Effect of Variation of Hard Segment Content and Graphene-Based Nanofiller Concentration on Morphological , Thermal , and Mechanical Properties of Polyurethane Nanocomposites

This study describes the development of a new class of high-performance polyurethane elastomer nanocomposites containing reduced graphene oxide (RGO) or graphene nanoplatelets (GNP). Two types of polyurethane elastomers with different contents of hard segments (HS) were used as a polymer matrix. The developed nanocomposites were characterized by thermal analysis (DSC, TG), dynamic mechanical testing (DMA), hardness testing, mechanical properties, rheology, FTIR spectroscopy, XRD, and microscopy investigation (TEM, SEM). Morphological investigation confirmed better compatibility of RGO with the polyurethane (PU) matrix compared to GNP. Both applied nanofillers influenced melting and crystallization of the PU matrix. The nonlinear viscoelastic behavior of the nanocomposites (Payne effect) was studied, and the results were compared with theoretical predictions.


Introduction
Many nanofillers are at present popular additives used to modify diverse properties of polymeric materials influencing their mechanical, physical, and chemical properties and durability parameters as well [1][2][3].Nanofillers are often distinguished by the shape as nanoplatelets (1D), nanofibres (2D), or nanoparticles (3D) [3].All of them should have at least one dimension down to 100 nm.Most commonly used nanofillers are silicates (includes the most popular montmorillonite), titanium dioxides, metallic nanoparticles, inorganic nanotubes, and all of the carbon nanofillers like carbon nanotubes, fullerenes, and finally graphene.When Geim and Novoselow in 2004 made a breakthrough in science with the discovery of graphene monolayers, the scientific community all over the world started focusing on these material properties and possible applications.Since then, improving methods of graphene synthesis and synthesis of different types of its derivatives were elaborated.Every single one of graphenes' derivatives, like graphene oxide (GO), chemically reduced graphene oxide (RGO), thermally reduced graphene (TRG), and chemically modified graphene (CMG), have different properties compared to pure graphene sheet.Every method of synthesis (from epitaxial methods, CVD, intercalation of graphite, exfoliation, etc.) provides a new material with specific characteristics.
Polyurethane elastomers are one of the most important polymeric materials, which can be applied in those areas, where use of classical elastomers is limited.Generally, polyurethane elastomer macromolecular chains are built up of soft (SS) and hard (HS) segments [31], and the proportion of hard to soft segments and the polarity of the segments determine especially elasticity, hardness [32,33], mechanical [33,34] or thermal [33,34] properties, and the morphology of PU.Intra-and intermolecular hydrogen bonds within polyurethane materials influence additionally the strength and thermal stability of these elastomers [35].The low miscibility between soft and hard segments leads to microphase separation of segmented polyurethanes.Increasing HS content above 30% often evidently changes long-range connectivity of HS, and this results in an increase in mechanical strength of polyurethanes.What is more, chemical components like isocyanates, polyols, and chain extenders (it is symmetric and presents reactive groups), used to obtain polyurethane materials, may introduce specific interactions between the segments, which determine the thermal and mechanical behavior [35].Properties of polyurethane elastomers can be also modified during processing by creating polymer nanocomposites [21,36,37], where the polyurethane material acts as a matrix for dispersed nanoparticles.
The present paper is aimed at reporting the influence of two carbon nanofillers (reduced graphene oxide (RGO) and graphene nanoplatelets (GNP)) on two polyurethane elastomeric matrices (possessing different hard segment (HS) contents).This also focuses on nanocomposite systems while gaining changes in some properties, such as improved mechanical and thermal properties.The morphology and viscoelastic behavior with changing stiffness of the matrix was also studied.Nanofillers used in this work were thermally reduced graphene (RGO) and graphene nanoplatelets (GNP).This is a new approach to obtaining nanomaterials based on the wide use of polyurethane raw materials, which should be suitable for a commercial application.The main goal was to obtain high-performance nanocomposites by simple methods of synthesis, such as in situ polymerization method, without solvents.Also, some of the synthesized nanocomposites contained nanofiller RGO derived from graphite.Compared to GNP, provided by the commercial manufacturer, but still expensive [38], usage of low-cost laboratory-made RGO is an economical approach to estimating the way towards mass scale production of graphene derivative-based nanocomposites.
RGO is an economical approach to estimating the way towards mass scale production of graphene derivativebased nanocomposites.Incorporation of these types of nanofillers can extend polyurethane thermoplastic elastomers' property limits, improving the mechanical and tribological performance of common TPU applications like footwear, closure o-rings and seals, adhesives, automotive interiors, and sport and leisure items.There here is a high-potential, still needed investigation for using GNP or RGO containing polyurethane nanocomposites in these technical goods, where electrostatic dissipation ability is of special importance, like conveyer belts and technical fabrics [38][39][40].
It is the first performed detailed study of the influence of the used nanofillers on hard and soft segment phase formation and property changes in polyurethane elastomer nanocomposites.

