3-Arm star pyrene-functional PMMAs for efficient exfoliation of graphite in chloroform: fabrication of graphene-reinforced fibrous veils

Sandra S. Gkermpoura ab, Konstantinia D. Papadimitriou ab, Emmanuel N. Skountzos ab, Ioannis Polyzos b, Maria Giovanna Pastore Carbone b, Athanasios Kotrotsos c, Vlasis G. Mavrantzas abd, Costas Galiotis ab and Constantinos Tsitsilianis *a
aDepartment of Chemical Engineering, Universty of Patras, GR – 26504, Patras, Greece. E-mail: ct@chemeng.upatras.gr
bFoundation of Research and Technology Hellas, Institute of Chemical Engineering Sciences (FORTH/ICE-HT), Stadiou Str., P.O. Box 1414, GR 26504, Rio-Patras, Greece
cApplied Mechanics Laboratory, Department of Mechanical Engineering and Aeronautics, University of Patras, 26504, Rio-Patras, Greece
dParticle Technology Laboratory, Department of Mechanical and Process Engineering, ETH-Z, CH-8093 Zürich, Switzerland

Received 25th August 2018 , Accepted 1st October 2018

First published on 3rd October 2018

3-Arm PMMAs end-functionalized by pyrene were designed as dispersing/stabilizing agents for the liquid-phase exfoliation of graphite in low-boiling point solvents like chloroform. The synthetic procedure comprised ARGET ATRP controlled polymerization, click chemistry and the quaternization reaction of triazole, ensuring tailor-made, well-defined pyrene-functional star PMMAs. Among a series of different pyrene-functional macromolecular topologies, the (PMMA-py2)3 proved the most efficient exfoliation agent giving relatively high graphene concentration (0.36 mg ml−1) at exceptionally low polymer/graphite mass ratio (mP/mGF = 0.003) and short sonication time (3 h). A 5-cycle iterative procedure relying on the redispersion of the sediment was developed yielding CG = 1.29 mg ml−1 with 14.8% exfoliation yield, under the favorable conditions of 10.5 h total shear mixing/tip sonication time and overall mP/mGF ratio as low as 0.15. In parallel, all-atom molecular dynamics simulations were conducted which helped understand the mechanism by which pyrene-functional macromolecular topologies act as efficient dispersing agents of graphene. Finally the G@(PMMA-Py)3 hybrids were well dispersed into the PMMA matrix by electrospinning to fabricate graphene-based nanocomposite fibrous veils. These graphene/polymer nanocomposites exhibited enhanced stiffness and strength by a factor of 4.4 with 1.5 wt% graphene hybrids as nanofillers.

1. Introduction

Graphene, a single layer of sp2 bonded carbon atoms, has received great attention in a wide variety of research fields.1 Due to its extraordinary electronic, chemical, mechanical, thermal and optical properties, it is considered as a promising material in several applications.2–5 Graphene was initially isolated by mechanical exfoliation with the so called ‘scotch tape’ method.6 Current improvements in the production of graphene include epitaxial growth, chemical vapor deposition (CVD) and liquid-phase exfoliation.7,8 The latter method is scalable and enables the production of high quality graphene in different solvents so it is best suited for the production of multifunctional advanced materials such as graphene based composites.9–11 Nevertheless, the liquid-phase production of graphene presents several challenges because graphene sheets tend to aggregate and restack due to strong inter-sheet van der Waals forces and π–π interactions. To overcome this problem, the dispersion of graphene should be carried out in specific solvents or the graphene surface should be altered through covalent or non-covalent modifications.

Many studies have reported the liquid-phase exfoliation of graphite using high surface tension solvents.12–15 The graphene–solvent interaction should be sufficiently strong to offset the van der Waals forces between graphene sheets, with solvents with surface tensions in the region of 40–50 mJ m−2 being the most suitable for the exfoliation. However, these solvents are characterized by high boiling points, which render their removal difficult when processing graphene into composites, since very high temperatures are needed. For the manufacturing of composite materials, in particular, with tailored properties for specific needs, the removal of high boiling point solvents is an unquestionable disadvantage. On the other hand, using low boiling point solvents with surface tension of much less than 40 mJ m−2 low concentration of graphene is normally obtained of poor exfoliation quality and stability.16

An alternative pathway to obtain concentrated graphene dispersions of high quality is the noncovalent modification of graphene surface, using specific dispersion/stabilization agents, either small organic molecules or macromolecules.17–26 The advantage of this type of modification compared to the covalent functionalization of graphene is that this method does not destroy the two-dimensional sp2 network of graphene, thereby preserving its outstanding properties. Especially, the functionalization of graphene through π–π interactions as the binding forces between graphene and stabilizers is one of the most effective non-destructive methods to prepare functional graphene of high quality. Many researchers have therefore recommended the adsorption of pyrene derivatives on graphene sheets through π–π stacking interactions as a very efficient method for the production of stable and uniform graphene dispersions.27–34 Recent studies have also suggested the use of pyrene-modified polymers as graphene exfoliation agents in water35–37 or other solvents27 such as chloroform.38,39 For example, in a recent article40 we showed that by using pyrene functionalized poly(methyl methacrylate) (py-PMMA-py) as dispersing agent, stable graphene dispersions can be produced in chloroform starting from relatively high concentrations of a few-layered graphene sheets. We also showed that the suggested telechelic macromolecular topology proved more efficient as a dispersing agent than the monofunctional Py-PMMA or the nonfunctionalized PMMA.

Compared with small pyrene derivatives, the main advantage of using pyrene-modified polymers for the functionalization of graphene is the good compatibility with different polymer hosts thus allowing for the fabrication of several graphene-based nanocomposite materials. Additionally, the non-destructive nature of the noncovalent modification through π–π interactions renders the method suitable and effective for producing composites that preserve the superior properties of graphene. It is for these reasons that the exfoliation of graphite in this study is pursued in the presence of pyrene-functionalized polymers as stabilizers and through the use of hybrid dispersions for preparing composite materials.

Among many kinds of composites, nanocomposites in fiber form have attracted considerable attention in the last years as very promising materials for new applications, due to their large surface areas and controllable size of nanofibers.41 Electrospinning is the most powerful technique for fabricating continuous nanofibers by the application of electrostatic forces to a jetting polymer solution.42,43 Adding nanofillers such as graphene into the electrospun polymer solution could enhance the mechanical and electrical properties of composite nanofibers rendering them very promising materials for several applications such as in the production of thin film batteries.44–48 Already, electrospun nanocomposite fibers are considered as candidate materials for several optoelectronic devices.49–52 However, the composite properties are highly influenced by the uniform dispersion of nanofiller in the polymer matrix.

Keeping all of the above considerations in mind, the present work is concerned with the production of stable graphene dispersions from the liquid-phase exfoliation of pristine graphite in chloroform using well-defined pyrene-end functionalized 3-arm star PMMAs as dispersing agents. We follow up on our previous work40 where linear pyrene-capped PMMA (py-PMMA-py) was used to explore the role of macromolecular topology on the exfoliation efficiency of graphite and we design new, pyrene-functionalized 3-arm star PMMAs bearing three, (PMMA-Py)3, or six, (PMMA-Py2)3, pyrene-end groups. The synthetic method pursued comprises ARGET ATRP controlled polymerization in combination with click chemistry (azide/alkyne cycloaddition) and the quaternization reaction on triazole, ensuring well-defined macromolecular topologies. In parallel, and in order to assess the capability of (PMMA-Py)3 to lead to stable dispersions of exfoliated graphite in chloroform, detailed atomistic-level molecular dynamics simulations have been carried out. These revealed many interesting conformations of the adsorbed star polymer on graphene sheets (GS) which play a major role in stabilizing the dispersion. Indeed, in the presence of the synthesized pyrene-functional star PMMAs stable graphene dispersions were obtained in our experiments, thus confirming the predictions of the simulations. Finally, as a potential application, nanocomposite fibrous veils were successfully prepared through electrospinning using graphene/PMMA(Py)n hybrids as nanofillers and PMMA as a matrix. Interestingly, these nanocomposites were found to exhibit enhanced mechanical properties compared to the neat PMMA fibers.

