Fundamental studies of isotactic polypropylene composites with brown coals: Structure–properties relationship

ABSTRACT Wood–polymer composites (WPCs) are commonly used in materials technology, and there is a reasonable amount of research for new kinds of green fillers. In this study, we have obtained the isotactic polypropylene (iPP) composites filled with native and modified xylite (brown coal fraction with a relatively high amount of cellulose). The modification of xylite has been based on processing in a mine, where the humic acids are rinsed with KOH solution for further use. The composites were subjected to a wide range of structural, thermal, functional, and mechanical characteristics. The results of the undertaken research have shown that the addition of xylite filler can be successfully used in obtaining of thermoplastic composites. The properties of those composites are similar to another types of WPCs. There also observed a strong impact of xylite filler addition into formation of β-iPP form.


Introduction
During the crystallization process, the polymers adopt different supermolecular structures. The nature and properties of these forms are influenced by thermal conditions, mechanical effects (e.g., external stress), and the presence of additional substances when we are observing the heterogeneous nucleation (Zhu et al. 2009;Jungblut and Dellago 2013;Jianzhu et al. 2019). In general, there are three crystal forms of isotactic polypropylene (iPP): monoclinic α-form, pseudohexagonal (triclinic) β-form, and triclinic γ-form (Huang, Xin-Gui, and Fang 1995;Lotz 2014;Przemyslaw et al. 2018;Varga 1983Varga , 1992. As noted during the injection molding process of polypropylene, shearing appears, which influences the formation of the beta phase in the injection mold. It is a shear-induced crystallization . A relatively low cost of iPP as well as its satisfying properties make polypropylene a good matrix for wood-polymer composites (WPCs) (Chattopadhyay et al. 2011;Chang-Mou, Lai, Wang 2016). They have been increasingly used in recent years (Peças et al. 2018;Pracella et al. 2006). Moreover, the cellulose-based fillers may occur as active or inactive nucleating agents for semicrystalline polymers (Nahar et al. 2012;Slouf et al. 2018). The subject of research, which is the identification of various nucleation mechanisms in crystallizing materials, and in particular, in polymers, has been discussed in detail by numerous research teams (Avérous and Le Digabel 2006;Yang et al. 2006).
To this day, brown coal has been usually used as a fossil fuel. However, it is a material with a relatively high number of impurities (being emitted during combustion and significantly unfavorable for the natural environment) and low calorific content (20-25 MJ/kg) and a high moisture content (up to 60 wt.%) (Fedyaeva et al. 2012). For this reason, our research team focused on the application of brown coal in composites technology, and the results of the preliminary tests are promising. To the best of our knowledge, no results of the application of brown coal are present in the literature.
In this work, the authors have analyzed the effect of the addition of brown coal (in the form of xylite -coal with a high amount of cellulose) on the iPP structure and mechanical properties of thermoplastic composites. The authors have analyzed the correlation between the type of xylite used, the structure of iPP in composites, and materials properties.

Materials
Polypropylene TATREN HG1007 supplied by Slovnaft (Slovakia) was used as a matrix in the experiment. It was characterized as follows: melting temperature 166°C and melt flow rate (MFR) 10 g/10 min (230°C/2,16 kg).
The brown coal in the form of xylite was supplied by the Polish Sieniawa Forestry -Kaminska Mine. The material had particle diameters from 1.0 to 1.5 mm. The material was dried in a convection dryer at 100°C during 24 h. The crushed xylite fraction was washed with 25%wt. aqueous KOH solution and then washed with distilled water until pH achieved a neutral value. The alkalization process was carried out in order to eluting the water-soluble humic acids, which were obtained for further applications. The obtained waste fractions after washing the water-soluble humic acids make it easy to develop an innovative method of their application. Native xylite was obtained directly from the deposits, and one modified with the addition of KOH was used as fillers.

