during the process of (co)polymerization. The strong barrier properties of the exfoliated graphene structures in the polymer composite thus consist in the considerable increase in tortuosity of the gas molecule pathways, since graphene nanosheets are commonly known to be completely impermeable for any gas molecules [15,17]. The presence of the graphene platelets also affected the ideal selectivities of the composites (Fig. 8). The strongest effect was observed for CO2/CH4 gas pair, where 2.5 wt.% of the filler caused 76% increase in ideal selectivity value when compared to the neat polymer matrix. This significant increase can be explained by the differences in diffusion and solubility coefficients for both gases. It is known that diffusion coefficients are dependent on the kinetic diameter of the molecules. The kinetic diameter of CO2 (0.33 nm) is lower than that of CH4 (0.38 nm), therefore, CO2 molecules can relatively easily diffuse through the highly tortuous path when compared to CH4 molecules. Moreover, CO2 exhibited about 2.5 times higher solubility in the nanocomposites than CH4, which further contributed positively to CO2 separation. In order to compare the extent of barrier properties for each set of the synthesized polymer composites, permeabilities were measured also for the composites containing NG and NG1 filler. Fig. 9 presents a comparison of the obtained results obtained for the neat matrix and for the composite membranes with initial 1.0 and 2.5 wt.% of NG, NG1 and NG2_A loading. For better clarity, individual data are summarized in Table S5. It can be seen that the unfilled 2-EHA/HEMA matrix exhibited the highest permeabilities for most of the gases when compared to the composites (Fig. 9a), apparently due to the absence of restrictions represented by any filler particles. For each filler type, permeability coefficients generally showed a decreasing trend with increasing filler loading. For 1.0 wt.% loading the barrier effect of NG2_A filler was generally more significant than for NG and NG1 fillers. On the other hand, composites with 2.5 wt.% of NG and NG1 particles exhibited significantly lower permeabilities of CO2 than the composite with the same content of mainly exfoliated NG2_A particles. Interestingly, the CO2/CH4 and CO2/N2 ideal selectivities significantly increased. Relatively high CO2 permeabilities of the PNG2 composite may be a consequence of a higher sorption capacity of the exfoliated filler 440

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

Fig. 9. Permeability coefficients (a) and ideal selectivities (b) for the polymer composites in dependence on the filler type and concentration.

when compared to the non-exfoliated one; nevertheless, further investigation is needed to support this idea. It should be also pointed out here that the real content of graphite/graphene structures in the PNG2_2.5 composite is significantly lower (1.8 wt.%) than those in the corresponding PNG_2.5 and PNG1_2.5 composites (2.5 wt.% and 2.0 wt.%, respectively; see Table 4), which makes the effect of the filler on CO2 permeabilities even more favorable for the NG2_A particles. Clearly, PNG as well as PNG1 composites exhibited significantly lower values of ideal selectivities for selected gas pairs when compared to the PNG2 composites with the same filler loading (Fig. 9b). This phenomenon may be caused by lower compatibility of polymer matrix and dispersed particles and/or by the presence of the agglomerating particles, potentially leading to a formation of additional Knudsen flow voids for gas permeation through such membrane [51]. As follows from this comparative study, the PNG2 composites containing well-dispersed NG2_A particles provided almost the lowest permeability coefficients and the highest ideal selectivities from all the presented composite types. These strong barrier properties indicate that the initial NG2_A filler is present in an exfoliated form, which supports the abovementioned results from other methods. Table 3 summarizes the results of some other works describing the effect of the addition of various graphene-like fillers into a polymer matrix on the gas transport properties of the resulting composite materials in comparison with the neat matrix. Different strategies in the composites preparation have also been taken into account, e.g. simple mixing of the filler with the polymer in solution, processing of the filler in a melt or in situ exfoliation of the filler during the polymerization (this work). For example, a membrane cast from solution of polystyrene and graphene particles (2.3 vol%) exhibited an oxygen permeability of 1.84 Barrer, representing 61% decrease in permeability when compared to the neat polystyrene membrane [52]. As follows from the Table 3, reduction of permeabilities for selected gases in the listed works ranges from 25 to 97% for composites containing up to 3.0 wt.% of the filler; these values are fully comparable with the results obtained in this study. Namely, 1.0 wt.% of NG2_A filler caused 57% decrease in oxygen permeability and 49% decrease in nitrogen permeability, in the case of 2.5 wt.% loading the permeability reduction for the oxygen and nitrogen was 52% and 55%, respectively. From separation efficiency perspective, our materials exhibited standard performance, because main separation contribution in

Table 3 Comparison of the effect of the addition graphene-like fillers into different matrices on the reduction of permeabilities. Polymer matrix

Filler type

Filler content

Processing

Gas

Permeability (Barrer)

Reduction (%)

Ref.