2. Experimental

2.1 Materials

1-Pyrenemethanol (Aldrich, 98%), 1-pyrenebutyric acid (Aldrich, 97%), 1-(bromomethyl) pyrene (Aldrich, 96%), α-bromoisobutyryl bromide (Aldrich, 98%), N,N,N′,N′′,N′′′-pentamethyldiethylenetriamine (PMDETA; Aldrich, 99%), copper(I) bromide (Aldrich, 99.999%), copper(II) bromide (Aldrich, 99.999%), sodium azide (Aldrich, ≥99.5%), 1-hydroxybenzotriazole hydrate (HOBt hydrate, Aldrich, ≥99.5%), 4-ethynylaniline (Aldrich, 97%), N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide hydrochloride (EDC hydrochloride, Aldrich, ≥99%), 1,1,1-tris(hydroxymethyl)propane (Aldrich, 97%), anhydrous solvents N,N-dimethylformamide (DMF) and Tetrahydrofuran (THF) (water <0.005%, Aldrich), methanol, anisole, dichloromethane, anhydrous magnesium sulphate and neutral alumina were used as received from Aldrich. MMA (Aldrich, 99%) was passed through a column filled with basic alumina and Poly(methyl methacrylate), PMMA with weight average molecular weight 120[thin space (1/6-em)]000 g mol−1 was purchased from Aldrich. Graphite with an average particle size of 500 μm and purity greater than 95% was supplied from NGS Naturgraphit GmbH (batch: large flakes).

2.2 Instrumentation

1H-NMR spectra were obtained on a Bruker Advance DPX spectrometer at 400 MHz. The samples were dissolved in deuterated chloroform or dimethyl sulfoxide with TMS as internal standard. Gel Permeation Chromatography (GPC) measurements were carried out using two PLgel MiniMix columns “C” and “D” and a refractive index (RI) or a UV (254 nm) detector with THF as eluent and a flow rate of 0.5 mL min−1 at 25 °C. The column system was calibrated with standard linear PMMAs in the 2000–500[thin space (1/6-em)]000 Mn range. A bath sonicator Branson 1200-E1 (30 W), a Qsonica Sonicator model Q55 tip sonicator at 10% amplitude and an IKA Ultra-Turrax model T18 digital high-shear mixer with motor speed of 6000 rpm were used for the exfoliation process. A Shimadzu UV 2500 absorption spectrophotometer was used to obtain absorption spectra between 200 and 700 nm. Spectra were recorded in a 1 cm path length quartz cuvette at 25 °C with a resolution of ±2 nm. TEM images were obtained with a JEM-2100 transmission electron microscope operating at 200 kV. SEM images of nanocomposite fibers were taken with a LEO Supra 35VP microscope, after the samples were gold-sputtered. Raman spectra were recorded by a commercial Raman system (Renishaw, Invia 2000) at ambient conditions. The laser wavelength was 514.5 nm (2.41 eV) and the irradiation power at the sample was kept below 1 mW to avoid local heating. A 100× objective lens (Olympus, MPLAN, NA 0.95) was used to focus laser light onto the sample with an estimated focal spot diameter less than 1 μm. Extended mapping measurements were taken using a high spatial resolution mapping facility with 2 μm step. SEM images of nanocomposite fibers were also taken using the LEO Supra 35VP microscope. Tensile test was performed using a micro-tensile tester (MT-200, Deben UK Ltd, Woolpit, UK) equipped with a 5 N load cell. Prior to the test, nonwoven mats consisting of neat PMMA fibers or of PMMA/G_PMMA(Py)3 nanocomposite fibers with graphene content 0.1, 0.5 and 1.5 wt% were cut into stripes having an overall length of 25 mm, a gauge length of 10 mm and a width of 3 mm. For each specimen, the thickness was determined as the mean of 10 measurements along the gauge length with a digital micrometer with a resolution of 0.0001 mm (Mitutoyo, Japan). All test specimens were secured onto paper testing cards using a two-part cold curing epoxy resin (Araldite 2011, Huntsman Advanced Materials, UK) to avoid damages to the gripping area of the test specimens by the clamps of the tensile tester, potentially leading to an earlier onset failure within the gripping zone of the test specimens. These specimens were subsequently loaded in tension with a crosshead displacement speed of 0.5 mm min−1 (corresponding to a test specimen strain rate of 0.05 s−1). Stress and strain were calculated based on the measured machine-recorded forces and displacements (knowing the initial cross-section area and gauge length, respectively). The Young's modulus was estimated through a linear regression analysis of the initial linear portion of the stress–strain curves. Average results of 10 test specimens are reported for each sample.

2.3 Synthesis of polymers

2.3.1. Synthesis of star PMMA. A tri-functional ATRP initiator was prepared for the synthesis of the 3-arm star PMMA via the esterification reaction between the commercially available 1,1-tris(hydroxymethyl)propane and 2-bromoisobutyryl bromide in the presence of trimethylamine following reported procedures.53 The experimental procedure for the synthesis of the initiator (1) can be found in the ESI, and the same for the 1H NMR spectrum (Fig. S1) which proves the chemical structure. The star PMMA (2) was obtained using the ARGET ATRP polymerization under the ratio of MMA/Initiator/Cu(II)Br2/PMDETA = 100/1/0.5/1 in 50% (v/v) anisole at 90 °C for 30 minutes. More specifically, a mixture of Cu(II)Br2 (0.3074 g, 1.37 mmol) with the trifunctional initiator (1) (1.6000 g, 2.75 mmol) was added to a dried round-bottom flask. The flask was degassed and back-filled with argon five times. To a two-necked flask a mixture of MMA (27.5600 g, 275.00 mmol), anisole (29.45 mL) and PMDETA (0.4771 g, 2.75 mmol) was bubbled with nitrogen for 30 min and transferred to the flask containing degassed Cu(II)Br2 and the initiator. The flask was degassed again and back-filled with argon, and then left to 90 °C for 30 min. The polymer precipitated in hexane, filtered off, dissolved in THF and then passed through a neutral alumina column to remove the copper complex. Then, the polymer precipitated in hexane, filtered off and dried at 70 °C under vacuum for 24 hours. The polymer structure was confirmed with the 1H-NMR spectrum (see Fig. S2).
2.3.2. Synthesis of pyrene-end functionalized star PMMA (5). Firstly, the star polymer (2) (7.0000 g, 0.89 mmol) was dissolved in DMF (130 mL), and NaN3 (1.1649 g, 1700 mmol) was added. The reaction left at 40 °C overnight under argon atmosphere. The azidated polymer (3) precipitated in a large amount of water (500 mL), filtered off and dried under vacuum at 40 °C for 24 hours. The yield was 93%.

The alkyne-functionalized-pyrene N-(4-ethynylphenyl)-4-(pyren-1-yl)butanamide (4) was obtained from the amidation reaction of 1-pyrenebutyric acid with 4-ethynylaniline in the presence of 1-hydroxybenzotriazole hydrate and N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide HCl in dry DMF, as reported elsewhere.54 Its 1H-NMR spectrum confirmed the successful reaction (see Fig. S3).

Finally, for the synthesis of PMMA (5) bearing three pyrene-end groups, the Copper-Catalyzed Azide–Alkyne Cycloaddition (CuAAC) click reaction was employed with a molar ratio of azide/alkyne/CuBr/PMDETA = 1/3.3/4.068/4.068 in DMF at 110 °C for two days.55 In detail, to a degassed two-necked flask equipped with a magnetic stirrer and a dropping funnel, the synthesized N-(4-ethynylphenyl)-4-(pyren-1-yl) butanamide (4) (0.9819 g, 2.53 mmol), CuBr (0.4481 g, 3.12 mmol) and PMDETA (0.5413 g, 3.12 mmol) were dissolved in nitrogen-purged dry DMF (70 mL). The flask was degassed again and back-filled with argon three times. Then, the azidated polymer (3) (6.0000 g, 0.76 mmol) was dissolved in 50 mL of dry nitrogen-purged DMF and added drop-wise to the reaction mixture at room temperature. The reacting mixture was stirred at 110 °C for two days under argon atmosphere. The polymer (5) precipitated in 1000 mL of water, filtered off and dried under vacuum at 50 °C for 24 hours. The 1H-NMR spectrum which confirms the polymer functionalization is shown in Fig. S4 of the ESI.

2.3.3. Synthesis of quaternized pyrene-end star PMMA (6). The click reaction resulted in the formation of 1,2,3 triazole rings which were quaternized using 1-(bromomethyl) pyrene in excess thus producing the quaternized PMMA (6).56 Particularly, a degassed round-bottomed flask equipped with a magnetic stirring bar was charged with pyrene-end star PMMA (0.5000 g, 0.0741 mmol) and 1-(bromomethyl) pyrene (0.1971 g, 0.667 mmol) in 7 mL of anhydrous DMF. The mixture was purged with argon for 15 min and heated at 70 °C for 72 h. After cooling to room temperature, the reaction mixture precipitated in water and dried under vacuum for 1 day at 50 °C. To remove the excess of 1-(bromomethyl) pyrene, the dried polymer was dissolved in THF, filtered off and the filtrate precipitated in n-hexane and dried under vaccum for 1 day at 70 °C.