Obtaining of composite materials
Polymer-filler mixtures obtained 75:25 wt. subjected to in-line extrusion with a Fairex single-screw extruder (McNeil Akron Repiquet L/D = 24, D = 25 mm) and a granulator. The temperatures in appropriate zones were (1) 140°C, (2) 180°C, (3) 195°C, and (4) head 200°C. The obtained composite granules were dried in a convection oven at 80°C for 12 h. The dried granulate was placed in the hopper of an ENGEL 80/20 HLS injection molding machine and subjected to the injection process. Table 1 shows the main parameters of the injection molding process.
In the injection process, the dumbbell samples were made following the PN-68/C-89034 standard ( Figure 1 a and b). As can be seen in Figure 1b, for sections of the dumbbell marked as 1, 2, and 3, there were prepared samples for the analysis. In addition, all analyses have been made into layers skin and core.

Structural investigations
The wide-angle X-ray scattering (WAXS) investigations were performed by means of a TUR M-62 (Germany) horizontal diffractometer (CuKα radiation, at 30 kV and 20 mA anode excitation; with nickel filtering). The diffraction curves were recorded for the angles in the range of 2Θ: 10-30° with a counting step of 0.04°.
The numeric analysis of diffraction curves was performed with the use of Hindeleh and Johnson's technique, modified by Rabiej (Rabiej, 2014). The profile of the diffraction line was appropriately normalized and described by a combination of Gauss' and Cauchy's functions with the use of WAXFit software. The procedure was described in detail in the literature (Paukszta et al. 2019). The exemplary diffraction image, after deconvolution and appropriate measurements, is presented in Figure 3. The X c parameter, degree of crystallinity, was determined using dependence: where ∑P c is the sum of the areas* of crystalline peaks and P a is the area of the amorphous peak (its presence is connected with the semi-crystalline model of the supermolecular structure of the polymer). The areas were determined using the graphical method, where the peaks are separated, deconvolution is applied, and the functions are integrated. The amount of pseudohexagonal (triclinic) β phase in an overall crystalline volume of ordered structures in samples was determined using a modified version Turner-Jones equation: k = P β • (P β + P α1 + P α2 + P α3 ) −1 where k is the quantity of the β-form in the crystalline phase of examined material, P β is the surface area of the diffractive maximum deriving from diffraction on a surface (300) of the β-form, P α1 is the surface area of the diffractive maximum deriving from diffraction on a surface (110) of the α-form, P α2 is the surface area of the diffractive maximum deriving from diffraction on a surface (040) of the αform, and P α3 is the surface area of the diffractive maximum deriving from diffraction on a surface (130) of the α-form. The P β1 , P α1 , P α2 , and P α3 values have been calculated using software elaborated by Rabiej (Rabiej, 2014). Additionally, we analyzed the crystallites' size D (hkl) using Scherrer' formula: where k -shape constant (0.9), λ -wavelength of CuKα radiation (1.5418 Å), β -full width at halfmaximum of the diffraction peak (FWHM), and Θ -the Bragg's angle.

Thermal characterization
Thermal analysis (differential scanning calorimetry, DSC) was performed using a Netzsch differential scanning calorimeter, model DSC 200 (Germany), in a nitrogen atmosphere. The samples were heated to 200°C (at a heating rate of 10°C/min) and kept at this temperature for 2 min. Then, the samples were quenched to °C at the rate of 5°C/min (using nitrogen flow). The second-cycle results of the procedure were presented in the paper and interpreted. The value of X c , the crystallinity degree, was determined with the use of the equation: where ΔH m is the experimental enthalpy of melting obtained from DSC melting endotherms, ΔH 0 is the enthalpy of melting of the 100% crystalline PP (209 J/g), and Y is the weight fraction of inclusions in the matrix (25 wt%).

Surface free energy (SFE)
The hydrophobic or hydrophilic character of the material surfaces can be measured by contact angle γ. The measurement of the properties of the filler and the composites has been evaluated by measuring the contact angle of polyethylene glycol (PEG) and water.
All tests were carried out using the optical goniometer Data Physics OCA20. The SFE was determined using the Owens, Wendt, Rabel, and Kaelble (OWRK) method.

