Spatial structure and aggregation of carbon allotrope nanofillers in isotactic polypropylene composites studied by small-angle neutron scattering
L.V. Elnikova, A.N. Ozerin, V.G. Shevchenko, P.M. Nedorezova, A.T. Ponomarenko, V.V. Skoi, A.I. Kuklin
SSpatial structure and aggregation of carbon allotrope nanofillers in isotacticpolypropylene composites studied by small-angle neutron scattering
L.V. Elnikova , A.N. Ozerin , V.G. Shevchenko , P.M. Nedorezova , A.T. Ponomarenko , V.V. Skoi , A.I. Kuklin NRC “Kurchatov Institute” – Alikhanov Institute for Theoretical and Experimental Physics,117218 Moscow, Russia, Enikolopov Institute of Synthetic Polymeric Materials, RAS, 117393 Moscow, Russia, Semenov Institute of Chemical Physics, RAS, 119991 Moscow, Russia, Joint Institute for Nuclear Research, 141980 Dubna, Russia Moscow Institute of Physics and Technology, 141701 Dolgoprudny, Russia
Abstract
We study the aggregation of carbon allotrope nanofillers in the matrix of isotactic polypropylenewith direct small-angle neutron scattering measurements. With the ATSAS software, we analyzedthe data and determined the fractal shape, dimension, and sizes of nanofiller aggregation in thebulk of isotactic polypropylene over the range of the scattering angles. We estimated the volumedistributions and aggregation of different types of carbon nanofillers at different concentrations:nanographite, graphene nanoplatelets (GNP), single-walled carbon nanotubes (SWCNT), multi-wall carbon nanotubes (MWCNT), binary fillers MWCNT/GNP and fullerenes. Wereconstructed the shape of nanoscale aggregates of all nanofillers and found that the systems arepolydisperse; nanofillers associate in the volume of iPP as fractal dense aggregates with ruggedsurface, their sizes exceeding original dimensions of nanofillers several times.
Key words : polypropylene, carbon allotrope nanofillers, Small-Angle Neutron Scattering, fractalobjects, aggregates, morphology
1. Introduction
The polymers filled with carbon allotropes have wide applications in different fields of humanactivity: industry, medicine, electronics and spintronics, sensing materials, lubricants, they areused as elements of electronic devices and actuators etc.[1,2]. The morphology of the fillers inpolymer matrix determines functional properties of such composites when carbon allotropes areadded, e.g. electrical and mechanical properties can change [3], these changes are also anifested due to size effects. Doping of polymer with carbon nanofillers is accompanied bymodification of surface, changes in polymer structure and interactions between nanofillers.Physical, chemical and structure properties of filled polymer nanocomposites can beinvestigated with different techniques, e.g. with dielectric spectroscopy, atomic force microscopy(AFM), differential scanning calorimetry (DSC), scanning electron microscopy (SEM), Ramanscattering, Small-Angle X-ray scattering (SAXS) [3,4,5] etc . However, during sample preparation one gets, as a rule, cleavages and cuts of thematerials, that damage and distort their bulk structure. These additional defects result in a loss ordistorted information about the test specimen. Whereas the SANS method providesnondestructive structure analysis, with length scale from 1 Å, while explicitly characterizing themorphology of nanofillers in the bulk of the material [4]. There is extensive SANS data for polymer nanocomposites [6-10], though very little ifany information exists about iPP nanocomposites, which have specific chemical composition andbulk configuration of carbon allotrope nanofillers.Various incorporated inhomogeneities in a polymer matrix correlate with the changes ofthe SANS curves (peak position, shape, slope, etc. ) as compared to basic polymer. Forinterpretation of the experimental SANS curves, many models for identification of structures areapplicable in frames of well-developed non-linear least square fitting models [4,11-14] for thePorod and Guinier scattering laws. In the bulk of polymers, nanofillers may form multilevelfractal objects [7-10, 13]; such aggregation changes the slope of the Guinier functions atappropriate scales of scattering angles.Also, modern methods of small-angle scattering curves processing are supplemented withcalculating procedures [11, 14] that make it possible to visualize the shape of particles in thebulk, to identify various morphologies of single nanoparticles as well as their aggregation, todistinguish mono- or polydispersity of the system etc . To reconstruct the shape of an object in thebulk from scattering spectra, the regularization methods are usually employed [12]. For example,in [11] and references therein, the capabilities the ATSAS program are reported.The goal of our SANS experiments is to characterize