Surface Modification of Carbon Nanofibers and Graphene Platelets Mixtures by Plasma Polymerization of Propylene

Carbon nanofibers (CNFs), graphene platelets (GPs), and their mixtures were treated by plasma polymerization of propylene. The carbon nanoparticles (CNPs) were previously sonicated in order to deagglomerate and increase the surface area. Untreated and plasma treatedCNPswere analyzed by dynamic light scattering (DLS), transmission electronmicroscopy (TEM), Fourier transform infrared spectroscopy (FTIR), Raman spectroscopy, and thermogravimetric analysis (TGA). DLS analysis showed a significant reduction of average particle size, due to the sonication pretreatment. Plasma polymerized propylene was deposited on the CNPs surface; the total amount of polymerized propylene was from 4.68 to 6.58wt-%. Raman spectroscopy indicates an increase in the sp hybridization of the treated samples, which suggest that the polymerized propylene is grafted onto the CNPs.


Introduction
Carbon atoms have a specific behavior that allows different electron configurations, known as hybridization, sp 2 hybridization being responsible for the formation of twodimensional sheets formed by carbon atoms arranged in hexagons, known as graphene [1].This hexagonal pattern is responsible for its unique properties such as high electrical and thermal conductivity and high mechanical strength [2][3][4], making it ideal to improve the properties of several polymer matrices.Graphene is considered the basic structure of various allotropes of carbon, such as carbon nanotubes (CNTs), fullerenes, carbon nanofibers (CNFs), and graphene platelets (GPs) [1,[5][6][7][8].
One of these allotropes that has generated particular interest, specifically for industrial scale, is the CNFs, due to their low cost and excellent physicochemical properties.Besides, they are easy to process and disperse in different polymer systems [9,10].Typically, these nanofilaments formed by stacked graphene sheets from the catalytic decomposition of certain hydrocarbons have diameters ranging from 2 to 100 nm and lengths from 5 to 100 m [11,12].Another allotrope of interest is the GPs, which are constituted by a low number of stacked graphene sheets (reaching up to 10 nm of thickness) [13] forming a planar structure.This structure and its ultrahigh aspect ratio provide to GPs extraordinary properties such as high thermal conductivity and high mechanical strength [6,14].
The addition of CNPs into polymer matrices has shown a marked improvement in the physicochemical properties of the nanocomposites; moreover, there is a tendency to add a mixture of two or more different carbon nanoparticles aiming to obtain a synergistic effect in the final properties of polymer nanocomposites [6,[15][16][17][18][19].For example, when one-dimensional carbon nanostructures (CNTs or CNFs) are intercalated with GPs or graphene (2D), a 3D network is promoted, which allows a greater interaction and a better dispersion and spatial distribution of the CNPs in the polymer matrix [18,20].On the other hand, surface modification or functionalization of carbon nanoparticles is a strategy to improve the interfacial adhesion with the matrix and also to get better dispersion in the polymer [21,22].In general, CNPs can be modified covalently or by physical interactions.In the first case, functional groups are bonded chemically on the surface of the CNPs [23], while electrostatic interactions or van der Waals forces [21] act as linkages in the second case.Considering the properties of the CNPs associated with its structure, a good selection of the modification method is necessary, to affect as little as possible the structure of the CNPs.In this sense, the surface modification of CNPs by means of plasma polymerization represents an innovative and sustainable alternative, because it is fast and simple and is performed in the absence of solvents (dry process).For example, CNPs have been coated with an ultrathin polypropylene layer when propylene monomer in the vapor phase was fed into and polymerized in a plasma reactor and deposited as an ultrathin film on the surface of the CNPs [24,25].On another study, Zhao et al. produced a change in the surface properties of CNTs via plasma treatment using an argon-oxygen gas mixture as feed into the plasma reactor.New chemical groups, such as C-O, grafted on the CNTs surface, and a reduction of the sp 2 -hybridization were observed [26].Neira-Velázquez et al. achieved the deposition of ultrathin polyacrylic acid film on the CNFs surface, and, according to the Raman results, it might be presumed that this nanofilm could be covalently bonded on the surface of nanofiber [27].However, a slight decrease in the sp 2 hybridization was observed after the plasma treatment, which, on the other hand, indicated that polymerized gases such as acrylic acid tend to produce a coating without significantly affecting the surface structure of the substrate.In a similar way, other researches have been treated CNPs by plasma polymerization using different monomers to coat carbon nanostructures [24,28,29].Plasma technology also has been used recently to modified graphene materials and novel properties have been found as a result of this plasma treatment such as photoluminescence [30] and semiconductivity [31] and selective detection for biosensors [32,33].
This study is focused on the surface modification of graphene platelets and carbon nanofibers mixtures by plasma polymerization using propylene monomer as feed into the plasma reactor in order to deposit an ultrathin polypropylene film on the surface of these carbon nanostructures.