PS PS

Graphene GO

2.3 vol% 2.0 wt.%

solution In situ

Functionalized GO GO Graphene Reduced GO Intercalated graphite (NG2_A)

3.0 wt.% 1.0 wt.% 0.5 wt.% 2.0 wt.% 1.0 wt.%

Solution Solution In situ Melt In situ

1.84 2.24 0.43 0.0014 0.97 0.81 0.041 0.31 0.08 0.15 0.35 0.07 0.11

61 25 59 97 79 70 91 57 49 59 52 55 70

[52] [53]

PET PU PMMA PU 2-EHA/HEMA copolymer

O2 O2 N2 O2 He O2 O2 O2 N2 CH4 O2 N2 CH4

2.5 wt.%

441

[54] [55] [56] [57] This work

This work

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

Fig. 10. Electrical conductivity at frequency 0.01 Hz of PNG, PNG1 and PNG2 polymer composites as a function of the filler concentration (a), and frequency dependence of electrical conductivity for PNG2 polymer composites (b).

the prepared composite materials comes from the matrix material, belonging in our case to the class of polymers with common gas transport behavior.

3.4.2. Electrical conductivity Electrical conductivity measurement is a powerful method providing besides the determination of the material conductivity also a possibility to estimate the filler morphology (aggregating or isolated particles) and its distribution in the material. Generally, to obtain a conductive/semiconductive composite material consisting of a highly conductive filler and nonconductive matrix, the specific percolation concentration of the filler has to be reached. At this concentration, the filler starts to create interconnected aggregates/agglomerates, which can form conductive pathways through the matrix, causing a significant decrease in the composite bulk resistivity. Electrical volume conductivities of the polymer composites with respect to the filler type and its concentration are displayed in Fig. 10a. Neat polymer matrix exhibited conductivity around 1.6 × 10−12 S/m, lying in the range of typical values for isolating polymer materials based on poly(meth)acrylates [6]. Addition of 1 wt.% of NG led to a significant increase in the composite conductivity (2.3 × 10−1 S/m), indicating that this NG content is already beyond a percolation threshold. Here, dispersion of bulky graphite particles in the polymer matrix ensures a good connection across the membrane, forming conductive pathways. A noticeable drop in volume conductivity (1.8 × 10−4 S/m) was observed for higher NG loading (2.5 wt.%), which suggest an interruption of the conductive pathways across the membrane resulting in the decrease in the overall volume conductivity. This effect can be explained by the sedimentation process of the filler aggregates at this relatively high NG content, obviously leading to a higher concentration of the filler particles in the bottom part of the membrane when compared to the upper part. The influence of sedimentation process was confirmed by surface conductivity measurement (not shown here) which revealed a few orders higher conductivity for bottom part of membrane compare to the top one. Composite containing 1 wt.% of the NG1 intercalate exhibited relatively high value of volume conductivity (5.7 × 10−3 S/m), suggesting that this filler content is also already above the percolation threshold. Contrary to the composites with unmodified graphite, rising the NG1 content up to 2.5 wt.% led to about two orders increase in the composite conductivity (2.5 × 10−1 S/m), which is a consequence of the higher amount of dispersed conductive aggregates forming a denser pathway for the charge transport across the membrane. In comparison with the PNG1 composites, samples containing NG2_A as a filler exhibited lower volume conductivities. This can be a consequence of at least two factors influencing the composites conductivity, in particular (i) lower content of the conductive graphite/graphene structures in the PNG2 composites when compared to the PNG1 samples having the same filler content (see Table 4), and (ii) in-plane distribution of smaller aggregates in PNG2 contrary to bigger aggregates in PNG1 (Figs. S1 and 6). A small drop in conductivity at the lowest filler concentration can be explained by a doping effect of well-distributed nanoparticles as it was already described for the polyethylene/graphene nanoparticles composite [58]. BDS measurements also revealed that the NG2_A percolation concentration lay between 1.0 and 2.5 wt.% (Fig. 10b); taking into account a fluctuation of conductivity with frequency for 2.5 wt.%, we can conclude the threshold is close to this value. At the same time, the increase in volume conductivity at low filler content and low frequency indicated a pronounced ionic conductivity originated from quaternary ammonium salts attached to graphene platelets [59].