Methods and Materials
2.1.Differential Scanning Calorimetry (DSC).Differential scanning calorimetry measurements have been carried out by a heat-flux Netzsch DSC 209F1 Proteus analyzer.The calibration process was performed for 6 points (C 10 H 16 , In, Sn, Bi, Zn, CsCl).The investigation temperature range was from −85 °C to 250 °C.5-10 mg samples were used in alumina pans.Measurement was performed in the N 2 atmosphere.During the investigation, a heating/cooling rate equal to 10 °C/min was used.A second heating scan was used for the all calculations.

Thermogravimetric Analysis (TG).
Thermogravimetry of obtained materials was conducted by a Netzsch TG 209F3 Tarsus analyzer.The calibration process was performed for 6 points (In, Sn, Bi, Zn, Al, and Ag).Investigations were performed using the Al 2 O 3 crucible.Mass loss was followed at a heating rate of 20 °C/min for samples (5-10 mg) in the temperature range from 37 °C to 600 °C, under N 2 atmosphere.

Dynamic Mechanical Analysis (DMA)
. Dynamical mechanical analysis has been carried out using a TA Instruments DMA Q800 analyzer.Investigation has been performed using tension mode, in the temperature range − 100-100 °C, frequency equal to 1 Hz, and heating rate at 4 °C/min.Samples were cooled using liquid nitrogen.Strain sweep analysis, for nonlinear viscoelastic investigation, has been conducted at temperature 35 °C, in the same tension mode and frequency of deformation but in the amplitude range from 0.5 to 6500 μm.

Fourier Transformation Infrared Spectroscopy (FTIR).
Fourier transformation infrared spectroscopy of polyurethane materials has been carried out using a Thermo Electron Corporation Nicolet 8700 Spectrometer (equipped with ATR mode).Samples were scanned from 4000 to 500 cm −1 .

Transmission Electron Microscopy (TEM). Microscopy analysis was performed using STEM-EDX Transmission
Electron Microscope FEI Europe, Tecnai F20 X-TWIN coupled with EDX spectrometer.All samples were cut in cryo-mode using an ultramicrotome (Leica EM UC7 Ultramicrotome).HRTEM (for nanocomposite systems) was from JEOL, model JEM-2100.The thickness of the samples The method of the previous publication, Strankowski et al. [42], was followed.

Preparation
3.1.Reduced Graphene Oxide (RGO) Preparation.Two-step synthesis was used to obtain reduced graphene oxide (RGO).Based on modified Hummer's method, the GO nanofiller was prepared [43].GO was thermally reduced to RGO at 200 °C, using the Ar atmosphere.3 International Journal of Polymer Science method with different hard segment (HS) contents of 20% (PU20) and 40% (PU40) by weight and at an isocyanate index equal to 1.05.The prepolymer was obtained at the first step by reacting PTMG polyol and MDI isocyanate at 70 °C during 1 h under low pressure.The obtained prepolymer was used to prepare unmodified polyurethanes and nanocomposites.For unmodified systems at the second step, 1.2-butanediol (BD) as a chain extender and 1,4-diazabicyclo[2.2.2]octane (DABCO) as a catalyst were used.For preparing nanocomposite systems, a calculated amount of the nanofiller was mechanically (2500 rpm) dispersed in the prepolymer with ultrasound support for 30 min.At the last stage, to complete the reaction, BD and DABCO were added to the prepolymer containing the dispersed nanofiller.The reacting mixtures, when still fluid, were poured into the steel mold and annealed at 100 °C for 24 h in a laboratory drier.After that, materials were compressed (2 tons) using a hydraulic press (Manual Hydraulic Press, Specac) to obtain thin 0.5 mm films, which were used for the investigations.

Unmodified Polyurethane Matrices (PU).
Based on spectroscopy investigation, it was confirmed that general absorption bands are typical for the polyurethane matrices (Figure 3) [44,45].This study was performed to check that all reactions were completed and there were no visible bands for unreacted NCO groups from isocyanate substrates.What is more, X-ray diffraction analysis shows absence of diffraction maxima characteristic for crystalline regions (Figure 4) [46].This study reveals that there are no long-range ordered structures in these polyurethane matrices.In Figure 4, a wide amorphous diffusion peak (at about 2θ = 20 °) can be observed which is typical for amorphous phase in amorphous and partially crystalline polymers [47][48][49].
For the polyurethane with higher hard segment content (PU40), higher temperature and enthalpy of melting point (T m ) were observed, in comparison to the softer matrix (PU20) (Table 1, Figure 5).The crystallization temperature for PU40 shifts to higher values (about 10 °C) comparatively to the PU20 matrix, and enthalpy of crystallization has    4 International Journal of Polymer Science higher values for the harder matrix (PU40) (Figure 5).These results are in good agreement with the DMA study (Figure 6), where a higher storage modulus (E ′ /MPa) was observed for the PU40 matrix in the applied temperature range for this investigation.