Experimental Section
2.1.Materials.The CNFs were purchased from Pyrograph Products, Inc., USA, with an average diameter of 100 nm and a specific surface area of 41 m 2 g −1 , while GPs were obtained from Cheap Tubes, USA, with 1-3 m in diameter, 10-12 layers of graphene, and a specific surface area of 600-750 m 2 g −1 .Propylene monomer from Sigma-Aldrich was used as received.

Methodology.
The CNPs were first sonicated in a glass reactor, to induce the deagglomeration of these carbon nanoparticles [34][35][36], as follows: 1 g of CNPs (CNFs and GPs, as well as the CNFs : GPs mixtures, in the following ratios by weight: 9 : 1, 8 : 2, or 7 : 3) was introduced in a glass reactor arranged with a sonotrode (see Figure 1) and sonicated at 20 kHz for 30 min in an air atmosphere.To improve the efficiency of functionalization, magnetic agitation was included.
A plasma polymerization reactor, reported by Neira-Velázquez et al. [27], was used for the surface modification of the CNPs.The plasma treatment was carried out as follows: 2 g of CNPs was introduced in the glass reactor and the system was put under vacuum.Then, a flow of propylene was fed into the reactor as a result of maintaining a constant pressure (vacuum) of 20 Pa inside the reactor.The plasma treatment condition was 60 W for 60 min.

Characterization.
A Zetasizer Nano S90, Dynamic Light Scattering (DLS) instrument was used for evaluating particle size distribution of CNPs before and after sonication. 1 mg of each CNPs was added to 10 mL of xylene and mechanically stirred for the dispersion test.
A Magna Nicolet 550 infrared spectrophotometer (FTIR) was used for analysis of untreated and plasma modified CNPs within 4000 to 400 cm −1 .Previously, the samples were dried in a vacuum oven at 100 ∘ C for 15 h, and thereafter they were supported in KBr disks.
A TA Instruments thermogravimetric analyzer (TGA) was used to examine the thermal stability of samples within a temperature range of 30 to 600 ∘ C, at 10 ∘ C/min, under nitrogen flow of 50 mL/min.From 600 to 800 ∘ C, the nitrogen flow was changed to oxygen flow, at the same 50 mL/min, in order to favor the total oxidation of the sample.
For analysis by transmission electron microscopy (TEM), a TITAN 80-300 kV microscope equipped with field emission gun and with a symmetrical condenser-objective lens type S-TWIN (Cs = 1.33 nm) was used.Images obtained by HRTEM have been registered in a CCD camera near the Scherzer focus.To assess the hybridization degree of the untreated and treated CNPs, a Horiba Scientific Micro Raman Xplora spectrophotometer was used from 1000 to 4000 cm −1 .A laser with 514 nm and 10x was used to realize the Raman analysis.