442

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

Table 4 DSC (Tg) and TGA (Td5%, Td10%, and char yield at 600 °C) results of neat polymer membrane and composite membranes with different amounts of natural graphite, NG1 and NG2_A intercalates. Sample

Filler type

Filler conc. (wt.%)

Tg (°C)

Td5% (°C)

Td10% (°C)

Char yield at 600 °C (%)

Matrix PNG_1.0 PNG_2.5 PNG1_1.0 PNG1_2.5 PNG2_0.05 PNG2_0.25 PNG2_1.0 PNG2_2.5 PNG2_5.0

– NG

0 1.0 2.5 1.0a (0.8) 2.5a (2.0) 0.05a (0.036) 0.25a (0.18) 1.0a (0.72) 2.5a (1.8) 5.0a (3.6)

58 64 44 46 48 65 69 70 63 68

321 329 344 293 308 330 331 335 316 307

341 356 368 347 336 347 349 359 346 337

1.1 3.4 6.6 2.0 5.0 1.4 1.7 2.5 12.4 12.9

NG1 NG2_A

Td5%, Td10% – the temperature of 5% and 10%, respectively, of the total sample mass loss (from TGA). a Content of GIC (NG1 or NG2_A). The total content of the graphite/graphene is given in brackets (determined from TGA).

3.4.3. Thermal properties Thermal stability of the prepared nanocomposites plays also an important role in the potential membrane applicability; therefore, DSC and TGA measurements were performed to determine thermal behavior of the prepared polymer composite membranes. Generally, Tg value of polymer composite membranes is highly affected by (i) content of residual low molecular weight components in the composite, (ii) conditions for the polymer preparation, (iii) physicochemical properties of the polymer (changes in tacticity and/or molecular weight), and (iv) interactions of the polymer chains with surface of the filler particles [60]. Here, the absence of undesired low molecular weight components (e.g. solvents or monomer, oligomer residues) both in the neat polymer and polymer composites was carefully checked by NMR and TGA methods. All membranes were prepared precisely under the same conditions, allowing the comparison of individual samples. In our developed PNG2 system, three types of comonomers participate in the polymer formation, in which the content of incorporated MQS is too low for being considered in polymer tacticity calculation. Moreover, as follows from the aforementioned NMR and TEM study on the PNG2 composites, a certain portion of the exfoliated graphene structures is bonded to terpolymer chains, which does not allow precise determination of terpolymer molecular weights and their comparison to that of the neat polymer matrix. On the other hand, ion-bonding of graphene platelets to polymer chains causes desired hardening of the polymer, resulting subsequently in the increase of Tg. With respect to the mentioned reasons it can be therefore assumed that Tg of the composites will be mainly affected by the interactions between the surface of the filler and polymer matrix. Tg values of the neat 2-EHA/HEMA membrane and related polymer composite membranes with different types of the filler are summarized in Table 4. The increased Tg values of the PNG2 composite membranes containing initial NG2_A filler in the range of 0.05–1 wt.% suggest the effective filler particle-polymer interactions that prevents the segmental motions of the polymer chains [61,62]. The lack of further increase beyond 2.5 wt.% originates from the decreased inter-particle distance and strong filler-filler interactions promoting filler agglomeration [63]; the presence of the filler aggregates or agglomerating filler particles at higher filler loading is in a good agreement with the results obtained from XRD, TEM and gas permeability measurements. The resulting Tg value (still higher than Tg of the neat matrix) was thus influenced by two antagonistic effects: the restricted copolymer mobility due to polymer-filler surface bonding and the increased free volume of polymer chains due to the presence of less compatible filler agglomerates. Yuan et al. reported the similar trend in Tg evolution of graphene/PMMA nanocomposites, where the Tg increased after addition of 0.07 wt.% graphene but no further increment was exhibited beyond this loading [64]. Much higher filler incompatibility with the polymer matrix and increased tendency to form filler agglomerates were, however, observed in the cases of composite membranes filled with NG graphite and NG1 intercalate. The phase separation and incompatibility with the matrix was even macroscopically visible (see photos in Fig. S1). The lack of the filler-polymer interactions and agglomerate formation decreased significantly Tg of PNG and PNG1 sample series (Table 4). Thermal stability of the prepared membranes was evaluated using TGA under nitrogen atmosphere (Table 4). All samples (neat matrix as well as composite membranes) exhibited similar weight one-step loss TG curve profile. The TGA results (expressed as Td5%, Td10% and% char yield at 600 °C in Table 4) showed relatively high thermal stability of the prepared nanocomposite membranes. Thermal stability of PNG2 membranes containing up to 1.0 wt.% of the filler was increased in comparison to the neat polymer matrix (Fig. S2). The improved thermal stability might be attributed to inorganic (GIC) phase homogeneously dispersed in the polymer matrix which acted as a barrier for the heat transfer from surroundings into the polymer matrix [65]. At higher filler loadings, the thermal stability slightly decreased due to agglomerates formation. The barrier effect of the intercalate can be also seen from the much higher char yield at 600 °C in the case of PNG2 membranes when compared to those filled with NG or NG1 intercalate. Thermal stability shifted to higher temperature was also observed in the case of PNG membranes containing neat graphite. Contrary to that, decreased thermal stability of PNG1 membranes when compared to the neat matrix might be due to the pyrolysis of intercalated [Na (en)]+.