Polyurethane Nanocomposites
4.1.1.FTIR Spectroscopy.Infrared spectroscopy was performed for all nanocomposite materials containing RGO or GNP nanofiller.By comparing the FTIR spectra, it was confirmed that the addition of the nanofiller did not influence the chemical structure of the polyurethane matrix (Figure 7).However, there was a possible reaction of isocyanate groups with functional groups located on the surface of reduced graphene oxide.Nevertheless, the number of these reactive groups is negligible (max.2%) and their influence of the nanocomposite creation is minor.
Based on the FTIR study where the absorption and diffraction of infrared radiation have low influence on waveforms (at such low content of the nanofiller) and, what is more, that the absorption coefficient of C=O and N-H bands   5 International Journal of Polymer Science did not change, the phase separation behavior should be connected only with disorders of the domain structure caused by nanofiller addition.Phase separation was calculated according to [50][51][52] where A 1 and A 2 are surface areas of oscillating bands of the C=O group, hydrogen-bonded and nonhydrogenbounded)], and the influence of nanofillers on this parameter is presented in Figure 8.It was confirmed that the phase separation degree (PSD) increased for PU40 nanocomposites and had lower values and decreased for PU20 materials (Figures 8(a) and 8(b)).

X-Ray Diffraction (XRD).
The obtained diffractograms (Figure 9) for nanocomposites PU20-GNP and PU40-GNP show wide halo patterns originating from amorphous polyurethane materials.Moreover, for nanocomposites containing a GNP nanofiller, a diffraction maximum at about 2θ = 27 °was visible, originating from crystalline layers of GNP.The intensity of these maxima increases with increasing content of the GNP nanofiller in the polyurethane matrix.For polyurethane nanocomposites, no diffraction maxima were observed, which confirmed (with TEM investigation support) that all the systems possess an exfoliated structure with a well-dispersed RGO nanofiller.

Transmission Electron Microscopy (TEM).
Based on transmission electron microscopy (TEM) analysis, good dispersion of the nanofiller (RGO or GNP) into polyurethane matrices with different hard segment (HS) contents was confirmed.This is a very important factor characterizing the morphology of the nanofiller (size, dispersion) in nanocomposite systems.Reduced graphene oxide (RGO) and graphene nanoplatelets (GNP) (Figure 10) are visible in the form of characteristic long, disordered dark lines into clear, gray, PU matrices.For PU/GNP systems (Figures 10(a The RGO nanofiller has better compatibility to the polyurethane matrix (what relates to the -Oor -OH group on the surface of the nanofiller), and therefore higher dispersion is achieved.
These results based on TEM analysis are in good correspondence to the XRD study where (no diffraction maxima; described in X-ray diffraction (XRD) paragraph) an exfoliated nanocomposite morphology was confirmed, especially for PU/RGO materials.