Results and Discussion
3.1.Deagglomeration of CNPs Analyzed by DLS. Figure 2 shows the particle size distribution (PSD) of the CNPs before and after sonication.Samples before sonication are labeled as CNFs, GPs, and the corresponding CNFs : GPs mixtures as 9 : 1, 8 : 2, and 7 : 3. Sonicated samples are labeled the same, but followed by [S].
It can clearly be observed that, in all cases, the particle size at the peak decreases with the sonication treatment; and, in all cases as well, the width of the particle size distribution curves decreases with the sonication treatment.This might indicate that sonication produces a certain degree of size homogenization as well as a smaller average particle size.This decrease in average particle size is associated with an increase in the particle Theoretical Surface Area (TSA).TSA was calculated using where  =  2 and "" is the TSA of 1 nanoparticle with diameter "". = 1/ and "" is the total volume of 1 g of CNPs with density "". = (4/3) 3 and "" is the volume of 1 nanoparticle with radius "".Thus, / is equal to the number of CNPs per 1 g of sample.
For calculations with (1), all particles were assumed as spherical, and densities were taken from the supplier as  CNFs = 1.95 × 10 6 g m −3 and  CPs = 0.04 × 10 6 g m −3 .The densities of CNFs and GPs mixtures were calculated from these.The results are summarized in Table 1.
Graphene deagglomeration by means of sonication in an solvent medium has been studied by several authors [37][38][39].However, in our air atmosphere sonicate system, the PSD of (a) TSA: theoretical surface area; (b) percent increase in TSA after sonication, in parenthesis.

GPs [S] [M]
GPs CNPs decreases, since the ultrasound waves are propagated through the gas (atmospheric air) and hit on the surface of the CNPs.The initial impact causes a suspension of the CNPs in the air and then a fluidized bed is formed.Within the fluidized bed, agglomerates absorbed the impact shock of ultrasound waves, creating a vibration in their body, resulting in exfoliation/fragmentation [36].
Such exfoliation is identified as CNPs with smaller average size in the DLS analysis (see Table 1).Also in Table 1, it is observed that the particle size decreases and the TSA increases, due to the increase in the number of particles per gram.With respect to the increment of TSA in percentage, the highest percentages were obtained for NPs combinations (specifically the combination of CNFs : GPs in a ratio of 7 : 3 reached an increase of 90%); this indicates a more effective deagglomeration treatment when CNFs and GPs are used together.It is suggested that impact shock of ultrasound waves occurs due to the increment in the interactions between the CNPs in the fluidized bed, probably due to the different morphology (size and shape) of the agglomerates in the system, encouraging the passage of the necessary ultrasound waves vibrations to deagglomerate all CNPs.

Infrared Analysis. FTIR spectra of sonicated CNFs and
GPs (marked with an [S] at the end), and sonicated plus plasma treated CNFs and GPs (marked with an [S] [M] at the end), are presented in Figure 3.The signals at 2800-3000 cm −1 , characteristic of a symmetric C-H stretching, appear in both spectra (with and without plasma treatment).However, a slightly higher intensity is observed in the plasma modified CNPs, due to a higher concentration of C-H bonding present on the surface of the CNPs.It is common to find signals of low intensity in the untreated CNPs, which are organic molecules (hydrocarbons) that remain on the surface of CNPs The band at 1633 cm −1 appears in the CNFs spectra, but not for the case of the GPs.This could be related to the adsorption of water from the environment [40].A noticeable increase is observed in the bands at 1450 and 1440 cm −1 after the plasma treatment, particularly in the GPs, attributed to the presence of -CH 2 -groups.Likewise, a noticeable increase is observed in the bands at 1255 cm −1 after the plasma treatment, particularly in the CNFs, which is associated with the appearance of tertbutyl groups.In addition, the band at 1098 cm −1 shows an increase, which is attributed to the vibrations produced by the stretching and swing modes of C-C, and bends and twists of C-H.Another increment is observed in the band at 660 cm −1 for trans-alkene disubstituted groups and at 800 cm −1 which indicates the vibrations of stretching of C-H and C-C, and oscillations of CH 2 [41].
All changes in these signals are associated with the chemical groups deposited on the surface on CNPs by the plasma treatment.Finally, a significant change is observed in the bands at 1740 cm −1 after the plasma treatment, particularly in the CNFs, indicating that they suffered oxidation either during or after the plasma treatment.This is a common phenomenon in plasma modification, because after treatment there are several active sites remaining on the surface of carbon nanoparticles, and, after treatment, they react with ambient oxygen.
The modification on the carbon nanoparticles made via the plasma treatment with propylene can be ascertained as a result of the changes in the FTIR signals above.However, none of these signals appears in the form of polypropylene (PP).According to the infrared analysis, it is suggested that the PP is in the form of a crosslinked homopolymer, due to the significant increase in the signal characteristic of the tertbutyl group.It is well known that the plasma polymerization promotes crosslinked structures, due to the multiple reactive species that are generated during the process.This is reflected in the decomposition temperature obtained by TGA, higher than the decomposition temperature of polypropylene (300 ∘ C) [42][43][44].