443

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

4. Conclusions Three types of graphite (natural and two synthetic ones) were used for the [Na(en)]+ intercalation and subsequent ion-exchange reaction with quaternary ammonium salt possessing methacrylamide group. Using XRD and TGA it was found that both reactions proceeded successfully in the case when natural graphite was employed as a starting material. Further, polymer composites were prepared by free radical copolymerization of 2-EHA and HEMA in the presence of both GICs derived from natural graphite. The process of the filler in situ exfoliation was significantly dependent on the type of the filler and its concentration. Below 2.5 wt.% content of the GICs bearing methacrylamide group, in situ exfoliated graphene nanoplatelets were homogenously dispersed in the polymer matrix. NMR and TEM study suggested that the in situ exfoliated graphene nanoplatelets were attached to copolymer chains. These composite membranes exhibited almost the lowest permeability coefficients and the highest ideal selectivities from all the presented composite types, confirming that the initial GICs was present in an exfoliated form. An increase in the exfoliated filler content caused a dramatic increase in the value of the composite volume conductivity, revealing simultaneously that the percolation concentration of the exfoliated filler lay closely to 2.5 wt.%. Acknowledgements The authors gratefully thank to the Grant Agency of the Czech Republic (project 17-08273S) and Ministry of Education, Youth and Sports, National Sustainability Program I - NPU I (project POLYMAT LO1507) for financial support. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.eurpolymj. 2017.07.033. References [1] W. Sirisaksoontorn, M.M. Lerner, Preparation of a homologous series of tetraalkylammonium graphite intercalation compounds, Inorg. Chem. 52 (2013) 7139–7144, http://dx.doi.org/10.1021/ic400733k. [2] W. Sirisaksoontorn, M.M. Lerner, The effect of surface passivation on the preparation and stability of the graphite intercalation compounds containing tetra-n-alkylammonium cations, Carbon N. Y. 69 (2014) 582–587, http://dx.doi.org/10.1016/j.carbon.2013.12.070. [3] M. Zhou, T. Tian, X. Li, X. Sun, J. Zhang, P. Cui, J. Tang, L.C. Qin, Production of graphene by liquid-phase exfoliation of intercalated graphite, Int. J. Electrochem. Sci. 9 (2014) 810–820. [4] W. Sirisaksoontorn, A.A. Adenuga, V.T. Remcho, M.M. Lerner, Preparation and characterization of a tetrabutylammonium graphite intercalation compound, J. Am. Chem. Soc. 133 (2011) 12436–12438, http://dx.doi.org/10.1021/ja2053539. [5] P. Vecera, J.C. Chacón-Torres, T. Pichler, S. Reich, H.R. Soni, A. Görling, K. Edelthalhammer, H. Peterlik, F. Hauke, A. Hirsch, Precise determination of graphene functionalization by in situ Raman spectroscopy, Nat. Commun. 8 (2017) 15192, http://dx.doi.org/10.1038/ncomms15192. [6] G. Chen, W. Weng, D. Wu, C. Wu, PMMA/graphite nanosheets composite and its conducting properties, Eur. Polym. J. 39 (2003) 2329–2335, http://dx.doi.org/10.1016/j. eurpolymj.2003.08.005. [7] K. Gotoh, C. Sugimoto, R. Morita, T. Miyatou, M. Mizuno, W. Sirisaksoontorn, M.M. Lerner, H. Ishida, Arrangement and dynamics of diamine, etheric, and tetraalkylammonium intercalates within graphene or graphite oxide galleries by 2H NMR, J. Phys. Chem. C 119 (2015) 11763–11770, http://dx.doi.org/10.1021/acs.jpcc. 5b03016. [8] D.-E. Liu, G. Xie, D. Guo, Z. Cui, L. Si, C. Wan, W. Zou, J. Luo, Tunable lubricity of aliphatic ammonium graphite intercalation compounds at the micro/nanoscale, Carbon N. Y. 115 (2017) 574–583. 