the nanoscale effects in iPP matrixinduced by the presence of carbon allotrope nanofillers (nanographites, GNPs, fullerenes, CNTsand their combinations) and to describe morphology of the formed nanoobjects. We are alsomotivated in possibility of further complex studies of physical properties of these compositesinduced by carbon nanofiller addition; basing on these SANS data, we can examine the dielectricand structure changes of these materials with electron spectroscopy, positron annihilationspectroscopy and other techniques. . Materials and methods The samples of isotactic polypropylene (iPP) are filled by in situ polymerization with GNP atconcentrations of 0.7 and 1.8 wt%, nanographite at 1.5 and 3.6 wt%, SWCNT at 1.2, 2.6 and 8wt%, MWCNT at 3.5 wt%, MWCNT/GNP at 0.48, 0.9, 1.16 and 3 wt% and fullerenes at 16.5 wt%(the volume ratio for binary filler MWCNT:GNP is 1:2).The chemical formula of iPP is (C H ) n , its density is 0.9–0.91 g/cm , and the degree ofcrystallinity is 60%. Nanographite particles are in the form of plates with diameter 112.7 nm andthickness of 47.3 nm.Graphene nanoplatelets (GNP) were produced by chemical or thermal reduction ofgraphite oxide (TRGO) [15].Graphite oxide was produced using modified method of Hammers - oxidizing graphite byKMnO in concentrated H SO [16-17].X-ray diffraction analysis of GNP and TRGO powders was made using ADP-1diffractometer [18]. For GNP and TRGO, the values of crystallite size were calculated to be1.127 and 1.003 nm, respectively. Accordingly, synthesized few-layer particles are estimated tocontain 3-5 layers of graphene. The approximate dimensions of individual GNP particle is 100nm × 100 nm × 1.127 nm. Pristine CVD-grown MWCNTs (purity C95%, average diameter < 10 nm, length range5–15µm) were purchased from Shenzhen Nanotech Port Co., Ltd., China (trade name of productis L-MWNTs-10). As-received MWCNTs were purified and mildly oxidized by boiling 30 wt%nitric acid for 1 h with subsequent settling at room temperature for 20 h. This procedure wascarried out in order to remove rest of amorphous carbon and impurities that might be poisonousfor metallocene catalysts and to increase content of carboxylic and hydroxyl groups onMWCNTs. The acid-treated MWCNTs were filtered and washed repeatedly with deionizedwater, dried in vacuum at 400°C for 5 h, and then stored in argon atmosphere. Diameter of SWCNTs is 1.4 nm, and length is more than 5 µm.Synthesis of nanocomposites was done in bulk propylene, as described in [18-19]. Thismetallocene catalyst used in the synthesis of composites is highly active and isospecific inpropylene polymerization, producing iPP of high molecular weight [20]. The process wasconducted at 60°С and pressure 2.5 MPa in a steel 200 cm reactor vessel, equipped with a high-speed stirrer (3000 rev./min). Nanocomposites were synthesized via the following routes: 1)powder of carbon nanofiller (GNP, nanotubes or C ), previously evacuated at 200°С, was fedinto reactor vessel, which was then filled with liquid propylene (100 ml), methylalumoxane andmetallocene catalyst; 2) GNP or TRGO was prepared as suspension in toluene and sonicated for10 min, then methylalumoxane was added and sonication continued for 10 more minutes. ltrasonic power was ~ 35W. Afterwards the suspension was fed into reactor vessel, filled withliquid propylene and catalyst was finally added. Concentration of filler in composites was varied by changing polymerization time. Finalproduct was unloaded from the reactor, washed successively by a mixture of ethyl alcohol andHCl (10% solution), ethyl alcohol and then dried in vacuum at 60°C until constant weight.Test specimen were cut from films 100–300 mm thick, pressure molded at 190°C andpressure 10 MPa at cooling rate 16 K·min -1 .The SANS measurements were performed using the YuMO spectrometer at the IBR2reactor in Dubna, Russian Federation [23].The neutron wavelength is λ = 0.7–6Å, the neutron flux on a sample was about 10 n/(s×cm ) [24], diameter of neutron beam on the sample was 14 mm. The solid film-like specimenswith different nanofillers iPP/GNP, iPP/nanographite, iPP/SWCNT, iPP/MWCNT,iPP/MWCNT/GNP, iPP/fullerene were fixed in the holder. The thickness values for thespecimens were normalized to thickness of iPP film 368 μm. Holder was put into thermo box. In the SANS measurements with YuMO, we recorded counts versus time of flight fromthe 16 rings of two detectors. Recalculation and normalization count using the gauge standard ofthe known cross section vs time of flight to the differential scattering cross section dΣ/dΩ(Q) andnormalization on the sample thickness was realized by program SAS[25 ] . The preliminarilyevaluated scattering length density (SLD) for the samples is in the range 3.84 × 10 – 5 × 10 cm -1 [4].