Thermal Stability.
In Figure 4, TGA are showed for the untreated and plasma treated CNFs, GPs, and the CNFs : GPs mixtures with an 9 : 1, 8 : 2, and 7 : 3 weight ratios.First, it is observed that there is a noticeable difference between the CNFs and GPs, either untreated or treated.With respect to the untreated carbon nanoparticles, when they are heated up to 600 ∘ C, the weight loss of the CNFs is almost 1%, whereas the weight loss of the GPs reaches up to 11%.The weight loss of the untreated GPs is assumed to be associated with organic material adsorbed on the particles surface during their synthesis.On the other hand, when the plasma treated carbon nanoparticles are heated up to 600 ∘ C, an additional weight loss of 4.68 and 6.58% is observed for CNFs and GPs, respectively.That is, when heated in a N 2 atmosphere, the weight loss of the GPs is greater than that of the CNFs.On the other hand, the weight loss of untreated mixtures of CNFs : GPs (9 : 1, 8 : 2, and 7 : 3) shows values of 9, 7, and 4%, respectively, while the additional weight losses observed in the treated samples are 4.92, 5.07, and 5.54%, respectively, probably due to the increase in the theoretical surface area of the mixtures (see Table 1).The additional weight loss in the plasma treated CNFs, GPs, and mixtures of them is entirely attributed to the polymeric material deposited on the particles surface during the propylene plasma treatment.5 presents TEM images of untreated and plasma treated CNPs. Figure 5(a) shows a CNFs segment in which a structure known as CNFs stacked cups (Endo et al. in 2003) can be observed [45].Shanmugam and Gedanken observed CNTs with similar structure, which they called "bamboo" structure [46].The flat surface is inclined with respect to the axial direction of the fiber, thereby exposing the edges of the plane of the CNFs, generating a greater surface area and facilitating the surface modification thereof.Furthermore, it is observed that the presence of a thin homogeneous layer of about 2 to 3 nm is associated with amorphous carbon on the surface of the CNFs, being this common due to the synthesis method used [47].At a lower magnification (Figure 5(b)), it can be observed that the surface of the CNFs is smooth and without any noticeable variations on the diameter along the fiber.On the contrary, Figure 5(c) shows a segment of plasma treated CNFs, where plasma polymerized PP is anchored on the CNFs surface and is denoted as clusters (marked with red arrows).However, these clusters have a higher density and break the apparent smoothness observed in the untreated CNFs (seen in Figure 5(b)) [48,49].Figure 5(d) shows an agglomerate of untreated GPs.Red circles indicate GPs with thickness less than 10 nm, whereas the squares indicate GPs with thickness more than 15 nm.Most GPs are below 10 nm in the agglomerate; the dark areas have a stack of GPs, while lighter areas indicate the presence of thinner stacked lamellae.In addition, it is noteworthy that the electron diffraction analysis of the CNPs, results in a characteristic interplanar distance of 3.5 Å between graphene layers [50,51].Figure 5(e) shows plasma treated GPs, where clusters (in red circles) might suggest the evidence of PP grafted on its surface.