10.1016/j.carbon.2017.01.049. [9] X. Li, G. Zhang, X. Bai, X. Sun, X. Wang, E. Wang, H. Dai, Highly conducting graphene sheets and Langmuir-Blodgett films, Nat Nano. 3 (2008) 538–542, http://dx.doi.org/ 10.1038/nnano.2008.210. [10] Z.-S. Wu, W. Ren, L. Gao, J. Zhao, Z. Chen, B. Liu, D. Tang, B. Yu, C. Jiang, H.-M. Cheng, Synthesis of graphene sheets with high electrical conductivity and good thermal stability by hydrogen arc discharge exfoliation, ACS Nano 3 (2009) 411–417, http://dx.doi.org/10.1021/nn900020u. [11] H. Chen, M.B. Müller, K.J. Gilmore, G.G. Wallace, D. Li, Mechanically strong, electrically conductive, and biocompatible graphene paper, Adv. Mater. 20 (2008) 3557–3561, http://dx.doi.org/10.1002/adma.200800757. [12] X. Xu, L.F.C. Pereira, Y. Wang, J. Wu, K. Zhang, X. Zhao, S. Bae, C. Tinh Bui, R. Xie, J.T.L. Thong, B.H. Hong, K.P. Loh, D. Donadio, B. Li, B. Özyilmaz, Length-dependent thermal conductivity in suspended single-layer graphene, Nat. Commun. 5 (2014) 3689, http://dx.doi.org/10.1038/ncomms4689. [13] A.a. Balandin, S. Ghosh, W. Bao, I. Calizo, D. Teweldebrhan, F. Miao, C.N. Lau, Superior thermal conductivity of single-layer graphene 2008, Nano Lett 8 (2008) 902–907, http://dx.doi.org/10.1021/nl0731872. [14] S. Alwarappan, S. Pillai, S.R. Singh, A. Kumar, Graphene-Based Biosensors and Gas Sensors, in: C. Wonbong, J.L. Lee (Eds.), Graphene Synth. Appl, CRC Press, Taylor & Francis Group, Boca Raton, FL, 2011, pp. 233–255. [15] A.M. Pinto, J. Cabral, D.A.P. Tanaka, A.M. Mendes, F.D. Magalhães, Effect of incorporation of graphene oxide and graphene nanoplatelets on mechanical and gas permeability properties of poly(lactic acid) films, Polym. Int. 62 (2013) 33–40, http://dx.doi.org/10.1002/pi.4290. [16] V. Berry, Impermeability of Graphene and its Applications, Carbon N. Y. 62 (2013) 1–10. 10.1016/j.carbon.2013.05.052. [17] Y. Cui, S.I. Kundalwal, S. Kumar, Gas barrier performance of graphene/polymer nanocomposites, Carbon N. Y. 98 (2016) 313–333, http://dx.doi.org/10.1016/j.carbon. 2015.11.018. [18] A. Naz, A. Kausar, M. Siddiq, Influence of graphite filler on physicochemical characteristics of polymer/graphite composites: a review, Polym. Plast. Technol. Eng. 55 (2016) 604–625, http://dx.doi.org/10.1080/03602559.2015.1098697. [19] H. Zhang, M.M. Lerner, Preparation of graphite intercalation compounds containing oligo and polyethers, Nanoscale 8 (2016) 4608–4612, http://dx.doi.org/10.1039/ C5NR08226A. [20] J.A. Pandit, K. Sudarshan, A.A. Athawale, Electrically conductive epoxy-polyester-graphite nanocomposites modified with aromatic amines, Polym (United Kingdom) 104 (2016) 49–60, http://dx.doi.org/10.1016/j.polymer.2016.09.084. [21] S.R. Kim, M. Poostforush, J.H. Kim, S.G. Lee, Thermal diffusivity of in-situ exfoliated graphite intercalated compound/polyamide and graphite/polyamide composites, Express Polym. Lett. 6 (2012) 476–484, http://dx.doi.org/10.3144/expresspolymlett.2012.50. [22] F.M. Uhl, Q. Yao, H. Nakajima, E. Manias, C.A. Wilkie, Expandable graphite/polyamide-6 nanocomposites, Polym. Degrad. Stab. 89 (2005) 70–84, http://dx.doi.org/10. 1016/j.polymdegradstab.2005.01.004. [23] S. Jiang, Z. Gui, C. Bao, K. Dai, X. Wang, K. Zhou, Y. Shi, S. Lo, Y. Hu, Preparation of functionalized graphene by simultaneous reduction and surface modification and its polymethyl methacrylate composites through latex technology and melt blending, Chem. Eng. J. 226 (2013) 326–335, http://dx.doi.org/10.1016/j.cej.2013.04.068. [24] J. Huang, Q. Tang, W. Liao, G. Wang, W. Wei, C. Li, Green preparation of expandable graphite and its application in flame-resistance polymer elastomer, Ind. Eng. Chem. Res. 56 (2017) 5253–5261, http://dx.doi.org/10.1021/acs.iecr.6b04860.