3. Results and analysis of SANS spectra
To analyze the experimental small-angle scattering curves, we used a number of the followingprocedures of the ATSAS 2.4 software package [11].Preliminary processing of the initial scattering curves and registration of scattering by thereference sample were performed using the PRIMUS procedure of ATSAS [11]. In this work, asthe reference scattering, which was subtracted from the experimental curve of small-anglescattering of samples
I(Q) , scattering from a sample of matrix polymer (iPP) was used. Thus,after taking into account reference scattering, the experimental small-angle scattering curves arecharacterized by scattering from only heterogeneous regions ("scattering particles") in the systemhaving a scattering length different from the scattering length of the polymer matrix.To calculate regularized scattering curves I reg (Q) , optimized over the entire range of scatteringangles, the particle distribution function, the integral values of the inertia radii of the particles ofthe scattering phase and the particle size distribution, we used the GNOM procedure of ATSAS ased on the regularization method according to Tikhonov [12].To determine the shape and spatial structure of scattering particles, we used an approach basedon the use of well-founded algorithms for reconstructing the shape of scattering particles fromsmall-angle scattering data, implemented in the DAMMIN and DAMMIF procedures of ATSAS,the algorithm of which uses the Monte Carlo method with the annealing procedure forreconstructing the shape of scattering particles in the framework of the model of "virtual"(dummy) atoms. The structures recovered in individual runs were averaged using theDAMAVER and SUPCOMB procedures [26]. The experimental values of the intensity
I(Q) of small-angle neutron scattering and theregularized small-angle X-ray scattering curves I reg ( Q ) calculated in accordance with the GNOMprocedure for samples nanographite 1.8, 3.6 wt% and GNP 0.7, 1.8 wt%; (excluding scatteringfrom iPP as a reference sample) are shown in Fig. 1.Fig. 1. The experimental SANS intensity I(Q) of the samples in the coordinates
I–Q (left) andlog( I )–log( Q ) (right). 1 – nanographite 3.6 wt%; 2 – nanographite 1.5 wt%; 3 – GNP 1.8 wt%; 4– GNP 0.7 wt%. The solid lines correspond to the regularized I reg ( Q ) curves. The scattering curveof the matrix polymer iPP is shown for comparison. The value k characterizes the slope of thelinear sections of the scattering curves in the log( I ) – log( Q ) coordinates.From Fig. 1. we conclude that scattering particles have an almost identical spatialstructure and their scattering pattern corresponds to scattering by a physical fractal object withdimension d s = 6 – │ k │ = 2.5, corresponding to a surface fractal. We reveal the dense compactaggregated particles with a rugged surface [13]. The upper size range of these physical fractalsexceeds the spatial resolution of the small angle neutron scattering method L max = 2π/ Q min = 94nm, implemented in the experiments of this work ( Q min = 0.0665 nm -1 ). s there are no interference effects and the curves are of diffuse nature, small-anglescattering by a dilute or polydisperse system of particles can be interpreted with minimal detailof the scattering system, [4] either in the scattering approximation from a polydisperse system ofparticles with a known form factor (balls, prisms, cylinders, volume ellipsoids of revolution, etc. ), or in the scattering approximation from identical particles of unknown shape and spatialstructure. In the first case, the resulting interpretation of the small-angle scattering data is thereconstructed particle size distribution function, and in the second case, the determination of theshape and size of the particles. Note, that the possibility of restoring the low-resolution structureof polydisperse and polymorphic nano-objects having more than one hierarchical level ofstructural organization (particles - particle aggregates) from small-angle scattering data using theATSAS software package was demonstrated earlier in [5,14].Since the GNP and nanographite samples used in the work contain particles in the formof plates, the regularized scattering curves from these samples were primarily used to calculatethe distribution function of particle thicknesses under the assumption of a polydisperse system ofplate-shaped particles with thickness T (the distance distribution function of thickness, assuminga polydisperse system of flat particles). The calculation results are shown in Fig. 2.Fig. 2. Normalized distribution functions of particle thicknesses D N (thin lines) under theassumption of a polydisperse system of plate-shaped particles with a thickness T calculated fromthe scattering curves for GNP and nanographite particles. The thick line is the smoothed curve,with adjacent averaging, weighted average value.Similar to the scattering curves, the distribution functions for the GNP and nanographitesamples turned out to be identical to each other. The system of scattering particles in the GNP nd nanographite samples is characterized by high polydispersity. The particle size distributioncontains wafers with a thickness of 1 to 20 nm, whereas the thickness of the initial plates are47.3 nm and 1 nm for nanographite and GNP samples, respectively.The radius of gyration R t = T /12 of the particle thickness determined from the slope ofthe linear part of the ln (Q I(Q))-Q plot (the Guinier plot) in the reciprocal space and thatcalculated by the indirect transform method [4] applied to the whole experimental scatteringcurve while using GNOM procedure, were close to each other and equal, on average, to 5.5 nm.