Differential Scanning Calorimetry (DSC).
Crystalline and amorphous domain structure changes in the obtained polyurethane nanocomposites were investigated using DSC.Figures 11 and 12 show DSC and cooling and heating scans.Based on cooling thermograms, it is visible that nanofiller addition (RGO and GNP) influences on polyurethane hard and soft domains.The temperatures of crystallization and enthalpies of this process change when the nanofiller is incorporated into the polyurethane matrix.For PU20-GNP and PU40-RGO, the nanofiller affects the rate of crystallization, where maximum temperature rates are lowest and enthalpy of crystallization takes the lowest values.This behavior is connected with a smaller amount of crystalline phase creation during cooling.However, samples PU20 (0.25 wt% GNP) and PU40 (2 wt% RGO) exhibit a shift in crystallization temperature of HS to higher temperatures, in comparison to the other modified materials.This behavior can be 6 International Journal of Polymer Science explained with possible presence of some impurities from the mold, which can infect the crystallization process of these materials.What is more, a reverse behavior exhibits PU20-RGO and PU40-GNP, where the crystallization temperatures are higher, which is related with the heterogenic nucleation process occurring at nanofiller surfaces.Additionally, for PU40-GNP nanocomposites, the observed crystallization peaks (shift to the higher temperatures) split with visible two maxima (Figure 11(c)), which is connected with creating crystalline areas of different size.Addition of the nanofiller also increases the crystallization rate of soft segments (SS), which is visible for all the PU40-GNP nanocomposites (Figure 11(c)), but not revealed for RGO-based nanocomposites, because this process could be extended in time after the controlled cooling step.After the cooling process, all samples were heated at a controlled speed (second heating step, after clearing the thermal history of the samples) to verify the phase transition-melting behavior.For PU20-GNP, PU20-RGO, and PU40-RGO, the addition of the nanofiller causes a decrease in enthalpy of melting, which is connected with the lower amount of crystalline phase in the hard segment domains.Not very clear melting peaks for PU20 nanocomposites (at temperature range 125-210 °C) were blurred in comparison to the nonmodified PU matrix.What is more, for PU40-RGO, there were revealed multiple peaks on the thermograms.The first maximum was visible at a temperature of about 90 °C and the other two peaks at 190 °C and 210 °C (Figure 12(b)).The first maximum occurs before the melting point, and the next two maxima are connected with melting and recrystallization of the crystalline phase.The addition of the reduced graphene oxide (RGO) nanofiller, into the polyurethane matrix, causes an increase of the peak at 90 °C, blurry peak at temperature 190 °C, and extinction   2).However, in the thermograms (Figure 12(b)), some configuration changes of melting peaks have been observed.At 190 °C, the melting peak is lower and disappears with the addition of the 2 wt% GNP nanofiller.What is more, the second peak at 210 °C is lower (up to 1 wt% GNP) and increases with the addition of 2 wt% in the nanofiller (GNP).Simultaneously for all nanocomposite systems, a third, well-defined, peak at 225 °C is observed.For all heating runs, it is visible that the melting temperature of soft segment (SS) domains can be estimated at about 13 °C (Figure 12(b)).Also, there was an observed decrease in the crystalline phase for nanocomposites PU20-GNP (above 0.5 wt% of the GNP nanofiller), PU40-RGO (for amount 0.25-1 wt% RGO), and PU20-RGO (for mount 0.5-1 wt% RGO), which is connected with lower values of melting enthalpy (Table 2).However, for PU40-GNP nanocomposites, this parameter (amount of crystalline phase within soft domains) remains at a similar level (Table 2).