2.4 Liquid-phase exfoliation of graphite

The exfoliation of pristine layered graphite in the presence of pyrene functionalized star PMMAs was attempted in the organic solvent chloroform (CHCl3). First, the pristine graphite (starting concentration of 8 mg mL−1) and the solvent (CHCl3) were subjected instantaneously to bath and tip sonication (10% amplitude) for 90 min. Then, a polymer solution of pyrene functionalized star PMMAs in CHCl3 was added, and the exfoliation process was continued for 90 more minutes using bath and tip sonication simultaneously. The suspensions were left to settle overnight and then the upper phase was taken and centrifuged at 500 rpm for 30 min. After centrifugation, the supernatant was gently extracted by pipetting. Stable black-colored dispersions of the exfoliated graphene were obtained and used for further measurements. For the iterative method, used to increase the exfoliation yield, graphene dispersions were prepared by the method of tip sonication (10% amplitude) combined with shear mixing (rotor speed 6000 rpm), which allows the use of much higher quantity of pristine graphite (see details in ESI).

2.5 Preparation of composite solutions

The selected solvent for the composite solutions was a 1[thin space (1/6-em)]:[thin space (1/6-em)]1 mixture of DMF and CHCl3. First, the hybrid dispersions of graphene with the pyrene functionalized PMMAs were stirred at 40 °C in order to evaporate the excess solvent. Then, the required volume of DMF was added to the dispersions, and bath sonication followed for 30 min. After that, a weighted amount of PMMA with molecular weight of 120 000 g mol−1, chosen as a polymer matrix, was added to the dispersions and stirred for 24 h in order to obtain homogeneity. Different solutions were prepared for the needs of the current investigation. In all cases, the concentration of PMMA and graphene in the solutions was 16 wt% and 0.1 wt%, respectively. For the purpose of comparison, polymer solutions without graphene were also prepared.

2.6 Electrospinning of nanocomposite fibers

The two types of solution (with and without graphene) were transferred into a 10 mL plastic syringe for electrospinning at room temperature, with the flow rate kept constant at 1.8 mL h−1. The applied voltage was also kept constant at 18 kV in order to achieve continuous jet formation. The distance between the tip of the nozzle and the collector was 10 cm. Disposable blunt-tiped needles (Jensen Global Industrial Needles) of an inner diameter of 0.603 mm were used. An aluminum foil was positioned onto a square aluminum collector (15 × 15 cm) to collect both types of electrospun nanofibers. Finally, the samples (electrospun nanofibrous veils) were removed from the collector and left for 24 h at ambient temperature.

3. Results and discussion

3.1 Synthesis and characterization of functionalized star PMMAs

The method followed to prepare pyrene-end 3-arm star PMMAs bearing from three, (PMMA-Py)3, to six, (PMMA-Py2)3, pyrene units at the arm ends is shown in Fig. 1. The polymers were synthesized by ARGET ATRP controlled polymerization using the synthesized tri-functional initiator (1), according to the core-first method. The bromine end groups of the star shaped polymer (2) were converted to azide derivatives using an excess of sodium azide. This step was adopted in order to achieve the insertion of three pyrene-end groups at the arm ends of the star polymer (2), using the CuAAC click reaction between the azidated polymer (3) and the synthesized pyrene-based alkyne (4) which leads to the formation of triazole rings. The chemical structure of the pyrene-functionalized star PMMA (5) which verifies the successful click reaction as above, was confirmed using 1H-NMR spectroscopy. The spectrum shows clearly the presence of the aromatic protons of the pyrene units between 7.6–8.7 ppm (see, Fig. S4 at ESI) which in combination with the disappearance of the proton of the alkyne group at 4.1 ppm (Fig. S3 at ESI) provides evidence for the successful pyrene-functionalization at the arm-ends of the star PMMA (5). The quantitative arm-end functionalization was also confirmed by UV spectroscopy and GPC, equipped with UV detector, as will be reported below (Fig. 3). 1H-NMR also allowed the determination of the true Mn of the star PMMA (see ESI). Integration of the aromatic protons of the pyrene groups and the protons of the methoxy groups (–OCH3) of MMA repeating units indicated a degree of polymerization equal to about 98. Moreover, the apparent molecular weights and polydispersity of the polymers were determined by GPC, using linear PMMA standards. The characterization data are presented in Table 1. The true Mn(1H-NMR) is higher than the apparent Mn(GPC), confirming the star-shaped macromolecular architecture. As it is known, star polymers exhibit smaller hydrodynamic volumes (higher monomer density) with respect to their linear counterparts, which results to higher retention times in GPC and therefore giving lower molecular weights (apparent) when analyzed with linear polymer standards as in the present case. To conclude, a well-defined telechelic pyrene-functionalized 3-armed star PMMA with low-polydispersity was obtained by the chosen synthetic procedure.
image file: c8nr06888g-f1.tif
Fig. 1 Synthetic procedure of quaternized pyrene-end 3-arm star PMMA using the ARGET ATRP polymerization and click chemistry.
Table 1 Molecular characteristics of the pyrene-functionalized star PMMA
Polymer M n[thin space (1/6-em)]a (g mol−1) M w[thin space (1/6-em)]a (g mol−1) M n[thin space (1/6-em)]b (g mol−1) DP (Mw/Mn)a
a Determined by GPC. b Determined by 1H-NMR (see ESI). c Determined by 1H-NMR taking in consideration that the quaternization degree is 46% (see ESI).
(PMMA-Py)3 6740 8130 12[thin space (1/6-em)]460 1.20
(PMMA-Py2)3 7600 8130 12[thin space (1/6-em)]820c 1.20

To insert more pyrene moieties in the arm-ends, the quaternization of the triazoles was attempted using the commercially available 1-(bromomethyl)pyrene. The 1H-NMR spectrum obtained for the quaternized star polymer (PMMA-Py2) is shown in Fig. 2. The degree of quaternization was determined by 1H-NMR spectroscopy from the integration of the distinct peaks (h) at 10 ppm and (l) at 5.65 ppm which are attributed to the single proton of the amide group and the pair of protons vicinal to the site of quaternization, respectively. According to this, the degree of quaternization was estimated to be about 46% (despite that an excess of the quaternizing agent was used to achieve full quaternization of the three azole rings). This partial quaternization can be probably attributed to the higher steric hindrance and rigidity of the triazole group located directly next to the main chain of PMMA. The presence of the peak k′ at 8.7 ppm corresponding to the aromatic proton of the 1,2,3-triazole provides evidence of the incomplete quaternization.

image file: c8nr06888g-f2.tif
Fig. 2 1H-NMR spectrum of the quaternized py-functional 3-arm star PMMA.

The degree of quaternization of pyrene-end star PMMAs was examined by Uv-Vis spectroscopy. For the calibration of the Uv-Vis spectrometer, a linear a-pyrene-functional PMMA (Py-PMMA) with Mn of 2900 g mol−1 was used, synthesized by pyrene-functionalized ATRP initiator, ensuring 100% pyrene functionalization.21 The Uv-Vis measurements showed that the absorbance of pyrene-end star PMMAs is a linearly increasing function of the polymer concentration (in mg mL−1), obeying the Beer–Lambert law (Fig. S5). By assuming quantitative functionalization/quaternization of the pyrene-end star PMMAs [3-py per (PMMA-Py)3 6-py per (PMMA-Py2)3] and plotting all the data as a function of pyrene concentration expressed in moles per Lt (M), the data collapse to a straight line with the data of the linear Py-PMMA standard (Fig. 3a). This suggests that the functionalization/quaternization of pyrene-end star PMMA was quantitative. In order to check whether all pyrene groups are covalently bonded on the star polymers or not, GPC measurements with a UV detector were performed which allows separating contributions from conjugated and entrapped free pyrene groups. Fig. 3b presents the GPC chromatograms of the (PMMA-Py)3 precursor and of the quaternized (PMMA-Py2)3, together with that of quaternized agent, 1-(bromomethyl) pyrene. For the (PMMA-Py)3 precursor, a single peak can be observed confirming quantitative pyrene conjugation in the 3-arm star. In the case of (PMMA-Py2)3, a second peak assigned to the quaternized agent is revealed, which suggests partial quaternization of the (PMMA-Py)3 and contamination of the sample with the free pyrene agent that could not be removed by the purification applied. The degree of partial quaternization could be estimated from the GPC traces, provided that 6 pyrene units (conjugated or not) correspond to each star polymer as implied from Fig. 3a. Thus, by integrating the areas of the two peaks corresponding to conjugated and free pyrene groups, the degree of conjugated pyrene units to (PMMA-Py2)3 star polymer was determined to be 71.5%.

image file: c8nr06888g-f3.tif
Fig. 3 (a) Absorbance at 345 nm vs concentration in moles of pyrene per Lt (M) of linear Py-PMMA and pyrene-end star PMMAs, (PMMA-Py)3 and (PMMA-Py2)3. The straight line is the linear fit of all the data (R = 0.97). (b) GPC chromatograms of (PMMA-Py)3, (PMMA-Py2)3 and 1-(bromomethyl) Pyrene, using a Uv-Vis detector at 256 nm.