The experimental values of the intensity
I(Q) of small-angle neutron scattering and theregularized small-angle X-ray scattering curves I reg (Q) of the sample with 16.5wt% fullerenecalculated with the GNOM procedure (excluding the scattering from iPP as a reference sample)are shown in Fig. 3 in the I – Q and log( I ) – log( Q ) coordinates.Fig. 3. The experimental SANS intensity I(Q) of the sample fullerene 16.5 wt%.From Fig. 3 (right), we conclude that the scattering by fullerene particlesof 16.5wt%concentration corresponds to scattering by a physical fractal object with dimension d s = 2.9, thisis a surface fractal, where dense compact particleshave a strongly rugged surface [13]). Theupper size range of this physical fractal significantly exceeds the spatial resolution of the small-angle neutron scattering method L max = 2π/ Q min = 94 nm, realized in the experiments of this work.The regularized scattering curve from the fullerene sample 16.5wt% was used to calculatethe particle size distribution function assuming a polydisperse system of particles of a sphericalshape with a radius R in the form of the volume distribution function of hard spheres. Thecalculation results are shown in Fig. 4. ig. 4. Volume size particle distribution functions D V (R) for the fullerene 16.5 wt% scatteringcurve calculated under the assumption of a polydisperse system of spherical particles with radius R . The thin line denotes the calculated curve; and the thick one is the smoothed curve withadjacent averaging, weighted average value.The radius of gyration R g of the particle determined from the slope of the linear part ofthe ln(I(Q))-Q plot (the Guinier plot) in the reciprocal space and that calculated by the indirecttransform method [4] applied to the whole experimental scattering curve while using the GNOMprocedure were close to each other and equal, on the average, 29 nm.The system of scattering particles in the fullerene 16.5wt% sample is characterized byhigh polydispersity. The particle size distribution contains spherical formations with a radius of 1to 40 nm. The particles are aggregated. Note, that diameter of fullerene C is 0.7 nm. The experimental values of the intensity
I(Q) of SANS and the regularized small-angle X-rayscattering curves I reg (Q) of the sample MWCNT 3.5wt% calculated with the GNOM procedure(excluding scattering from iPP as a reference sample) are shown in Fig. 5. ig. 5. The experimental SANS intensity I(Q) of the sample MWCNT 3.5 wt% in the coordinates I – Q (left) and log( I )–log( Q ) (right). The label “1” denotes scattering from the MWCNT particles;“2” denotes scattering from aggregates of the MWCNT particles.According to the procedure [5, 14], the MWCNT 3.5 wt% scattering curve (Fig. 5.) wasdivided into two regularized components, which relate to scattering by MWCNT particles (curve“1”) and scattering from MWCNT aggregates (curve “2”).The relative content of aggregated and non-aggregated components in the system wasestimated from the integrated intensity of the scattering components in the coordinates of thescattering intensity ( IQ )–wave vector ( Q ). The calculated volume fractions of MWCNTs in theparticle and aggregate forms were found to be equal to each other (~ 0.5). Both components ofthe SANS curve were analyzed independently. The scattering curves “1” and “2” have been used to calculate the distance distributionfunction of the cross-section, assuming a polydisperse system of rod-like particles, as cylinderswith radius R (Fig. 6). ig. 6. The distribution function D N (R) of the MWCNT 3.5 wt% particle cross sections under theassumption of a polydisperse system of cylindrical particles with a radius R , calculated forscattering on the particles ”1” and on the aggregates of particles ”2” . The thin line denotes thecalculated curve; and the thick one is the smoothed curve with adjacent averaging, weightedaverage value.The gyration radius R c of the particle cross-section determined from the slope of thelinear part of the Guinier plot ln (qI(q))-q in the reciprocal space and calculated by the indirecttransform method [5] applied to the whole experimental scattering curve while using the GNOMprocedure, were close to each other and equal, on the average, to 3.8 and 11.4 nm for particlesand aggregates, respectively, whereas the initial MWCNT diameter is 10 nm. The results of reconstructing the shape and spatial structure of the system of particles andparticle aggregates calculated from the regularized scattering curves of the MWCNT 3.5 wt%sample using the DAMMIN and DAMMIF procedures are presented in Fig. 7. Particle shapereconstuction was performed without any additional restrictions imposed on the expectedsymmetry and anisometry of the particles. ig. 7. The shape of the scattering particles calculated from the scattering curve of particles ”1”(left) and aggregates of particles “2” (right). Visualization was performed with the model of bulkvirtual ("dummy") atoms (right) and with the model of the surface accessible to solvent (left).It is known [27] that small-angle scattering of a CNT system dispersed in a liquidmedium or a polymer matrix is usually considered as scattering from some disordered two-levelstructure, in which the level of large characteristic sizes is referred to as CNT aggregates, and thelevel of small characteristic sizes, in turn, to straightened CNT fragments (analogue of the kineticsegment for the polymer chain), the persistent