Thermogravimetric Analysis (TG).
Based on the performed TGA analysis, there was a confirmed slight improvement in the thermal stability of the nanocomposite system by the addition of the nanofiller (RGO or GNP).The onset degradation temperature of obtained materials is comparable for unmodified and nanocomposite systems (Figure 13, Table 3).For the modified system, this parameter shifts to higher temperatures at about T 5% = 5 -7 °C in comparison to the polyurethanes without the nanofiller.Visible improvement in thermal stability (11 °C) is observed for the PU40-RGO system with the 2 wt% nanofiller (Figure 13(f)).Maximum degradation temperatures (T DTGmax ) of hard segments (HS) for PU20-RGO (Figure 13(d)) and PU40-RGO (Figure 13(h)) are shifted to higher values by 27 °C and 14 °C, respectively (Table 3).For all nanocomposite systems, the degradation temperature of soft segments possesses similar values.For filled systems, there is a visible higher mass content [5% for PU20-2.0%GNP/RGO (Figures 13(a) and 13(b)) and 9% for PU40-2% GNP (Figure 13(e))] after degradation (mass residue), which is connected with higher thermal stability of nanofillers in comparison to the pure polyurethane matrices (2% for PU20 and 5% for PU40) (Figure 13, Table 3).Addition of the nanofiller does not cause significant changes in DMA characteristic parameters: storage modulus (E′), loss modulus (E″), and loss tangent (δ) in the glassy state temperature rage (Figures 14 and 15).Visible changes were observed above glass transition temperatures, where soft segments are at viscoelastic state and their movement is possible.It was confirmed that for polyurethane nanocomposites slightly lower values of the loss tangent were registered in comparison to the nonmodified PUs (Figures 14(e), 14(f) and 15(e), 15(f)).The glass transition temperature (calculated from the maximum of (tanδ)) changed with the addition of the nanofiller with a decrease of 6 °C for PU20-2% RGO and an increase of about 4 °C for PU20-2% GNP.For stiffer nanocomposite matrices with RGO and GNP nanofillers, a decrease in glass transition temperature in the range (1-9 °C) was evident (Figures 14 and 15, Table 4).For PU40-0.5% GNP glass transition, the temperature shifts to the temperature of −51 °C, with a value lower by 9 °C compared to the nonmodified PU40 system (−42 °C) 9 International Journal of Polymer Science (Figure 14, Table 4).A higher decrease in glass transition temperature was observed for graphene nanoplatelet-(GNP-) based nanocomposites parallel to nanocomposites with reduced graphene oxide (RGO).Table 2: Enthalpy of melting for PU nanocomposites (heating scans).
Heating scans % nanofiller PU20-GNP PU20-RGO PU40-GNP PU40-RGO ΔH m,HD (J/g) ΔH m,SD (J/g) ΔH m,HD (J/g) ΔH m,SD (J/g) ΔH m,HD (J/g) ΔH m,SD (J/g) ΔH m,HD (J/g) ΔH m,SD (J/g) 0.00 5.5 13.5 5.  11 International Journal of Polymer Science soft segment phase.What is more, at this region the mobility of SS is limited for both amorphous phase and crystalline phase of SS.However, with the temperature increase, the SS crystalline phase melts going completely into the amorphous phase at 20 °C.
Addition of the nanofiller causes visible strain changes in comparison to stress behavior, defined as tanδ.For PU20-RGO nanocomposites (Figure 14(f)), values of tanδ are lower with increasing RGO nanofiller content, which is connected with a more elastic behavior of these materials at the investigated temperature range.This is similar for PU20-GNP series (Figure 14(e)), but with the difference that for a 2 wt% addition of GNP there is a rapid increase in tanδ values, which is higher in comparison to the other nanocomposite series and the nonmodified PU matrix.This distinct change probably is related with higher aggregation of the GNP nanofiller, and the increase in loss tangent parameter is the result of the sliding of nanoplatelets, resulting in a decrease in the elasticity of the system, which causes an increase in tanδ values.What is more, the decrease of this parameter for other systems (besides described above) probably is connected with a hydrodynamic effect, where nanofiller particles increase the viscosity of the nanocomposites causing the stiffening of the system.
In the case of PU20-RGO and PU20-GNP nanocomposites, there are noticeable storage modulus (E′) changes (Figures 14(a) and 14(b)) in comparison to the PU20 matrix.These E′ values are lower up to 1 wt% of the nanofiller content and increase for 2 wt% (RGO of GNP amount), reaching a similar level as for the PU20 matrix.Lower values of the storage modulus (E′) (despite increased stiffness) are connected with the weaker adhesion of the nanofiller to the soft segment phase, with movement that is easier on the nanofiller surface.
For polyurethane nanocomposites (PU40-GNP and PU40-RGO), there was a visible (at all investigation temperature range) increase in storage modulus (E ′ ) values with the addition of the nanofiller (Figures 15(a) and 15(b)).On the other hand, tanδ changes are different, and for both series this parameter decreases.At up to 0 °C for PU40-GNP nanocomposites, a high decrease in tanδ values is observed in comparison to the results for PU40-RGO.This behavior probably means that PU40-GNP possessed higher elasticity.These differences in tanδ values for the two types of the nanocomposite systems evidence a more effective interaction between bordering graphene nanoplatelets (GNP) than in the case of reduced graphene oxide (RGO).It can be expected that the GNP nanofiller creates a more effective network composed of nanofiller particles which are in contact, resulting in a drop of tanδ.
However, it should be highlighted that the stiffness of the nanocomposites, within the discussed temperature range, influences the crystalline phase of the soft domain, which contributes to the stiffening of the pure PU matrix, acting as the physical cross-link of the network.The addition of nanofillers can cause changes in the crystalline structure of these areas, which affect the thermomechanical properties.The fragmentation of the crystalline structure of the soft domain (SS) and the lowering of the crystallinity of these systems may cause an additional decrease in the stiffness of the nanocomposites.At this elastic area, all crystallite structures represented by soft segments (SS) are in the amorphous state, above melting temperature.Soft segments have high a degree of free movement, which is limited only by the physical cross-linking network of the hard domain (HS).For all obtained nanocomposite systems, there was a confirmed increase in storage modulus (E ′ ) at the elastic plateau region (normal use of PU elastomers) with increasing concentration of the nanofiller.PU20-RGO nanocomposites exhibit a drop at tanδ for the nanofiller content up to 1 wt%, which is connected with a more stiffer behavior of these systems.Further addition of the nanofiller up to 2 wt% results in higher values of tanδ (Figure 14(f)) with a simultaneous increase in loss modulus (E″) (Figure 14(d)), which is the result of the sliding motion of the RGO resulting in the increase of the viscoelasticity.The decrease in tanδ values for the PU20-GNP series was also observed (Figure 14(e)).Likewise, in this case for 2 wt% GNP contents in the polyurethane matrix, there was a noticeable E″ increase, related with a sliding motion (Figure 14(c)).For PU40-RGO and PU40-GNP nanocomposites, the presence of the nanofiller causes an increase in storage modulus (E′) values (Figures 15(a) and 15(b)) contributing to the mechanical strength of the nanocomposite.Some differences were also visible for tanδ curves (Figures 15(e) and 15(f)).For PU40-GNP nanocomposites, the shape of the tanδ curve for temperatures above 50 °C is similar to the shape of the curve obtained for nonmodified PU (Figure 15(e)).These values also increase with a simultaneous increase in loss modulus (E ″ ), which is correlated with a sliding motion of the GNP nanofiller (in contact with each other) and it causes an increase in E ″ (Figure 15(c)).PU40-RGO nanocomposites exhibit an interesting behavior of the tanδ curves, where a maximum of loss tangent appears, connected with a relaxation transformation of melting behavior of the hard domains, which in this case melt at the lower temperature.This is correlated with a lower size of the crystallites, causing a shift of the melting temperature towards lower values.What is more, in comparison to the PU20 nanocomposites, for stiffer matrix (PU40), loss modulus (E″) values increase independently, regardless the amount of the nanofiller (Figure 15(c)).
(2) DMA (Nonlinear Viscoelastic Investigation).A nonlinear viscoelastic behavior of nanocomposite materials was investigated using dynamic mechanical analysis (DMA).All measured series possess three characteristic nonlinear regions (low-amplitude, transition, and high-amplitude).These regions were revealed for all pure polyurethane matrices PU20 and PU40 (Figures 16 and 17).Reinforcement of storage modulus (ΔE/MPa) was calculated as a difference between modulus values at E 0 1% ′ and E 10% ′ .All curves (storage modulus versus strain for PU20 matrix) are presented in Figures 16(a For polyurethane nanocomposite PU20 containing a minimum 2 wt% GNP nanofiller, a reinforcement effect is observed (higher values of E ′ ).In turn, for RGO nanocomposites, all E ′ values are similar in comparison to the pure polyurethane matrix (Figures 16(a) and 16(b)).In Figure 17, the difference between modulus change (ΔE) and the increasing amount of nanofiller content into the polyurethane matrix is presented.
A higher effect of the nanofiller content was observed for a harder polyurethane matrix (PU40) in comparison to the softer one (PU20) (Figures 18(a) and 18(b)).Both GNP and RGO nanofillers influence the PU40 matrix which is visible as higher storage modulus (E ′ ) values.For PU40 nanocomposites, storage modulus changes (ΔE) are most visible even at low contents of both nanofillers (Figure 19).A sudden drop of E ′ modulus at the transitional area was not observed, which is characteristic during breakdown of the nanofiller network.What is more, GNP shows a higher interaction with a harder polyurethane matrix (PU40), and a nonlinear viscoelastic behavior is visible for this type of nanofiller.