This means that the degree of quartenization of (PMMA-Py2)3 is 43%, which is in excellent agreement with the value (46%) obtained from the 1H-NMR data. The above analysis that combines Uv-Vis spectroscopy with GPC, equipped with UV detector, seems to be necessary for fully elucidating the pyrene funtionalization/quaternization of the 3-arm star PMMA.

3.2 Molecular dynamics simulations

From a computational point of view, several studies have demonstrated the strong interaction of aromatic pyrene with the surface of graphene sheets (GS) via π–π stacking.57–59 The Molecular Dynamics (MD) simulations of Xu and Yang57 for example, have provided strong evidence for a very favorable interaction of pyrene–polyethylene glycol (py–PEG) chains with GS. The final, equilibrium conformation was one in which the pyrene group of the py–PEG chains was lying flat on the GS at a stable separation approximately equal to 3.5 Å. A calculation of the free energy showed that the main contributor to adsorption was the π–π interaction between the pyrene group and the carbon atoms on the surface of the GS. Similar MD simulations were used to study the mechanism of solvent exfoliation and stabilization of graphene in the presence of (compressed) carbon dioxide (cpCO2), p-xylene and pyrene–polyethylene glycol (Py–PEG) polymers.59 Both of these MD studies are consistent with an earlier study by Zhang et al.60 who used single molecule force spectroscopy to directly measure the strength of the interaction between a pyrene molecule and a graphite surface.

In our previous contribution40 MD studies had demonstrated that linear pyrene-PMMA-pyrene (py-PMMA-py) chains can effectively stabilize GS in chloroform, forming some unique structures (dangling ends, loops and bridges) that provide the steric hindrance needed for GS to avoid agglomeration. In the present work we extend these studies to the case of pyrene-end 3-arm star PMMA. We carried out a long atomistic MD simulation of a model system, representative of the experimental samples, in which functionalized (PMMA-Py)3 chains were added into a chloroform solution in which a certain number of pristine graphene sheets (GSs) had been dispersed. The MD simulation was performed with a fully periodic simulation cell containing 14[thin space (1/6-em)]000 CHCl3 molecules in which 5 hydrogen-terminated single-layered GSs with lateral dimensions 39 Å × 39 Å had been dispersed. In this simulation cell, 75 (PMMA-Py)3 chains were added consisting of 10 methyl methacrylate monomers per PMMA arm resulting in a concentration equal to 1.75 wt% for GSs, equal to 16.91 wt% for (PMMA-Py)3 chains and equal to 81.34 wt% for CHCl3 molecules. The total number of interacting particles in the simulation cell was 123[thin space (1/6-em)]185. Clearly, the dimensions of GSs employed in the simulation are significantly smaller than those encountered in the corresponding experimental study because of computational limitations. Working with larger GSs would result in simulation cells that would be intractable by current computational resources. However, it is important to mention that (not a single but) several (= five) GSs were used in our simulations, and that their small sizes, commensurate with the size (radius of gyration) of the simulated PMMA chains, favors their mobility in the solution within the limits of available simulation time so that if they exhibited any tendency to self-assemble, this would be immediately realized in the course of the MD run. The MD simulation was executed in the isothermal-isobaric (NPT) statistical ensemble at T = 298.15 K and P = 1 atm. Additional technical aspects regarding the forcefield employed in the simulation for the description of all intra- and inter-atomic interactions, the types of thermostat and barostat chosen to keep the temperature and pressure fixed during the simulation, and the method used to integrate the equations of motion can be found elsewhere.40,61

The MD simulation lasted for about 800 ns and allowed us to monitor the displacement of the GSs in the solution to check whether or not these tend to get close to each other and form graphitic structures or stay away thus remaining well dispersed in the solution. To this, we followed the time evolution of the distance between the centers-of-mass of all GS pairs present in the simulation cell, and the results are shown in Fig. 4.

image file: c8nr06888g-f4.tif
Fig. 4 Time evolution of the distance between the centers-of-mass of all pairs of GS during our NPT MD simulation.

For all GS pairs, this distance remains constantly (i.e., for the entire duration of the MD simulation) well above the characteristic value of 3.4 Å for graphene self-assembly [i.e., the typical equilibrium distance between successive GS layers in a graphite flake]. We also clarify here that in the course of our 800 ns – long MD simulation, each GS was displaced (diffused) by about 1800 Å2, which is comparable to the lateral dimensions of the studied GSs (surface area equal to 1521 Å2). This indicates significant motion (spatial displacement) of the dispersed GS in the chloroform solution; however, this is not accompanied by any tendency for self-assembly, as can be clearly seen also from Fig. 5.

image file: c8nr06888g-f5.tif
Fig. 5 A simulation cell illustrating dispersion of GS (yellow) in CHCl3 (molecules not shown) in the presence of pyrene-functional 3-arm PMMA [(PMMA-Py)3] after several nanoseconds of simulation time.

To get a better understanding of the mechanism(s) preventing GS agglomeration in the chloroform solution in the presence of the (PMMA-Py)3 chains, we examined the environment around each GS which revealed significant adsorption of (PMMA-Py)3 chains on the available area on the two faces of the five GSs. A more careful inspection of the trajectory file showed that functionalized (PMMA-Py)3 chains exhibit a strong tendency to adsorb on the dispersed GSs thereby forming a variety of conformations which can be categorized as follows (see Fig. 6):

Type-1 (dangling ends): This conformation corresponds to a (PMMA-Py)3 chain adsorbed on a GS by only one of its three ends, the other two ends remaining free (Fig. 6a).

Type-2 (loop and dangling end): This corresponds to a (PMMA-Py)3 chain adsorbed by two of its end pyrene groups on the same face of a GS, with the third remaining free (Fig. 6b).

Type-3 (extended loop & dangling end): This corresponds to a (PMMA-Py)3 chain adsorbed on the two different faces of the same GS by two of its three ends, leaving the other end free (Fig. 6c).

Type-4 (double loops): This corresponds to a (PMMA-Py)3 chain all three end-pyrene groups of which have been adsorbed on the same face of the same GS (Fig. 6d).

Type-5 (extended double loops): This type of conformations corresponds to a (PMMA-Py)3 chain adsorbed on both faces of the same GS (one pyrene adsorbed on one face and the other two pyrenes adsorbed on the other face) (Fig. 6e).

Type-6 (bridges): This type of conformations corresponds to a (PMMA-Py)3 chain adsorbed with two of its end-pyrene groups to two different GSs, the third group remaining free (Fig. 6f).

Type-7 (loop & bridges): This type of conformations corresponds to a (PMMA-Py)3 chain adsorbed to two GSs with two of its end-pyrene groups at the same face of one GS and the third to another GS (Fig. 6g).

Type-8 (extended loop & bridges): This type of conformations corresponds to a (PMMA-Py)3 chain adsorbed to two GSs with two of its end-pyrene groups at the two different faces of the same GS and the third one to a different GS (Fig. 6h).

Type-9 (triple bridges): This conformation corresponds to a (PMMA-Py)3 chain adsorbed to three different GSs (Fig. 6i).

image file: c8nr06888g-f6.tif
Fig. 6 Typical atomistic conformations (a–i) of adsorbed (PMMA-Py)3 chains on the faces of GS that have been dispersed in a chloroform solution. Carbon (PMMA), carbon (pyrene), and oxygen atoms are represented in gray, yellow and red colors, respectively, while hydrogen atoms have been omitted for clarity.