length of which is significantly less than thecontour length of the CNT.In this regard, the reconstructed form presented in Fig. 7 (left) for a scattering particle (aMWCNT fragment) in the form of an elongated cylinder with D = 5–10 nm and a length L = 30nm is the expected result. In turn, according to results shown at Fig. 7 (right), the shape ofscattering particles (aggregates) can most easily be interpreted as the “entanglements” ofneighboring CNTs, analogous to similar “entanglements” of polymer macromolecules atconcentrations higher than the crossover concentration. According toFig.5, the shape ofscattering particles (aggregates) can most simply be interpreted as the “entanglements” ofneighboring CNTs, by analogy with similar “entanglements” of polymer macromolecules atconcentrations higher than the crossover concentration. In short, MWCNT system is similar tonon-woven fabric structure. The experimental values of the SANS intensity I ( Q ) of and the regularized SAXS curves I reg (Q) of the SWCNT 1.2, 2.6 and 8wt% samples(excluding scattering from iPP as a reference sample)calculated with the GNOM procedure are shown in Fig. 8. ig. 8. The experimental SANS intensity I ( Q ) for the samples SWCNT 1.2wt% (curves “1”),2.6wt%(curves “2”) and 8wt% (curves “3”) in the coordinates I – Q (left)and log( I )–log( Q )(right). The labels “1'” and “2'” denote scattering from SWCNT particles and from aggregates ofSWCNT particles respectively.Fig. 8 shows, that scattering particles have an almost identical spatial structure.According to the procedure [5, 14], we divided the scattering curves of the samplesSWCNT 1.2 wt%, 2.6 wt% and 8 wt% (Fig. 8., right) into two regularized components that relateto scattering by SWCNT particles and scattering by SWCNT aggregates. To simplify thepresentation, separation of the scattering curves into two components is shown in Fig. 8, right,only for sample SWCNT 8 wt%.The relative content of aggregated and non-aggregated components in the system wasestimated from the integrated intensity of the scattering components in the coordinates of thescattering intensity ( IQ ) – wave vector ( Q ). The calculated values of the volume fraction ofSWCNT in the form of particles and aggregates for all samples SWCNT 1.2 wt%, 2.6 wt% and 8wt% turned out to be equal to each other (~ 0.5). Both components of the scattering curve wereindependently analyzed.The regularized scattering curves for SWCNT particles and particle aggregates were usedto calculate the distribution function of particle cross sections under the assumption of apolydisperse system of cylindrical particles with a cylinder radius R (the distance distributionfunction of the cross-section assuming a polydisperse system of rod-like particles). Thecalculation results are shown in Fig.9. ig. 9. The distribution function D N (R) of the cross sections of the particles of the samplesSWCNT 1.2 wt%, (1), 2.6 wt% (2), and 8 wt% (3) under the assumption of a polydispersesystem of cylindrical particles with a radius R , calculated for scattering on particles (left); and foraggregates of particles (right). The thin line denotes the calculated curve; and the thick one is thesmoothed curve with adjacent averaging, weighted average value.The gyration radius R c of SWCNT particle cross-section determined from the slope of thelinear part of the Guinier plot ln (qI(q)) – q in the reciprocal space and calculated by the indirecttransform method [4] applied to the whole experimental scattering curve using the GNOMprocedure were equal to 5.6, 5.1, 5.8 nm for the samples SWCNT 1.2 wt%, SWCNT 2.6 wt%,SWCNT 8 wt%, respectively. Similarly, for the gyration radius R c of SWCNT particle aggregatescross-section, we have 12.3, 17.2, 18.2 nm for the same samples. The results of reconstructing of shape and spatial structure of the SWCNT particle systemand SWCNT particle aggregates calculated from the regularized scattering curves of the samplesSWCNT 1.2 wt%, SWCNT 2.6 wt%, SWCNT 8 wt% using the DAMMIN and DAMMIFprocedures are presented in Fig. 10. Particle shape reconstruction was performed without anyadditional restrictions imposed on the expected symmetry and anisometry of the particles. ig. 10. The shape of the scattering particles calculated from the scattering curve of SWCNTparticles (a–c) and SWCNT particle aggregates (d–e) of the samples SWCNT 1.2 wt% (a, d),SWCNT 2.6 wt% (b, d), SWCNT 8 wt% (c, f). Visualization with the model of bulk virtual("dummy") atoms and the model of a surface accessible to a solventThe reconstructed shape of the scattering particles reflects the quite expected compression of thespatial structure of the scattering system with increasing filler concentration. The SWCNTsystem can be described as a grid of individual tubes and tubes "stuck together" by the lateralsurfaces and rather large knots. The experimental values of the SANS intensity I ( Q ) of and the regularized SAXS curves I reg (Q) of the MWCNT/GNP 0.48 and 3 wt% samples (excluding scattering from iPP as a referencesample) calculated with the GNOM procedure are shown in Fig. 11. ig. 11. The experimental SANS intensity I(Q) for samples MWCNT/GNP 0.48 wt% (1) andMWCNT/GNP 3 wt% (2) in the coordinates I – Q (left) and log ( I ) – log ( Q ) (right). The label“1'” denotes scattering from particles MWCNT/GNP 0.48 wt%; “1''” is scattering from particleaggregates MWCNT/GNP 0.48w%; “2'” is scattering from particles MWCNT/GNP 3wt%; “2''”is scattering