Discussion
The performed investigations of polyurethane nanocomposites, with different hard segment (HS) contents and modified with two types of graphene derivatives (RGO and GNP), revealed the explicit influence of applied nanofillers on morphological, thermal, and mechanical parameters on these systems.
The nanofillers (GNP and RGO) possess a nonpolar character; however, the PU matrix incorporates polar hard segments and the adhesion between these nanofillers and the polyurethane matrix is low.This restricted compatibility of both components (matrix and nanofiller) could be responsible for bigger particle aggregation or agglomeration (100-300 nm length) within the polyurethane matrix, especially for the GNP nanofiller (Figures 10(a) and 10(d)).FTIR investigation did not deliver any clues for the effective interaction between nanofiller (GNP) and matrix (PU) since not even a slight shifting in the position of characteristic bands was observed.There was some agglomeration of the nanofiller with its increasing amount.Based on XRD analysis, some diffraction maxima (representative for GNP nanoplatelets) are visible (Figures 9(a) and 9(c); 2θ = 27 °), which are connected with multilayer structures, especially for higher content of the GNP (PU40-GNP and PU20-GNP samples).On the other hand, for nanocomposites PU40-RGO and PU20-RGO, there were not visible diffraction maxima (Figures 9(b) and 9(d)), which shows good dispersion and 15 International Journal of Polymer Science better compatibility of this type of nanofiller.This result was confirmed using TEM analysis (Figures 10(b), 10(c), 10(e), and 10(f)), where good dispersion of the nanofiller in the polyurethane matrix was achieved.For all obtained nanocomposites (especially for GNP-based systems), it was observed that some part of the nanofiller is highly dispersed into the PU matrix and with higher concentrations of the nanofiller, more agglomerated structures were formed.
Dynamic mechanical analysis allows verifying changes of the basic viscoelastic parameters as a function of temperature for polyurethane nanocomposites with different amounts and different kinds of nanofiller (GNP, RGO).DMA analysis shows that GNP and RGO do not have the same compatibility with the polyurethane matrix.These nanofillers affect both areas of soft and hard segments, where dimensions are similar to the nanofiller surface, but in the glassy state segmental movements of the PU chains disappear and the addition of the nanofiller does not cause significant incensement of storage modulus (E′).
Above the glass transition temperature (T g ), a relaxation of soft segments allowing free movement of the nanofiller within the liquid elastic phase was observed.Depending on the kind and amount of the nanofiller, a different effect on the polyurethane matrix was observed.At low nanofiller content (0.5-1.0 wt%), there was a visible increase in the stiffness, connected with the hydrodynamic effect, where stiff nanofiller particles cause higher viscosity of the "liquid" soft domain and that results in higher E ′ values (Figures 14(a), 14(b) and 15(a) and 15(b)).For higher content of the nanofiller (1.0-2.0 wt%), a further increase in storage modulus (E ′ ) and also higher values of the loss modulus (E ″ ) were registered (Figures 14(c), 14(d) and 15(c) and 15(d)).This behavior is the result of a sliding motion of the nanoplatelets (at high nanofiller content) which can be submitted to aggregation and agglomeration.These nanoparticles can clash with one another, causing an increase in mechanical friction, which results in energy dissipation as a result of mechanical loss (increase of loss modulus, E ″ ).
The Payne effect is not visible for all obtained nanocomposite systems.The big role of creating nanofiller network improvement plays as a type of the nanofiller and hard segment content of the polyurethane matrix.For PU20-RGO nanocomposites, the addition of the nanofiller did not change visibly the storage modulus (E′), which is probably connected with the absence of aggregation enlargement after exceeding the threshold of percolation, which allows for effective interactions between the platelets.Comparing PU20-RGO and PU20-GNP samples with 2 wt% nanofiller addition, it was observed that the percolation threshold for 16 International Journal of Polymer Science RGO is higher than for the GNP nanofiller into the PU20 matrix.In PU40 samples, an increase in hard segment (HS) content causes a lower percolation threshold for both nanofillers, where an increase in E ′ is visible.Addition of these nanofiller leads to an instant increase in E ′ .The graphene nanoplatelet nanofiller was a more effective modificator, for the PU40 matrix, resulting in a higher E ′ than the RGO additive.The hard domain behavior and crystallization process of polyurethane strongly depend on the type of introduced nanofiller and on hard segment (HS) content.Addition of the nanofillers impacts on the crystallization rate, degree of crystallization, and size of the crystallites formed.Table 5 shows the effect of the nanofiller additive on the following: crystallization rate-based on cooling thermograms, resulting from changes in the temperature at which they start to crystallize; effect on crystallinity-decrease in crystallinity was found by the lower values of the melting enthalpy; and influence on the size of forming crystallites-presented on the basis of shifting of melting temperatures (T m ).
Based on observed investigations for nanocomposite systems (PU20-RGO, PU20-GNP, and PU40-RGO), a decrease in crystallinity within the hard domain (Table 5) was conformed.This points to disruption of the hard domain crystallization process caused by the nanofiller, which acts as a physical barrier, reducing the tendency to crystallization.For PU40-GNP polyurethane nanocomposites, a significant influence on the crystallization degree within hard and soft domains was not conformed, but the rate of crystallization in both cases increased.For these samples, creation of hard domain crystallites, which can have larger sizes, was confirmed, which is connected with a shift to higher temperatures of melting temperature peaks with simultaneous disappearance of the preheating area.From the described changes in the crystalline structure of hard domain PU40 nanocomposites, it can be presumed that the crystallization process within these domains is different.Presented differences within HS can be explained by changes in the crystallization process, which depends on the type of nanofiller and also hard segment content in the polyurethane matrices.
For nanocomposites described above, a hard domain melting point is influenced by the fact that all segments (soft and hard) are mixed, while the nanofiller is dispersed between them.At cooling from the melt, the occurring phase separation becomes better with increasing concentration of hard segments.During crystallization of HS, nanoplatelets probably migrate into the soft domain.This behavior can be confirmed by no changes in HS crystallization parameters for PU40-GNP nanocomposites compared to the PU-40  Lower crystallization enthalpy values for PU40-RGO can be connected with a higher content of RGO, which remains in HS domains causing a decrease in crystallinity and higher crystallite fragmentation.RGO's less tendency to migrate to a soft segment domain (SS) can be due to its folding, smaller flake thickness, and higher polydispersity of its surface area.Also, polar groups on the RGO surface increase the affinity to the rigid segments of PU.
The nanoplatelets migrating into the hard segment domains can accumulate within these structures, and their impact can be enhanced by the hydrodynamic effect and by the creation of the Payne effect but only at a fitting amount of this nanofiller.In the case of nanocomposites PU20-GNP and PU20-RGO, a decrease in crystallinity was observed, indicating that nanoparticles are dispersed in HS domains.As explained above, the migration of RGO is more difficult due to its defected surface structure.The migration of nanofiller particles into hard segment domains for PU20-RGO and PU20-GNP can be less intense or may not occur, because of the small size of the crystalline domains forming from hard segments, which could be connected with a lower amount of HS in these matrices.
Based on the previous investigation, the influence of the nanofillers on phase separation was shown.It can be assumed that the used nanofillers do not have significant influence on the degree of phase separation, with parameter changes fluctuating around 3%.The increasing content of RGO and GNP in the PU20 matrix results in a decrease in the degree of phase separation.What is more, the addition of these nanofillers to the PU40 matrix results in the opposite effect and the increase in the degree of phase separation.The character of these changes correlates with the influence of the nanofillers on the crystallization behavior of HS.For PU20 nanocomposites, a decrease in crystallinity was observed, whereas for PU40 an increasing effect was found only for the GNP nanofiller, not for PU40-RGO samples.With a lower HS content in the PU20 matrix, nanoparticle segregation is smaller, causing more nanoparticles to be present in the hard segment domain thus reducing hydrogen bonding within these crystalline structures.A reverse effect is found for PU40 nanocomposites, where phase separation is higher causing more well-structured hard domains.However, this situation does not coincide for nanocomposites PU20-RGO and PU40-RGO, where phase separation is higher than for PU40-GNP and PU40-RGO.This could be the reason for the heterogeneity of the nanocomposites confirmed by FTIR investigation and of some agglomeration of the nanofiller on the surface of the samples.