Mechanical properties of the composites
The mechanical properties of iPP-brown carbon composites were prepared according to the ASTM D 638 specification. All tests were carried out using a mechanical testing machine -Zwick Roell Z020. The parameters set for the machine during the tests were as follows: the load cell capacity was 20 kN at a cross-head speed of 5 mm/min. Figure 2 shows the diffraction curves obtained during the WAXS analysis of native xylite as well as the xylite chemically modified. The occurrence of simple, commonly occurring in fossil fuels depositions substances, e.g.: aluminum phosphate or calcium sulfate was observed in examined native xylite. After the treatment with KOH solution, the maxima of inorganic compounds are not present in the diffraction patterns, for the reason of leaching them as a soluble potassium salts. However, the effect of the mercerization process (transition from cellulose I into cellulose II) was not observed in the current experiment conditions.

Structural and surface properties of xylites
In order to calculate the SFE for PP and appropriate xylite and modified xylite, the contact angle for water and PEG was measured. The contact angle for the water drop is as follows: PP 105,3° > xylite 93,4° > modified xylite 80,2° as shown in Figure 3.
The contact angles for the drops of nonpolar PEG are equal to PP 56,8° > xylite 47,1° > modified xylite 46,4°. The SFE has been calculated according to the Owens-Wendt-Rabel-Kaelble (OWRK) method. Calculation of the SFE with this model requires the measurement of the contact angles with two known liquids (here, polar water and nonpolar PEG). The calculated SFEs were 44,3 mN/m for xylite, 33,4 mN/m for modified xylite, and 30,8 mN/m for PP. The low and high SFEs correspond to low and high molecular adsorption powers, which are reflected in the properties of the material. The PP exhibits properties of the nonpolar compounds because of a methyl group on its backbone. The contact angles measured for water drops on the surfaces of native and modified xylite show the increase of polar character of used filler with its modification with KOH. It proves the leaching of organic substances deposited on the brown coal powder surface and formation of polar groups. The PP matrix spread well on the surface of the xylite composites, and it would be possible to form better interfacial adhesion between the polymer and the unmodified xylite than that between the modified one.