444

European Polymer Journal 94 (2017) 431–445

L. Poláková et al.

[25] X. Zeng, J. Yang, W. Yuan, Preparation of a poly(methyl methacrylate)-reduced graphene oxide composite with enhanced properties by a solution blending method, Eur. Polym. J. 48 (2012) 1674–1682, http://dx.doi.org/10.1016/j.eurpolymj.2012.07.011. [26] W. Zheng, S.C. Wong, Electrical conductivity and dielectric properties of PMMA/expanded graphite composites, Compos. Sci. Technol. 63 (2003) 225–235, http://dx.doi. org/10.1016/S0266-3538(02)00201-4. [27] H. Shioyama, K. Tatsumi, N. Iwashita, K. Fujita, Y. Sawada, On the interaction between the potassium—GIC and unsaturated hydrocarbons, Synth. Met. 96 (1998) 229–233, http://dx.doi.org/10.1016/S0379-6779(98)00098-8. [28] K. Xu, D. Erricolo, M. Dutta, M.A. Stroscio, Electrical conductivity and dielectric properties of PMMA/graphite nanoplatelet ensembles, Superlattices Microstruct. 51 (2012) 606–612, http://dx.doi.org/10.1016/j.spmi.2012.03.001. [29] L. Poláková, H. Beneš, P. Ecorchard, E. Pavlová, Z. Sedláková, J. Kredatusová, V. Štengl, Nanocomposite preparation via in situ polymerization of quaternary ammonium salt ion-bonded to graphite platelets, RSC Adv. 6 (2016) 353–357, http://dx.doi.org/10.1039/C5RA22419E. [30] S.W. Rutherford, D.D. Do, Review of time lag permeation technique as a method for characterisation of porous media and membranes, Adsorption 3 (1997) 283–312, http://dx.doi.org/10.1007/BF01653631. [31] R.R. Haering, Band structure of rhombohedral graphite, Can. J. Phys. 36 (1958) 352–362, http://dx.doi.org/10.1139/p58-036. [32] P. Novak, J. Ufheil, H. Buqa, F. Krumeich, M.E. Spahr, D. Goers, H. Wilhelm, J. Dentzer, R. Gadiou, C. Vix-Guterl, The importance of the active surface area of graphite materials in the first lithium intercalation, J. Power Sources 174 (2007) 1082–1085, http://dx.doi.org/10.1016/j.jpowsour.2007.06.036. [33] F. Béguin, F. Chevallier, C. Vix-Guterl, S. Saadallah, V. Bertagna, J.N. Rouzaud, E. Frackowiak, Correlation of the irreversible lithium capacity with the active surface area of modified carbons, Carbon N. Y. 43 (2005) 2160–2167, http://dx.doi.org/10.1016/j.carbon.2005.03.041. [34] A.J. Cooper, N.R. Wilson, I.A. Kinloch, R.A.W. Dryfe, Single stage electrochemical exfoliation method for the production of few-layer graphene via intercalation of tetraalkylammonium cations, Carbon N. Y. 66 (2014) 340–350, http://dx.doi.org/10.1016/j.carbon.2013.09.009. [35] D.J. Haloi, N.K. Singha, Synthesis of poly(2-ethylhexyl acrylate)/clay nanocomposite by in situ living radical polymerization, J. Polym. Sci. Part A Polym. Chem. 49 (2011) 1564–1571, http://dx.doi.org/10.1002/pola.24577. [36] J.J. George, A.K. Bhowmick, Ethylene vinyl acetate/expanded graphite nanocomposites by solution intercalation: preparation, characterization and properties, J. Mater. Sci. 43 (2008) 702–708, http://dx.doi.org/10.1007/s10853-007-2193-6. [37] A. Yasmin, J.J. Luo, I.M. Daniel, Processing of expanded graphite reinforced polymer nanocomposites, Compos. Sci. Technol. 66 (2006) 1179–1186, http://dx.doi.org/10. 1016/j.compscitech.2005.10.014. [38] A. Yasmin, I.M. Daniel, Mechanical and thermal properties of graphite platelet/epoxy composites, Polymer (Guildf) 45 (2004) 8211–8219, http://dx.doi.org/10.1016/j. polymer.2004.09.054. [39] M. Xiao, Synthesis and properties of polystyrene/graphite nanocomposites, Polymer (Guildf). 