As an additional measure of proper GS dispersion in our simulation, we calculated the radial pair distribution function gji(r) between atoms i and j. We are particularly interested in gji(r) curves where the atomistic unit i corresponds to GS atoms and the atomistic unit j to atoms in the three pyrene groups of (PMMA-Py)3chains; let us symbolize this function as gpyrenegraphene(r). Fig. 7 shows a series of gpyrenegraphene(r) graphs computed as ensemble averages over all configurations accumulated in the course of the MD simulation between two different time instances. It provides information about the time evolution and the type of adsorbed pyrene group conformations on the available surfaces of GSs. A clear peak at distance ∼4 Å (corresponding to the intermolecular distance between carbon atoms in GS and carbon atoms in pyrene groups) is immediately discernible in the curves even from the very first 10 ns of the simulation. This indicates that pyrene groups belonging to functionalized (PMMA-Py)3 chains, as they rapidly explore space in their vicinity, find GSs onto which they tend to adsorb strongly. As the simulation goes on, this peak becomes sharper and sharper, since more and more pyrene groups adsorb on the GSs that have been dispersed in the chloroform solution. This increase continues up to ∼600 ns beyond which the computed gpyrenegraphene(r) vs. r curves do not change any more. This is due to GS face saturation: all available area on the two faces of all GSs has been completely covered by pyrene groups, thus leaving no space for other pyrene groups to adsorb. An interesting feature to note in the graphs of Fig. 7 is the appearance of a 2nd peak that slightly forms at ∼7 Å after about 100 ns. This reveals that a 2nd layer of adsorbed pyrene groups forms on GS, implying multi-layer adsorption. It is also interesting that, similar to the peak at ∼4 Å, the height of this 2nd peak remains constant after ∼600 ns. This is because, by this time, all pyrene groups have found GSs to adsorb, thus there is no more population left to adsorb.

image file: c8nr06888g-f7.tif
Fig. 7 Pair distribution function gpyrenegraphene(r) at different time intervals during the MD simulation. The curve marked as “initial” represents the gpyrenegraphene(r) curve obtained from the starting (at time t = 0) system configuration. Similar, the curve marked as “final” represents the gpyrenegraphene(r) curve obtained from the analysis of the finally stored (at time t = 800 ns) system configuration.

The observations drawn from the analysis of the computed gpyrenegraphene(r) vs. r curves in time agree with the time evolution of the relative population of the nine types of adsorbed conformations of (PMMA-Py)3 chains, shown in Fig. 8. As time goes by, the population of non-adsorbed (free) chains decreases continuously while that of adsorbed conformations increases. This continues until an asymptotic behavior is reached beyond which no further increase is observed, implying constant populations in time. We also observe that the dominant conformation of adsorbed segments is that of dangling ends, which is also the easiest to form. But clearly a non-negligible fraction of all other conformations is also recognized in Fig. 8. Yet, as we already explained, after ∼600 ns, the population of all nine conformations fluctuates around a constant value, confirming that the system reached thermodynamic equilibrium.

image file: c8nr06888g-f8.tif
Fig. 8 Time evolution of the relative population of the various types (see Fig. 5) of adsorbed (PMMA-Py)3 chain conformations and of the corresponding free (PMMA-Py)3 chains, in the course of our NPT MD simulation.

3.3 Exfoliation of graphite

The liquid-phase exfoliation of graphite in chloroform was pursued by using the synthesized py-functional star PMMAs as dispersing/stabilizing agents. Chloroform was selected as the liquid medium due to its low boiling point, which ensures easy removal for the preparation of nanocomposites. The graphite exfoliation was attempted by using concurrently bath and tip sonication following the procedure described in section 2.4. As reported in a previous work of our group40 by the simultaneous combination of different methods of exfoliation a more effective exfoliation can be achieved. It is interesting to note the GPC data indicate clearly that the polymers remained intact after this treatment (Fig. S6).

To investigate the exfoliation efficiency of the synthesized py-functional polymers, a number of dispersions were prepared by varying the polymer concentration and keeping all other factors constant (i.e., CGF = 8 mg ml−1, V = 6 ml, sonication time 3 h). The graphene concentrations (CG) of the stabilized dispersions were determined by UV-vis spectroscopy as described in ESI. In Fig. 9(a), the CG is plotted as a function of concentration for the (PMMA-Py)3 and (PMMA-Py2)3 stabilization agents. As shown, the CG depends strongly on polymer concentration and degree of pyrene functionalization of the star PMMAs. Particularly, (PMMA-Py2)3 promotes higher exfoliation efficiency because it exhibits remarkably higher CG (∼237%) than (PMMA-Py)3 for the same initial polymer concentration (0.05 mg mL−1). More importantly, a high CG value of 0.36 mg ml−1 was achieved using a very low polymer concentration of 0.025 mg ml−1, which corresponds to an unusually very low polymer/graphite mass ratio of mP/mGF = 0.003, demonstrating that the (PMMA-Py2)3 star is a very efficient dispersing agent of graphene. However, further increasing the stabilizer's concentration causes a decrease in CG. This is possibly attributed to the fact that pyrene groups at high concentrations tend to interact with each other, thus leaving less free pyrene moieties to interact with graphene flakes through π–π interactions. It is possible that apart from the (PMMA-Py2)3, the presence of the quaternized agent in the polymer sample contributes to some extent to the high dispersibility of graphite. It is also interesting to note that the graphene dispersions exhibited high stability even after several weeks in the presence of both (PMMA-Py)3 and (PMMA-Py2)3, as demonstrated in Fig. 9(b) and (c), respectively.

image file: c8nr06888g-f9.tif
Fig. 9 (a) Graphene concentration, CG, in CHCl3 and in the presence of (PMMA-Py)3 and (PMMA-Py2)3 at different initial polymer concentration (mg mL−1). Digital images after several weeks of hybrid dispersions in the presence of polymers (b) (PMMA-Py)3 and (c) (PMMA-Py2)3 at different initial concentration.

The exfoliation yield for the maximum concentration achieved was relatively low (i.e., 0.94% for (PMMA-Py)3 and 4.5% for (PMMA-Py2)3). In order to increase the exfoliation yield for (PMMA-Py)3 as the dispersing agent, an iterative method was attempted relying on the redispersion of the sediment, followed by successive centrifugation, separation and redispersion cycles. After the exfoliation of graphite in the presence of (PMMA-Py)3 and removal of the supernatant, the sediment was re-dispersed many times by adding polymer at each repeated cycle (Scheme S1). As we have shown previously,40 combination of high shear mixing with tip sonication provides better exfoliation capability, and hence it was used in this process. Indeed, as can be seen from Fig. S8, high CG was obtained after each cycle: for instance, CG = 2.25 mg ml−1 in the first supernatant and CG = 1.8 mg ml−1 after the first cycle. Overall, the 5-cycle iterative procedure resulted in a larger quantity of dispersed graphene nanosheets with (average value) CG = 1.29 mg ml−1 and total yield of 14.8% under conditions of 10.5 h [shear mixing (6000 rpm)/tip sonication (30% amplitude)] total time and mP/mGF = 0.15.

Finally, in order to evaluate the importance of the macromolecular topology of the pyrene-functional PMMAs to exfoliate graphite in chloroform forming graphene/PMMA hybrids, the data of the present work are plotted together with those of our previous work40 in Fig. 10. The CG values have been measured from stable dispersions prepared under the same conditions (i.e., CGF = 8 mg ml−1, V = 6 ml, bath & tip sonication, sonication time 3 h, settle overnight – upper phase centrifugation 500 rpm, 30 min). Obviously, the most efficient exfoliation/stabilization agent is the (PMMA-Py2)3 3-arm star PMMA at very low polymer concentration with the second best stabilizer being the telechelic py-PMMA-py which exhibits the maximum CG at a higher polymer concentration. Expressing the concentration of the stabilizer in pyrene molar concentration, it comes out that it is the molecular topology (and not the overall number of pyrene functions) the factor that promotes exfoliation of graphite. We note here that the polymer concentration (and, in turn, the mass ratio mP/mGF) is lower than 0.075 in this series of our experiments with the functional PMMAs, which to the best of our knowledge is the lowest concentration reported so far.

image file: c8nr06888g-f10.tif
Fig. 10 Concentration of dispersed GS in chloroform, CG, exfoliated under similar conditions, as a function of stabilizer concentration expressed in polymer mg ml−1 (a) or pyrene functions M (b), for various pyrene-functional PMMAs.

3.4 Dispersion quality

The exfoliation efficiency of graphene dispersions was evaluated through TEM microscopy and Raman Spectroscopy (RS). Representative TEM images of hybrids graphene/(PMMA-Py)3 and graphene/(PMMA-Py2)3 can be seen in Fig. 11(a) and (b) respectively. The size of the graphene flakes varies between 0.5 to 2 μm, but the majority of the flakes have size of the order of 1 μm.
image file: c8nr06888g-f11.tif
Fig. 11 TEM images of hybrid graphene/polymer dispersions after centrifugation at 500 rpm in the presence of polymers (a) (PMMA-Py)3 and (b) (PMMA-Py2)3 (scale bar: 500 nm).