from MWCNT/GNP 3wt% particle aggregates.The scattering curves of the MWCNT/GNP 0.48 wt% and MWCNT/GNP 3 wt% samples(Fig. 11, right part) were divided into two regularized components [5, 14], which relate toscattering by particles and scattering by particle aggregates.The relative content of aggregated and non-aggregated components in the system wasestimated from the integrated intensity of the scattering components in the coordinates of thescattering intensity ( IQ ) – wave vector ( Q ). The calculated value of the volume fraction ofparticles and particle aggregates was 0.25/0.75 and 0.5/0.5 for samples MWCNT/GNP 0.48 wt%and MWCNT/GNP 3 wt% respectively. Both components of the scattering curve wereindependently analyzed for each of the samples.Since the shape of the scattering particles was not known for “mixed” MWCNT/GNPsamples, the regularized scattering curves for MWCNT/GNP particles and particle aggregateswere used to calculate the most generalized characteristic: the particle volume distributionfunction under the assumption of a polydisperse system of spherical particles with a cylinderradius R (the volume distribution function of hard spheres). The calculation results are shown inFig. 12. ig. 12. Volume distribution function of the D N (R) particles of MWCNT/GNP 0.48 wt% samples(1) and MWCNT/GNP 3 wt% (2) under the assumption of a polydisperse system of particles of aspherical shape with a radius R , calculated for scattering: on particles (left) and on aggregates ofparticles (right). Thin and thick lines mean respectively calculated and smooth curves, adjacentaveraging, weighted average.The results of reconstructing the shape and spatial structure of the system of particles andparticle aggregates calculated from the regularized scattering curves of the samplesMWCNT/GNP 0.48 wt% and MWCNT/GNP 3 wt% using the DAMMIN and DAMMIFprocedures are presented in Fig. 13. Particle shape reconstruction was performed without anyadditional restrictions imposed on the expected symmetry and anisometry of the particles. ig. 13. The shape of scattering particles calculated from the scattering curve of particles (a, b)and particle aggregates (c, d) of MWCNT/GNP 0.48 wt% (a, c) and MWCNT/GNP 3 wt% (2)samples (b, d). Visualization with a model of bulk virtual ("dummy") atoms and a model of asurface accessible to a solvent.The reconstructed shape of scattering particles in both samples refers to nanotubes; theshape of aggregates can be most likely ascribed to GNP. I.e., CNTs and GNP are separated in theiPP matrix. Moreover, the addition of GNP improves the dispersion of nanotubes.
4. Conclusions
The SANS method allowed us to characterize morphology of different carbon allotropesnanofillers in the volume of iPP in detail and to reconstruct their shape using numerical modelingexcluding scattering from iPP. All nanofillers are found to be polydisperse, they form self-similar aggregates with high surface fractality. In iPP volume, CNTs twist in coils and knots,become more packaged, GNP and nanographite form new flat particles. The system of scattering particles in the GNP and nanographite samples is characterizedby high polydispersity. The particle size distribution contains wafers with a thickness of 1 to 20 m, whereas the thickness of the initial plates is 47.3 nm and 1 nm for nanographite and GNPsamples, respectively.In the system of MWCNT the shape of scattering particles (aggregates) can most simplybe interpreted as the “entanglements” of neighboring CNTs, by analogy with similar“entanglements” of polymer macromolecules at concentrations higher than the crossoverconcentration. In short, MWCNT system is similar to non-woven fabric structure.The SWCNT system can be described as a grid of individual tubes and tubes "stucktogether" by the lateral surfaces and rather large knots.In case of binary filler (MWCNT+GNP) the shape of scattering particles refers tonanotubes, while the shape of aggregates can be most likely ascribed to GNP. I.e., CNTs andGNP are separated in the iPP matrix. Moreover, the addition of GNP improves the dispersion ofnanotubes.With modified geometry, nanofillers influence electron density distribution, electrical,optical, mechanical properties of functional basic polymers etc. , and this phenomenon requiresfurther careful studies for practical applications Acknowledgements
This work was supported by the Ministry of Science and Higher Education of the RussianFederation.
References [1] Polymer Composites. V. 2.: Nanocomposites. Eds. T. Sabu, J. Kuruvilla, M. Sant Kumar, etal., Wiley-VCH Verlag GmbH & Co. KGaA.,2013.[2] T. Hasan, V. Scardaci, P.H. Tan, F. Bonaccorso, A.G. Rozhin, Z. Sun and A.C. Ferrari,Nanotube and Graphene Polymer Composites for Photonics and Optoelectronics, in: O. Hayden,K. Nielsch (eds.), Molecular- and Nano-Tubes, Springer, 2011, pp. 279-354.DOI 10.1007/978-1-4419-9443-1_9.[3] V.G. Shevchenko, S.V. Polschikov, P.M. Nedorezova, A.N. Klyamkina, A.N. Shchegolikhin,A.M. Aladyshev, V.E. Muradyan,
In situ polymerized poly(propylene)/graphene nanoplateletsnanocomposites: dielectric and microwave properties, Polymer. 53 (2012) 5330–5335.[4] L.A. Feigin, D.I. Svergun, Structure Analysis by Small-Angle X-ray and Neutron Scattering,Plenum Press, New York, 1987.[5] A.N. Ozerin, T.S. Kurkin, L.A. Ozerina, V.Yu. Dolmatov, X-ray Diffraction Study of theStructure of Detonation Nanodiamonds, Crystallogr. Rep. 53 (1) (2008) 60–67.DOI:10.1134/S1063774508010070.