43 (2002) 2245–2248, http://dx.doi.org/10.1016/S0032-3861(02)00022-8. [40] M. Fang, K. Wang, H. Lu, Y. Yang, S. Nutt, Covalent polymer functionalization of graphene nanosheets and mechanical properties of composites, J. Mater. Chem. 19 (2009) 7098, http://dx.doi.org/10.1039/b908220d. [41] P. Song, Z. Cao, Y. Cai, L. Zhao, Z. Fang, S. Fu, Fabrication of exfoliated graphene-based polypropylene nanocomposites with enhanced mechanical and thermal properties, Polymer (Guildf) 52 (2011) 4001–4010, http://dx.doi.org/10.1016/j.polymer.2011.06.045. [42] J.R. Potts, S.H. Lee, T.M. Alam, J. An, M.D. Stoller, R.D. Piner, R.S. Ruoff, Thermomechanical properties of chemically modified graphene/poly(methyl methacrylate) composites made by in situ polymerization, Carbon N. Y. 49 (2011) 2615–2623, http://dx.doi.org/10.1016/j.carbon.2011.02.023. [43] J.M. Herrera-Alonso, Z. Sedlakova, E. Marand, Gas barrier properties of nanocomposites based on in situ polymerized poly(n-butyl methacrylate) in the presence of surface modified montmorillonite, J. Memb. Sci. 349 (2010) 251–257, http://dx.doi.org/10.1016/j.memsci.2009.11.057. [44] A. Tabe-Mohammadi, A review of the applications of membrane separation technology in natural gas treatment, Sep. Sci. Technol. 34 (1999) 2095–2111, http://dx.doi.org/ 10.1081/SS-100100758. [45] A.M.W. Hillock, W.J. Koros, R. V September, V. Re, M. Recei, V. No, Cross-Linkable Polyimide Membrane for Natural Gas Purification and Carbon Dioxide Plasticization Reduction 100 (2007) 0–4. [46] P. Bernardo, E. Drioli, G. Golemme, Membrane gas separation: a review/state of the art, Ind. Eng. Chem. Res. 48 (2009) 4638–4663, http://dx.doi.org/10.1021/ie8019032. [47] P. Bakonyi, N. Nemestóthy, K. Bélafi-Bakó, Biohydrogen purification by membranes: an overview on the operational conditions affecting the performance of non-porous, polymeric and ionic liquid based gas separation membranes, Int. J. Hydrogen Energy. 38 (2013) 9673–9687, http://dx.doi.org/10.1016/j.ijhydene.2013.05.158. [48] P. Bakonyi, G. Kumar, N. Nemestóthy, C.Y. Lin, K. Bélafi-Bakó, Biohydrogen purification using a commercial polyimide membrane module: studying the effects of some process variables, Int. J. Hydrogen Energy 38 (2013) 15092–15099, http://dx.doi.org/10.1016/j.ijhydene.2013.09.133. [49] V. Giel, B. Galajdová, D. Popelková, J. Kredatusová, M. Trchová, E. Pavlova, H. Beneš, R. Válek, J. Peter, Gas transport properties of novel mixed matrix membranes made of titanate nanotubes and PBI or PPO, Desalin. Water Treat. 56 (2014) 1–9, http://dx.doi.org/10.1080/19443994.2014.981931. [50] C.T. Wright, D.R. Paul, Gas sorption and transport in poly(tertiary-butyl methacrylate), Polymer (Guildf) 38 (1997) 1871–1878, http://dx.doi.org/10.1016/S00323861(96)00724-0. [51] W.J. Koros, G.K. Fleming, Membrane-based gas separation, J. Memb. Sci. 83 (1993) 1–80, http://dx.doi.org/10.1016/0376-7388(93)80013-N. [52] O.C. Compton, S. Kim, C. Pierre, J.M. Torkelson, S.T. Nguyen, Crumpled graphene nanosheets as highly effective barrier property enhancers, Adv. Mater. 22 (2010), http:// dx.doi.org/10.1002/adma.201000960. [53] Y.-H. Yu, Y.-Y. Lin, C.-H. Lin, C.-C. Chan, Y.