Raman spectroscopy, which is one of the most useful and versatile tools to characterize and investigate graphitic materials,62,63 was used to examine the quality and thickness of the produced graphene nanosheets. The following statistical analysis is based on the Raman 2D peak shape (more details in ESI).64,65 This approach provides an accurate estimation for the mean number of layers in graphene flakes at least up to maximum 10 graphene layers. Fig. 12 shows the Raman spectra of graphene flakes with different number of layers, and the percentage of graphene flakes with the number of layers of hybrid dispersions graphene/(PMMA-Py)3 and graphene/(PMMA-Py2)3. The characteristic G and 2D graphene peaks are clearly visible in all cases. An intense D peak is also shown revealing the presence of defects in the crystal lattice and/or the presence of edges due to the relatively small lateral size of exfoliated flakes. The polymer (PMMA-Py2)3 is clearly more efficient dispersing agent than its precursor (PMMA-Py)3 since the relative percentage of few-layer graphenes (2–3 layers) is 10% and below 1%, respectively. Furthermore, in the case of (PMMA-Py2)3 single layer graphenes are also detected. Additionally, both methods produce many multilayer graphene flakes (5–10 layers) and a small amount of nanographites.

image file: c8nr06888g-f12.tif
Fig. 12 Representative Raman spectra of graphene flakes with various numbers of layers and percentage of graphene flakes with number of layers of hybrid dispersions graphene/(PMMA-Py)3 (a and a′) and graphene/(PMMA-Py2)3 (b and b′).

3.5 Electrospun PMMA/G@PMMA(Py)n nanocomposite fibers

3.5.1 Morphology and microstructure characterization. The potential of the hybrid dispersions of graphene to be used for the fabrication of nanocomposite fibers through electrospinning technique was also investigated. For this purpose, PMMA/G@(PMMA-Pyn)3 nanocomposite fibers were prepared with 16 wt% PMMA and 0.1 wt% graphene content. As already mentioned, PMMA matrix was chosen due to its compatibility with the functionalized star PMMAs.

The morphology of the nanocomposite fibers was examined by SEM and TEM microscopy as shown in Fig. 13(c) for the PMMA/G@(PMMA-Pyn)3 derivative. For comparison, a SEM image of neat PMMA fibers is also presented in Fig. 13(a). As can be seen, all electrospun fibers form a fibrous morphology without the presence of beads or gross defects and are randomly oriented. This indicates that the employed manufacturing parameters are suitable for the fabrication of uniform fibers. In the case of pure PMMA fibers, their surface is smooth whereas the nanocomposite fibers exhibit a wrinkled structure due to the incorporation of graphene nanosheets but they appear highly continuous and homogeneous (Fig. 13(e)). As can be observed from Fig. 13(d) and (f), the average diameter of nanocomposite fibers is slightly smaller compared to neat PMMA fibers (Fig. 13(b)). This can be attributed to the increase in the conductivity of the solution due to the added graphene nanosheets.42 Thus, improvement of conductivity contributes to the electrostatic charge build-up during electrospinning, which in turn leads to the fabrication of thinner fibers as a result of the narrower distribution of the fiber diameter.

image file: c8nr06888g-f13.tif
Fig. 13 SEM images and diameter histogram frequency chart of PMMA (a and b), PMMA/G@(PMMA-Py)3 (c and d) and PMMA/G@(PMMA-Py2)3 (e and f), with 0.1 wt% graphene content (scale bar: 10 μm).

To investigate the internal structure of the graphene nanosheets in the PMMA fibers, the PMMA/G@(PMMA-Py2)3 nanocomposite fibers were characterized by TEM microscopy and the use of copper grids for the collection of electrospun fibers. In Fig. 14(a–d), the TEM images show quite clearly the incorporation of graphene nanosheets in the PMMA fibers which are broadly oriented along the axial direction of the fiber. This can be explained by the stresses imparted on the formed fibers during the electrospinning process, which causes their alignment along the fiber axis. However, Fig. 14(a) shows that in some cases graphene nanosheets are oriented perpendicular to the fiber direction.

image file: c8nr06888g-f14.tif
Fig. 14 TEM images of PMMA/G@(PMMA-Py2)3 (scale bar: 0.5 μm).
3.5.2 Mechanical properties. The mechanical properties of the electrospun PMMA/G@(PMMA-Py)3 nanocomposite fibers with hybrid graphene content equal to 0.1, 0.5 and 1.5 wt% were investigated by uniaxial tensile testing. Pristine electrospun PMMA mat was tested and adopted as reference. As described above, the tensile loading was applied to rectangular specimens prepared from the electrospun non-woven fiber mats mounted on testing cards. In order to avoid severe damage, extreme care was taken when preparing and then gripping the specimens to the tensile apparatus. Representative stress–strain curves for the pristine PMMA mat and the PMMA/G@(PMMA-Py)3 nanocomposites are shown in Fig. 15a. For each specimen, the Young's modulus was calculated through linear regression analysis of the initial linear parts of the stress–strain curves. The pristine PMMA reveals the typical yield behavior with the corresponding Young's modulus and tensile strength being equal to 31.8 MPa and 0.3 MPa, respectively. The addition of hybrid graphene remarkably changes the tensile behavior of PMMA mats: yielding becomes hardly discernable especially at the higher graphene content and the elongation at the break gradually decreases. The average values of Young's modulus, tensile strength, elongation at break and modulus of toughness, along with the associate errors, are plotted as a function of the weight-percent hybrid graphene loading in Fig. 15b–d. For each sample, the results reported are the averages obtained from measurements from 10 individual specimens. The trend reveals a noteworthy increase in Young's modulus and strength of the nanocomposites with graphene content. For the nanocomposite fiber mat with 1.5 wt% in hybrid graphene, the Young's modulus and the tensile strength are found to be 139.5 MPa and 1.3 MPa, respectively, which are higher by a factor of 4.4 compared to those for the pristine PMMA mat. The incorporation of small amounts of hybrid graphene leads also to the enhancement of the modulus of toughness of the nonwoven mats, evaluated as the area under the stress–strain curve (Fig. 15d). It is interesting noting that for these nanocomposite fiber mats with 0.1 and 0.5 wt% in hybrid graphene, the elongation at break remains approximatively similar to the value of the neat PMMA mat (Fig. 15c). Instead, the nanocomposite with 1.5 wt% in hybrid filler presents an earlier failure, and thus no further increase of modulus of toughness is observed for this systems, albeit the value is higher by a factor of 2.3 compared to the pristine PMMA mat. This improvement of the mechanical properties can be attributed to the high stiffness and high aspect ratio of the incorporated hybrid graphene flakes and their preferred orientation along the fiber axis. Overall, as revealed by Raman spectroscopy, the exfoliation in the presence of pyrene-end star PMMAs effectively leads to high aspect ratio nanofillers (with the number of graphene layers ranging between 4 and 8), which are more effective load bearers than nanographites.66 Also, the presence of graphene oriented along the fiber axis promotes the toughening of the nanocomposites through different mechanisms (e.g. crack bridging and crack deflection).67 In addition, the adsorbed (PMMA-Py)3 chains tend to promote both the interfacial interactions (which further contributes to the effectiveness of the load transfer) and the toughen effect, and to provide improved dispersion of the hybrid graphene in the polymer matrix (which results in the enhancement of the tensile strength of the nanocomposite).
image file: c8nr06888g-f15.tif
Fig. 15 Typical stress–strain curves (a), tensile properties (b), elongation at break (c) and modulus of toughness (d) of electrospun PMMA/G@(PMMA-Py)3 nanocomposite mats as a function of graphene content.

4. Conclusions

Liquid-phase exfoliation of graphite in the low-boiling point chloroform has been conducted by using pyrene-functionalized 3-arm star PMMAs bearing 3 [(PMMA-Py)3] and 6 [(PMMA-Py2)3] pyrene-end functionalities. The designed 3-arm stars were synthesized by ARGET ATRP controlled polymerization and their pyrene-end functionalization was implemented by click chemistry. The synthetic strategy followed, together with detailed characterization, ensured tailor-made, well-defined py-functional PMMA dispersing/stabilizing agents of graphene.