6] J.S. Pedersen, Analysis of small-angle scattering data from colloids and polymer solutions:modeling and least-squares fitting, Advances in Colloid and Interface Sci. 70 (1997) 171–210.[7]
Y. Dror, W. Salalha, W. Pyckhout-Hintzen, A.L. Yarin, E. Zussman, Y. Cohen, From carbonnanotube dispersion to composite nanofibers, Progr. Colloid Polym Sci. 130 (2005) 64–69. DOI10.1007/b107346.[8] E.M. Milner, N.T. Skipper, C. A. Howard, M.S.P. Shaffer, D.J. Buckley, K.A. Rahnejat, P.L.Cullen, R.K. Heenan, P. Lindner, and R. Schweins, Structure and Morphology of ChargedGraphene Platelets in Solution by Small-Angle Neutron Scattering, J. Am. Chem. Soc. 134(2012) 8302–8305. dx.doi.org/10.1021/ja211869u.[9] A. Manta, M. Gresil, C. Soutis, Predictive Model of Graphene Based PolymerNanocomposites: Electrical Performance, Appl. Compos. Mater. 24 (2017) 281–300. DOI:10.1007/s10443-016-9557-5.[10] P.S. Singh, V.K. Aswal, Probing polymer nanocomposite morphology by small angleneutron scattering, Pramana – J. Phys. 71 (2008) 947–952.[11] M.V. Petoukhov, D. Franke, A.V. Shkumatov, G. Tria, A.G. Kikhney, M. Gajda, C. Gorba,H.D.T. Mertens, P.V. Konarev, D.I. Svergun, New developments in the ATSAS program packagefor small-angle scattering data analysis, J. Appl. Cryst. 45 (2012) 342–350.doi:10.1107/S0021889812007662.[12] A.N. Tikhonov, V.Ya. Arsenin, Solution of Ill-Posed Problems, Wiley, NY, 1977.[13] J.E. Martin, A.J. Hurd, Scattering from Fractals, J. Appl. Cryst. 20 (1987) 61–78.[14] E.V. Shtykova, Shape Determination of Polydisperse and Polymorphic Nanoobjects fromSmall-Angle X-Ray Scattering Data (Computer Simulation), Nanotechnologies in Russia. 10(2015) 408–419. DOI: 10.1134/S1995078015030155.[15] A.A. Arbuzov, V.E. Muradyan, B.P. Tarasov, Synthesis of graphene like materials bygraphite oxide reduction, Russ. Chem. Bull.62 (2013)1962–1966.[16] W.S. Hummers, R.E. Offeman, Preparation of Graphitic Oxide, Journal of the AmericanChemical Society 80 (1958) 1339-1339. http://dx.doi.org/10.1021/ja01539a017.[17] V.E. Muradyan, V.S. Romanova, A.P. Moravsky, Z.N. Parnes, Yu.N. Novikov, A graphiteoxide-based nickel catalyst for reductive dechlorination of polychlorinated aromatichydrocarbons, Russian Chemical Bulletin. 49 (2000) 1017-1019. DOI: 10.1007/BF02494886.[18] S.V. Polschikov, P.M. Nedorezova, A.N. Klyamkina, A.A. Kovalchuk, A.M. Aladyshev,A.N. Shchegolikhin, V.G. Shevchenko, V.E. Muradyan, Composite materials of graphenenanoplatelets and Polypropylene, prepared by in situ polymerization, J. Applied Polymer Sci.127 (2013) 904-911.DOI: 10.1002/app.37837.[19] P.M. Nedorezova, V.G. Shevchenko, A.N. Shchegolikhin, V.I. Tsvetkova, and Yu.M. orolev, Polymerizationally filled conducting polypropylene graphite composites prepared withhighly efficient metallocene catalysts, Polym. Sci. Ser. A. 46 (2004) 242–249. [20] W. Spaleck, F. Kuber, A. Winter, J. Rohrmann, B. Bochmann, M. Antberg, V. Dolle, E.F.Paulus, The Influence of Aromatic Substituents on the Polymerization Behavior of BridgedZirconocene Catalysts, Organometallics. 13 (1994) 954–963. [20] S.V. Polschikov, P.M. Nedorezova, O.M. Palaznik, A.N. Klyamkina, D.P. Shashkin, A.Ya.Gorenberg, V.G. Krasheninnikov, V.G. Shevchenko, A.A. Arbuzov, In situ polymerization ofpropylene with carbon nanoparticles. Effect of catalytic system and graphene type, PolymerEngineering & Science. 