Dispersion Test. Dispersion tests in xylene can provide
an idea if the plasma modification has been achieved on the CNPs surface.Figure 6 shows images of the untreated and plasma treated carbon nanoparticles analyzed in this study (CNFs, GPs, and the CNFs : GPs mixtures at 9 : 1, 8 : 2, and 7 : 3 ratios), taken after 24 h in xylene of 5 vials with untreated CNPs treated and another 5 with CNPs untreated.The untreated CNPs show no interaction with the solvent, and they sediment immediately.The plasma treated CNPs, on the other hand, remain dispersed for much longer times, suggesting an increase in interaction between the CNPs and the solvent, where probably methyl groups of polypropylene (PP) increase this interaction.
In addition, an increase was observed in the interaction of CNPs with the solvent when concentration of GPs gets higher, despite having similar degree of modification.

Hybridization
Degree by Raman Spectroscopy.The hybridization degree of the untreated and plasma treated samples was calculated via the  (in 1345 cm −1 ) and  (1575 cm −1 ) bands, using Raman spectroscopy.Several authors have proposed that the ratio of these intensity bands (  /  ) can be used for characterizing the degree of structural disorder present on the surface of the CNPs, which is important because it may indicate the degree of modification [52][53][54][55][56][57]. Figure 7 shows the Raman spectra, in the high frequency region (1100-1800 cm −1 ) of the untreated and plasma treated CNFs and GPs.CNFs signal shows a slight increase in the intensity of the  band, which is related to the amount of defects on the surface of the CNPs, while, for GPs, the  band remains without any significant change; however, a considerable decrease of the  band is produced, denoting a change of hybridization from sp 2 to sp 3 .Table 2 shows the values obtained by Raman spectroscopy for untreated CNPs (denoted by the [S] extension) and plasma     sites).Therefore, the type of surface modification (covalent or physical) depends of the number of sites that can be chemically modified.
The hybridization change and the amount of polymer deposited are related by the TSA of the CNPs.A higher TSA not only represents a larger area where the polymer can be deposited but also represents a greater exposure of sp 2 hybridization points by the CNPs allowing covalent bonds between the polymer and the surface of the CNPs (denoted by the hybridization change).For this reason, GPs with larger TSA than CNFs generate more covalent interaction with the polymer, as well as a higher percentage of polymer to be deposited on its surface.

Conclusions
The results obtained by sonication in gas phase of the CNPs indicated a lower agglomeration, generating CNPs with more TSA; specifically the combination of CNFs : GPs in a ratio of 7 : 3 reached an increase of 90%, indicating a more effective deagglomeration treatment when CNFs and GPs are used together.
A thin film of PP was covalently deposited by plasma polymerization on the surface of carbon nanoparticles.This is reflected in the hybridization change from sp 2 to sp 3 by Raman spectroscopy, among other techniques.The degree of deposition is analyzed with the ratio of   /  bands.The band analysis shows greater covalent interaction for the GPs and the CNFs : GPs mixtures, due to the higher TSA exposed (or exposed sp 2 hybridization sites) in relation to the CNFs.The amount of deposited polymer coating is not reflected directly in the hybridization change (from sp 2 to sp 3 ).This change reflects only the covalent interaction of the coating and CNPs, regardless of the amount.

Figure 3 :
Figure 3: FTIR spectra of untreated and plasma treated of CNPs.

Figure 7 :Figure 8 :
Figure 7: Raman spectroscopy of treated and untreated CNFs and GPs.

Table 1 :
Particle size at the peak of the size distribution curves of Figure2and theoretical surface area of each particle group.

Table 2 :
Summary of   /  values of all samples obtained by Raman spectroscopy.