The presence of dispersing agents led to stable few-layered graphene sheets in chloroform. The polymer (PMMA-Py2)3 exhibited higher dispersing efficiency than its non-quaternized analogue, (PMMA-Py)3 under the same conditions. More importantly, high CG of 0.36 mg ml−1 was achieved using a very low polymer concentration (0.025 mg ml−1). To the best of our knowledge, this corresponds to remarkable very low polymer/graphite mass ratio mP/mGF = 0.003, the lowest reported so far, demonstrating the efficiency of (PMMA-Py2)3 star as a dispersing agent of graphene. A comparison of the present results with those reported previously40 reveals the importance of macromolecular topology in the design of polymer-based pyrene-functional dispersing agents. Among the various polymers synthesized and tested (py-PMMA, py-PMMA-py, (PMMA-Py)3 and (PMMA-Py2)3), (PMMA-Py2)3 proved to be the most efficient dispersing agent. To the advantages of the present dispersion procedure we should add the low (low-power) sonication time (3 h) required for the exfoliation of graphenes. Note that for analogous times in the presence of stabilizers, CG never exceeded 0.05 mg ml−1 in chloroform.

In order to increase the exfoliation efficiency in terms of CG combined with mG/mGF % exfoliation yield, a 5-cycle iterative procedure relying on the redispersion of the sentiment was developed. With a mean CG = 1.29 mg ml−1, a total yield of 14.8% was achieved for 10.5 h total shear mixing/tip sonication time and overall mP/mGF ratio as low as 0.15.

All-atom MD simulations were also performed to evaluate the dispersing ability of the py-functional 3-arm (PMMA-Py)3 in chloroform. Our results indicated significant motion (spatial displacement) of the dispersed GS in the chloroform solution but without any tendency for agglomeration. We also noticed that GS remained well dispersed even after chloroform was substituted by PMMA chains (after solvent evaporation towards graphene-reinforced PMMA composites), as will be reported in a forthcoming paper. According to the MD simulations, the pyrene functions of the functionalized (PMMA-Py)3 chains tend to adsorb strongly on the dispersed GSs thereby contributing to the formation of a variety of novel conformations that help keep GS apart and well dispersed in the solution or in a melt with other, non-functionalized PMMA chains.

Finally, the G@(PMMA-Pyn)3 hybrids were well incorporated into the PMMA matrix by an electrospinning technique to obtain graphene-based nanocomposite fibrous veils. The produced PMMA/G@(PMMA-Pyn)3 elecrospun fiber mats exhibit relatively uniform morphology, with thinner fibers compared to pure PMMA. Interestingly, a modulus enhancement by a factor of 4.4 for 1.5 wt% G@(PMMA-Py)3 nanofiller was observed. This reinforcement could be attributed to the preferred axial orientation of the nanofiller, the good dispersion provided by the (PMMA-Py)3 star stabilizer along with the improved interfacial interactions (caused by the pyrene-end functions adsorbed on the GS surface) and the PMMA chains of the star (that are entangled with the matrix PMMA chains).

Conflicts of interest

There are no conflicts to declare.


This work was supported by the European Union (European Social Fund-ESF) and Greek national funds through the research Funding Program: ERC-10 “Deformation, Yield and Failure of Graphene and Graphene-based Nanocomposites”. The financial support of the Graphene FET Flagship (Grant Agreement No: 604391) is also acknowledged by CG. ENS and VGM acknowledge support provided by the General Secretariat for Research and Technology (GSRT) and the Hellenic Foundation for Research and Innovation (HFRI) in the context of the action ‘1st Proclamation of Scholarships from ELIDEK for PhD Candidates’ – Scholarship Code: 2043. The work was supported by computational time granted from the Greek Research & Technology Network (GRNET) in the National HPC facility-ARIS-under project pr005013. To this, we feel indebted to Dr Dimitrios Dellis from GR-NET, Greece, for his invaluable technical support regarding MD run execution on ARIS. We also thank Panagiotis Mermigkis for valuable discussions regarding the identification of the different types of adsorbed polymer conformations on GSs in the MD part of our study, and the calculation of their relative populations.