58 (2018) 1461-1470. DOI: 10.1002/pen.24644[21] O.M. Palaznik, P.M. Nedorezovaa, S.V. Pol’shchikova, A.N. Klyamkina, V.G. Shevchenko,V.G. Krasheninnikov, T.V. Monakhova, and A.A. Arbuzov, Production by In Situ Polymerizationand Properties of Composite Materials Based on Polypropylene and Hybrid Carbon Nanofillers,Polymer Science. Series B. 61 (2) (2019) 200–214.[23] A.I. Kuklin, A.K. Islamov, V.I. Gordeliy, Scientific reviews: Two-detector system for small-angle neutron scattering instrument, Neutron News. 16(3) (2005) 16–18. [24] A.I. Kuklin, A.D. Rogov, Yu.E. Gorshkova, P.K. Utrobin, Yu.S. Kovalev, A.V. Rogachev,O.I. Ivankov, S.A. Kutuzov, D.V. Soloviov, and V.I. Gordeliy, Analysis of neutron spectra andfluxes obtained with cold and thermal moderators at IBR-2 reactor: Experimental and computer-modeling studies, Physics of Particles and Nuclei Letters. 8.2 (2011) 119–128. [25] A.G. Soloviev, T.M. Solovjeva, O.I. Ivankov, D.V. Soloviov, A.V. Rogachev, and A.I.Kuklin, SAS program for two-detector system: seamless curve from both detectors, Journal ofPhysics: Conference Series. 848 (2017) 012020-1-012020-7. doi :10.1088/1742-6596/848/1/012020.[26] M.B. Kozin and D.I. Svergun, Automated matching of high- and low-resolution structuralmodels, J. Appl. Crystallogr. 34 (2001) 33–41.[27] D.W. Schaefer, R.S. Justice, How Nano Are Nanocomposites?, Macromolecules. 40 (2007)8501–8517.orolev, Polymerizationally filled conducting polypropylene graphite composites prepared withhighly efficient metallocene catalysts, Polym. Sci. Ser. A. 46 (2004) 242–249. [20] W. Spaleck, F. Kuber, A. Winter, J. Rohrmann, B. Bochmann, M. Antberg, V. Dolle, E.F.Paulus, The Influence of Aromatic Substituents on the Polymerization Behavior of BridgedZirconocene Catalysts, Organometallics. 13 (1994) 954–963. [20] S.V. Polschikov, P.M. Nedorezova, O.M. Palaznik, A.N. Klyamkina, D.P. Shashkin, A.Ya.Gorenberg, V.G. Krasheninnikov, V.G. Shevchenko, A.A. Arbuzov, In situ polymerization ofpropylene with carbon nanoparticles. Effect of catalytic system and graphene type, PolymerEngineering & Science. 58 (2018) 1461-1470. DOI: 10.1002/pen.24644[21] O.M. Palaznik, P.M. Nedorezovaa, S.V. Pol’shchikova, A.N. Klyamkina, V.G. Shevchenko,V.G. Krasheninnikov, T.V. Monakhova, and A.A. Arbuzov, Production by In Situ Polymerizationand Properties of Composite Materials Based on Polypropylene and Hybrid Carbon Nanofillers,Polymer Science. Series B. 61 (2) (2019) 200–214.[23] A.I. Kuklin, A.K. Islamov, V.I. Gordeliy, Scientific reviews: Two-detector system for small-angle neutron scattering instrument, Neutron News. 16(3) (2005) 16–18. [24] A.I. Kuklin, A.D. Rogov, Yu.E. Gorshkova, P.K. Utrobin, Yu.S. Kovalev, A.V. Rogachev,O.I. Ivankov, S.A. Kutuzov, D.V. Soloviov, and V.I. Gordeliy, Analysis of neutron spectra andfluxes obtained with cold and thermal moderators at IBR-2 reactor: Experimental and computer-modeling studies, Physics of Particles and Nuclei Letters. 8.2 (2011) 119–128. [25] A.G. Soloviev, T.M. Solovjeva, O.I. Ivankov, D.V. Soloviov, A.V. Rogachev, and A.I.Kuklin, SAS program for two-detector system: seamless curve from both detectors, Journal ofPhysics: Conference Series. 848 (2017) 012020-1-012020-7. doi :10.1088/1742-6596/848/1/012020.[26] M.B. Kozin and D.I. Svergun, Automated matching of high- and low-resolution structuralmodels, J. Appl. Crystallogr. 34 (2001) 33–41.[27] D.W. Schaefer, R.S. Justice, How Nano Are Nanocomposites?, Macromolecules. 40 (2007)8501–8517.