Conclusions
A clear influence of the nanofillers on the structure, morphology, and thermal and mechanical behavior of the polyurethane nanocomposites was found for two polyurethane matrices possessing different hard segment contents (20% and 40% HS) modified with graphene derivatives (reduced graphene oxide (RGO) and graphene nanoplatelets (GNP)) (Table 6).Differences in the properties of the obtained nanocomposites result mainly from the change in the morphological structure of these nanocomposites as a result of the introduction of nanofillers.Obtaining nanocomposites with a different content of the nanofiller allowed following the changes in the structure and properties of nanocomposites.The addition of both types of graphene forms influenced the crystalline structure of hard and soft domains.The use of the polyurethane matrix with a different content of rigid segments revealed a different way in which crystallization and melting processes occur within both domains present.It has been found, depending on the filled matrix, that the graphene form may cause a decrease in the degree of Cooling scans % nanofiller PU20-GNP PU20-RGO PU40-GNP PU40-RGO ΔH c,HD (J/g) ΔH c,SD (J/g) ΔH c,HD (J/g) ΔH c,SD (J/g) ΔH c,HD (J/g) ΔH c,SD (J/g) ΔH c,HD (J/g) ΔH c,SD (J/g) 0.00 −7.Based on changes in crystallinity, the crystallization process occurring in the presence of nanofillers in polyurethane matrices was examined.A properly selected type of graphene nanofiller and its appropriate amount allows improving the elastic properties of polyurethane materials, as a result of hydrodynamic reinforcement connected with the creation of Payne's reinforcing aggregation network.