  1. D. R. Dreyer, R. S. Ruoff and C. W. Bielawski, Angew. Chem., Int. Ed., 2010, 49, 9336 CrossRef CAS.
  2. T. Ramanathan, A. A. Abdala, S. Stankovich, D. A. Dikin, M. Herrera-Alonso, R. D. Piner, D. H. Adamson, H. C. Schniepp, X. Chen, R. S. Ruoff, S. T. Nguyen, I. A. Aksay, R. K. Prud'Homme and L. C. Brinson, Nat. Nanotechnol., 2008, 3, 327 CrossRef CAS PubMed.
  3. X. G. Li and G. B. McKenna, ACS Macro Lett., 2012, 1, 388 CrossRef CAS.
  4. K. Hu, D. D. Kulkarni, I. Choi and V. V. Tsukruk, Prog. Polym. Sci., 2014, 39, 1934 CrossRef CAS.
  5. M. Zhang, Y. Li, Z. Su and G. Wei, Polym. Chem., 2015, 6, 6107 RSC.
  6. K. S. Novoselov, A. K. Geim, S. V. Morozov, D. Jiang, Y. Zhang, S. V. Dubonos, I. V. Grigorieva and A. A. Firsov, Science, 2004, 306, 666 CrossRef CAS PubMed.
  7. K. S. Novoselov, V. I. Fal'ko, L. Colombo, P. R. Gellert, M. G. Schwab and K. Kim, Nature, 2012, 490, 192 CrossRef CAS.
  8. Y. L. Zhong, Z. Tian, G. P. Simon and D. Li, Mater. Today, 2015, 18, 73 CrossRef CAS.
  9. J. N. Coleman, Acc. Chem. Res., 2013, 46, 14 CrossRef CAS.
  10. H. Kim and C. W. Macosko, Macromolecules, 2008, 41, 3317 CrossRef CAS.
  11. H. Kim, Y. Miura and C. W. Macosko, Chem. Mater., 2010, 22, 3441 CrossRef CAS.
  12. U. Khan, H. Porwal, A. O'Neil, K. Nawaz, P. May and J. N. Coleman, Langmuir, 2011, 27, 9077 CrossRef CAS PubMed.
  13. Y. Hernandez, V. Nicolosi, M. Lotya, F. M. Blighe, Z. Sun, S. De, I. T. McGovern, B. Holland, M. Byrne, Y. K. Gun'Ko, J. J. Boland, P. Niraj, G. Duesberg, S. Krishnamurthy, R. Goodhue, J. Hutchison, V. Scardaci, A. C. Ferrari and J. N. Coleman, Nat. Nanotechnol., 2008, 3, 563 CrossRef CAS PubMed.
  14. C. E. Hamilton, J. R. Lomeda, Z. Sun, J. M. Tour and A. R. Barron, Nano Lett., 2009, 9, 3460 CrossRef CAS.
  15. L. Niu, J. N. Coleman, H. Zhang, H. Shin, M. Chhowalla and Z. Zheng, Small, 2016, 12, 272 CrossRef CAS PubMed.
  16. A. O'Neil, U. Khan, P. N. Nirmalraj, J. Boland and J. N. Coleman, J. Phys. Chem. C, 2011, 115, 5422 CrossRef.
  17. A. Ciesielski and P. Samori, Chem. Soc. Rev., 2014, 43, 381 RSC.
  18. M. T. Popescu, D. Tasis and C. Tsitsilianis, ACS Macro Lett., 2014, 3, 981 CrossRef CAS.
  19. T. Skaltsas, N. Tagmatarchis and S. Pispas, Curr. Org. Chem., 2015, 19, 1800 CrossRef CAS.
  20. M. T. Popescu, D. Tasis, K. D. Papadimitriou, S. Gkermpoura, C. Galiotis and C. Tsitsilianis, RSC Adv., 2015, 5, 89447 RSC.
  21. A. Ciesielski and P. Samori, Adv. Mater., 2016, 28, 6030 CrossRef CAS PubMed.
  22. A. Amiri, M. Naraghi, G. Ahmadi, M. Soleymaniha and M. Shanbedi, FlatChem, 2018, 8, 40 CrossRef CAS.
  23. Y. T. Liang and M. C. Hersam, J. Am. Chem. Soc., 2010, 132, 17661 CrossRef CAS PubMed.
  24. L. Xu, J. W. McGraw, F. Gao, M. Grundy, Z. Ye, Z. Gu and J. L. Shepherd, J. Phys. Chem. C, 2013, 117, 10730 CrossRef CAS.
  25. R. Shah, C. Wolfe, K. M. Lewis, M. Washington, S. K. Nayak, S. Talapatra and S. Kar, Nano Lett., 2010, 10, 4295 CrossRef.
  26. A. Schlierf, H. Yang, E. Gebremedhn, E. Treossi, L. Ortolani, L. Chen, A. Minoia, V. Morandi, P. Samori, C. Casiraghi, D. Beljonne and V. Palermo, Nanoscale, 2013, 5, 4205 RSC.
  27. M. Varenik, M. J. Green and O. Regev, Langmuir, 2016, 32, 10699 CrossRef CAS.
  28. D. Parviz, S. Das, H. S. Tanvir Ahmed, F. Irin, S. Bhattacharia and M. J. Green, ACS Nano, 2012, 6, 8857 CrossRef CAS.
  29. X. Zheng, Q. Xu, J. Li, L. Li and J. Wei, RSC Adv., 2012, 2, 10632 RSC.
  30. A. Schlierf, H. Yang, E. Gebremedhn, E. Treossi, L. Ortolani, L. Chen, A. Minoia, V. Morandi, P. Samorì, C. Casiraghi, D. Beljonne and V. Palermo, Nanoscale, 2013, 5, 4205 RSC.
  31. H. Yang, Y. Hernandez, A. Schlierf, A. Felten, A. Eckmann, S. Johal, P. Louette, J. J. Pireaux, X. Feng, K. Mullen, V. Palermo and C. Casiraghi, Carbon, 2013, 53, 357 CrossRef CAS.
  32. M. Zhang, R. R. Parajuli, D. Mastrogiovanni, B. Dai, P. Lo, W. Cheung, R. Brukh, P. L. Chiu, T. Zhou, Z. Liu, E. Garfunkel and H. He, Small, 2010, 6, 1100 CrossRef CAS PubMed.
  33. C. Backes, C.D Schmidt, K. Rosenlehner, F. Hauke, J. N. Coleman and A. Hirsch, Adv. Mater., 2010, 22, 788 CrossRef CAS PubMed.
  34. A. Liscio, K. Kouroupis-Agalou, A. Kovtun, E. Gebremedhn, M. El Garah, W. Rekab, E. Orgiu, L. Giorgini, P. Samori, D. Beljonne and V. Palermo, ChemPlusChem, 2017, 82, 358 CrossRef CAS.
  35. J. Liu, W. Yang, L. Tao, D. Li, C. Boyer and T. P. Davis, J. Polym. Sci., Part A: Polym. Chem., 2010, 48, 425 CrossRef CAS.
  36. F. Liu, J. Y. Choi and T. S. Seo, Chem. Commun., 2010, 46, 2844 RSC.
  37. D. W. Lee, T. Kim and M. Lee, Chem. Commun., 2011, 47, 8259 RSC.
  38. H. Wang, Z. Chen, L. Xin, J. Cui, S. Zhao and Y. Yan, J. Polym. Sci., Part A: Polym. Chem., 2015, 53, 2175 CrossRef CAS.
  39. D. Duan, H. Ye, N. Meng, C. Xu, B. Han, Y. Chen, M. Zhong and L. Xu, Carbon, 2018, 136, 417 CrossRef CAS.
  40. K. D. Papadimitriou, E. N. Skountzos, S. S. Gkermpoura, I. Polyzos, V. G. Mavrantzas, C. Galiotis and C. Tsitsilianis, ACS Macro Lett., 2016, 5, 24 CrossRef CAS.
  41. Z. Xu and C. Gao, Mater. Today, 2015, 18, 480 CrossRef CAS.
  42. I. S. Chronakis, J. Mater. Process. Technol., 2005, 167, 283 CrossRef CAS.
  43. F. Barzegar, A. Bello, M. Fabiane, S. Khamlich, D. Momodu, F. Taghizadeh, J. Dangbegnon and N. Manyala, J. Phys. Chem. Solids, 2015, 77, 139 CrossRef CAS.
  44. S. Ghobadi, S. Sadighikia, M. Papila, F. C. Cebeci and S. A. Gursel, RSC Adv., 2013, 5, 85009 RSC.
  45. S. Panzavolta, B. Bracci, C. Gualandi, M. L. Focarete, E. Treossi, K. Kouroupis-Agalou, K. Rubini, F. Bosia, L. Brely, N. M. Pugno, V. Palermo and A. Bigi, Carbon, 2014, 78, 566 CrossRef CAS.
  46. Y. Wang, J. Tang, S. Xie, J. Liu, Z. Xin, X. Liu and L. A. Belfiore, RSC Adv., 2015, 5, 42174 RSC.
  47. T. Huang, M. Lu, H. Yu, Q. Zhang, H. Wang and M. Zhu, Sci. Rep., 2015, 5, 13942,  DOI:10.1038/srep13942.
  48. Y. Li, H. Porwal, Z. Huang, H. Zhang, E. Bilotti and T. Peijs, J. Nanomater., 2016, 4624976,  DOI:10.1155/2016/4624976.
  49. Q. Bao, H. Zhang, J. X. Yang, S. Wang, D. Y. Tang, R. Jose, S. Ramakrishna, C. T. Lim and K. P. Loh, Adv. Funct. Mater., 2010, 20, 782 CrossRef CAS.
  50. A. Moayaeri and A. Ajji, J. Nanosci. Nanotechnol., 2017, 17, 1820 CrossRef.
  51. W. S. Khan, R. Asmatulu, V. Rodriguez and M. Ceylan, Int. J. Energy Res., 2014, 38, 2044 CrossRef CAS.
  52. A. A. Madhavan, S. Kalluri, D. K. Chacko, T. A. Arun, S. Nagarajan, K. R. V. Subramanian, A. S. Nair, S. V. Nair and A. Balakrishnan, RSC Adv., 2012, 2, 13032 RSC.
  53. H. Durmaz, A. Dag, U. Tunca and G. Hizal, J. Polym. Sci., Part A: Polym. Chem., 2012, 50, 2406 CrossRef CAS.
  54. C. Ahlborn, K. Siegmund and C. Richert, J. Am. Chem. Soc., 2007, 129, 15218 CrossRef CAS.
  55. H. Durmaz, A. Dag, O. Altidas, T. Erdogan, G. Hizal and U. Tunca, Macromolecules, 2007, 40, 191 CrossRef CAS.
  56. R. Tejero, D. López, F. López-Fabal, J. L. Gómez-Garcés and M. Fernández-García, Polym. Chem., 2015, 6, 3449 RSC.
  57. L. Xu and X. Yang, J. Colloid Interface Sci., 2014, 418, 66 CrossRef CAS.
  58. X. Xu, L. Cai, X. Zheng and Q. Xu, Phys. Chem. Chem. Phys., 2017, 19, 16062 RSC.
  59. S. Lin and M. J. Buehler, Nanotechnology, 2013, 24, 165702 CrossRef.
  60. Y. Zhang, C. Liu, W. Shi, Z. Wang, L. Dai and X. Zhang, Langmuir, 2007, 23, 7911 CrossRef CAS PubMed.
  61. E. N. Skountzos, A. Anastassiou, V. G. Mavrantzas and D. N. Theodorou, Macromolecules, 2014, 47, 8072 CrossRef CAS.
  62. L. M. Malard, M. A. Pimenta, G. Dresselhaus and M. S. Dresselhaus, Phys. Rep., 2009, 473, 51 CrossRef CAS.
  63. A. C. Ferrari, J. C. Meyer, V. Scardaci, C. Casiraghi, M. Lazzeri, F. Mauri, S. Piscanec, D. Jiang, K. S. Novoselov, S. Roth and A. K. Geim, Phys. Rev. Lett., 2006, 97, 187401 CrossRef CAS.
  64. Y. Shin, M. Lozada-Hidalgo, J. L. Sambricio, I. V. Grigorieva, A. K. Geim and C. Casiraghi, Appl. Phys. Lett., 2016, 108, 221907 CrossRef.
  65. Y. Hernandez, V. Nicolosi, M. Lotya, F. M. Blighe, Z. Sun, S. De, I. T. McGovern, B. Holland, M. Byrne, Y. K. Gun'Ko, J. J. Boland, P. Niraj, G. Duesberg, S. Krishnamurthy, R. Goodhue, J. Hutchison, V. Scardaci, A. C. Ferrari and J. N. Coleman, Nat. Nanotechnol., 2008, 3, 563 CrossRef CAS PubMed.
  66. D. G. Papageorgiou, I. A. Kinloch and R. J. Young, Prog. Mater. Sci., 2017, 90, 75 CrossRef CAS.
  67. N. Domun, H. Hadavinia, T. Zhang, T. Sainsbury, G. H. Liaghat and S. Vahid, Nanoscale, 2015, 7, 10294–10329 RSC.


Electronic supplementary information (ESI) available. See DOI: 10.1039/c8nr06888g

This journal is © The Royal Society of Chemistry 2019