-C. Huang, L. Li, S.H. Chan, J.H. Zhao, W.J. Ding, Y. Wei, J.M. Yeh, J. Park, R.S. Ruoff, High-performance polystyrene/graphenebased nanocomposites with excellent anti-corrosion properties, Polym. Chem. 5 (2014) 535–550, http://dx.doi.org/10.1039/C3PY00825H. [54] S.H. Shim, K.T. Kim, J.U. Lee, W.H. Jo, Facile method to functionalize graphene oxide and its application to poly(ethylene terephthalate)/graphene composite, ACS Appl. Mater. Interfaces 4 (2012) 4184–4191, http://dx.doi.org/10.1021/am300906z. [55] P. Kaveh, M. Mortezaei, M. Barikani, G. Khanbabaei, Low-temperature flexible polyurethane/graphene oxide nanocomposites: effect of polyols and graphene oxide on physicomechanical properties and gas permeability, Polym. Plast. Technol. Eng. 53 (2014) 278–289, http://dx.doi.org/10.1080/03602559.2013.844241. [56] C.-H. Chang, T.-C. Huang, C.-W. Peng, T.-C. Yeh, H.-I. Lu, W.-I. Hung, C.-J. Weng, T.-I. Yang, J.-M. Yeh, Novel anticorrosion coatings prepared from polyaniline/graphene composites, Carbon N. Y. 50 (2012) 5044–5051, http://dx.doi.org/10.1016/j.carbon.2012.06.043. [57] J.N. Gavgani, H. Adelnia, M.M. Gudarzi, Intumescent flame retardant polyurethane/reduced graphene oxide composites with improved mechanical, thermal, and barrier properties, J. Mater. Sci. 49 (2014) 243–254, http://dx.doi.org/10.1007/s10853-013-7698-6. [58] Z. Jing, C. Li, H. Zhao, G. Zhang, B. Han, Doping effect of graphene nanoplatelets on electrical insulation properties of polyethylene: from macroscopic to molecular scale, Materials (Basel) 9 (2016) 680, http://dx.doi.org/10.3390/ma9080680. [59] H.M. Ng, S. Ramesh, K. Ramesh, Exploration on the P(VP-co-VAc) copolymer based gel polymer electrolytes doped with quaternary ammonium iodide salt for DSSC applications: electrochemical behaviors and photovoltaic performances, Org. Electron. Phys., Mater. Appl. 22 (2015) 132–139, http://dx.doi.org/10.1016/j.orgel.2015.03. 020. [60] B.J. Ash, L.S. Schadler, R.W. Siegel, Glass transition behavior of alumina/polymethylmethacrylate nanocomposites, Mater. Lett. 55 (2002) 83–87, http://dx.doi.org/10. 1016/S0167-577X(01)00626-7. [61] H.J. Salavagione, G. Martínez, M.a. Gómez, Synthesis of poly(vinyl alcohol)/reduced graphite oxide nanocomposites with improved thermal and electrical properties, J. Mater. Chem. 19 (2009) 5027, http://dx.doi.org/10.1039/b904232f. [62] K.H. Liao, S. Aoyama, A.A. Abdala, C. Macosko, Does graphene change T g of nanocomposites? Macromolecules 47 (2014) 8311–8319, http://dx.doi.org/10.1021/ ma501799z. [63] J.S. Sefadi, A.S. Luyt, J. Pionteck, F. Piana, U. Gohs, Effect of surfactant and electron treatment on the electrical and thermal conductivity as well as thermal and mechanical properties of ethylene vinyl acetate/expanded graphite composites, J. Appl. Polym. Sci. 132 (2015) 1–10, http://dx.doi.org/10.1002/app.42396. [64] X.Y. Yuan, L.L. Zou, C.C. Liao, J.W. Dai, Improved properties of chemically modified graphene/poly(methyl methacrylate) nanocomposites via a facile in-situ bulk polymerization, Express Polym. Lett. 6 (2012) 847–858, http://dx.doi.org/10.3144/expresspolymlett.2012.90. [65] S.K. Yadav, J.W. Cho, Functionalized graphene nanoplatelets for enhanced mechanical and thermal properties of polyurethane nanocomposites, Appl. Surf. Sci. 266 (2013) 360–367, http://dx.doi.org/10.1016/j.apsusc.2012.12.028.

445