Figure 3 :
Figure 3: FTIR spectra for nonmodified PU with different hard segment ratios.

Figure 6 :
Figure 6: Storage modulus (E′/MPa-black line) and loss tangent (black dot line) versus temperature/ °C for nonmodified PU matrices with different hard segment contents.
) and 10(d)), long lines (about 500 nm) are observed, which are aggregated into PU matrices (PU20 or PU40).For PU/GNP systems (Figures10(a) and 10(d)), larger agglomerates were observed in comparison to the RGO-filled systems.The incorporation of RGO into the PU matrix causes good dispersion, and a characteristic disordered structure (short, thin, black lines) was obtained, as clearly observed in Figures 10(b), 10(c), 10(e), and 10(f).The RGO nanofiller is affected by the hard and soft segment ratio in the PU matrix, and better exfoliation is archived for the PU matrix with lower content of hard segments (Figures 10(b) and 10(c)).

Table 3 :
TG data for PU nanocomposites.

Table 4 :
Glass transition temperatures for PU and nanocomposite materials.

Table 6 :
18mparison of the nanofiller's influence on PU hard and soft domains' behavior.Designation: (↑) increase parameter, (↓) decrease parameter, (<−>) parameter remains unchanged.18InternationalJournal of Polymer Science crystallinity and delay the crystallization process or may exhibit a nucleating behavior.