Investigations of nucleating ability of xylite
It is known from the literature (Garbarczyk and Borysiak 2004) that some defined phenomena may be observed in flowing composite melt. The various phases matrix and filler are characterized by different rates of flow, which cause the increase of shearing on the border matrix/fibers. In the research literature (Szkudlarek et al. 2013), this is also called the "shear amplification effect." Varga and coauthors (Varga and Karger-Kocsis 1996) proposed an explanation for this effect: the α phase nuclei deposited on flowing fibers have a main role in the β-nuclei formation. The α row nuclei induce βphase formation, which has a higher growth rate than the monoclinic α-iPP phase (an appropriate temperature). The conclusions were confirmed by Garbarczyk and Borysiak (Garbarczyk and Borysiak 2004). Moreover, they stated that the growth of adhesion between fibers and matrix results in an increase in the amount of β-phase of iPP.
Considering the results presented in Figure 4, the amount of β-iPP phase (k = 16%) confirms proposed explanations. The amount of β-iPP in foils from virgin iPP and its composite obtained by compression, as well as in injected neat polypropylene, is constant and consists of~5%.
It may be stated that the formation of the β-iPP phase in iPP during cooling, after compression molding, takes place independent of the filler presence. It confirms the role of α-iPP row nuclei deposited on moving fibers, resulting in inducing metastable β-phase formation during injection molding. The iPP melt is not exposed to shearing during compression molding, which results in the absence of pseudohexagonal crystallites in composite volume. Moreover, the differences between matrix and filler thermal expansion, as well as thermal conductivity, do not cause the additional shear stress, inducing β-iPP formation.
The values of X c of examined samples ( Table 2) are worthy of attention. The percentage of crystalline phase in the materials is higher after injection molding than after compression. It proves that the shearing on the (1) matrix/filler border and (2) matrix/mold wall border has a beneficial effect on iPP nucleation. The dispersion of filler in injected samples is far superior to the dispersion in compressed materials, where the formation of agglomerates cannot be excluded. Figure 5 shows the three-layer of the injected dumbbell: skin, interlayer, and core in which can be observed a different degree of material ordering.
In order to exclude the active role of brown coal in the iPP crystallization, the structural characterization of samples within or with various content of additive was performed: 0 (neat polypropylene), 0.5, 5, and 25 wt% of unmodified xylite, respectively. The results presented ( Figure 6) state that the xylite fraction is not an active β-nucleating agent, and it does not beneficially influence the α-iPP crystalline phase amount in the samples' volume. Moreover, the higher amounts of xylite may destabilize the crystallization process and decrease the degree of crystallinity. This is connected with the reduction of iPP chains' mobility. The small content of β-phase is observed in the samples;  however, it occurs spontaneously, and its formation is not connected with brown coal presence in quiescent conditions -without external stress. This fact unequivocally proves the role of external shearing (e.g., during injection molding) on the formation of β-iPP in composites with lignocellulosic filler (Odalanowska andBorysiak 2018, Le Digabel et al. 2004). The DSC results, obtained for compressed foils presented in Table 3 and in Figure 7, conclusively show the influence of brown coal on the iPP crystallization kinetics in composites. The increase of crystallization temperature T c , typical for polymers blended with active nucleating agents, is not observed in this case.  This confirms the thesis proposed above -brown coal is not an active nucleating agent for any crystallographic form of iPP. Moreover, the absence of an endothermic peak at ~152°C suggests that β-iPP does not occur in both samples or occurs in small concentrations. This confirms the thesis: the simultaneous affecting of shear stress (1) on the border between the mold wall and the composite and (2) on the border between the polymer matrix and filler particles of the material induces the formation of a metastable pseudohexagonal form. Nevertheless, the changes of X c suggest the influence of the described filler on iPP crystallization through reduction of free volume and, obviously, the possibility of movement (rotations and translations of chains segments).
The crystallization kinetics are changed by the presence of modified and unmodified coals (Table 3 and Figure 7). The half-time crystallization t 1/2 value is the lowest for the iPP composite with unmodified, native xylite (90 s). The native xylite probably accelerates the primary nucleation processes but simultaneously makes further crystallization limited. Moreover, a comparison of the values of t 1/2 measured for pure iPP (102 s) and iPP with chemically modified xylite (111 s) suggests the unbeneficial influence of coal modification on the course of iPP nucleation as well as on further crystallization (and secondary nucleation). Modification with a potassium peroxide solution, and in consequence occurrence of a more hydrophilic surface of modified xylite, may hamper the chain segments' deposition and the heteronucleation process.

The skin-core structure in dumbbells
Further investigations were focused on nucleation and crystallization processes in the skin and core layers of injected dumbbells. The amounts of β-iPP in samples (k) as well as their degrees of crystallinity (X c ) can be seen in Table 4. Analyzing the values of the k parameter, the strong influence of the  effect of intense shearing in the regions situated close to the mold wall can be observed. A small amount of β-form, created spontaneously (similar like in compressed materials) or induced by shearing, occurs in the skin layer of neat polypropylene. In the skin layer of composites, the percentage of β-iPP is significant -up to 25%. Two factors play the main role: (1) the formation of highly-oriented α-row nuclei, on which the β-iPP grows, and (2) simultaneous local shearing between filler and matrix -the shear amplification effect. Moreover, the β-iPP phase also crystallized in the core regions of composites, but it was not observed in the core of neat polypropylene dumbbells. It only shows the important role of local shearing in the formation of pseudohexagonal form of iPP -the formation of oriented row nuclei should not be possible in core regions (Huihui et al. 2003). Analysis of the degree of crystallinity (X c ) for a sample of unfilled polypropylene shows that this parameter is at a similar level in the entire dumbbell volume obtained by injection. It may be stated that there are two influencing factors: (1) the presence of the unmodified filler and (2) the action of shearing causes the increase of X c in the polypropylene matrix in composites. The presence of the modified filler has a negative effect on the crystallization process because X c inside the dumbbell obtained from PP_Cm composite amounts to only 48%.
From analysis of the crystallites' size D (hkl) ( Table 5) and its distribution against the dumbbell region (1, 2, or 3 and skin-core, respectively), it can be concluded that the crystallites are rather similar in overall sample volume. Moreover, the D (hkl) measured for β-iPP crystallites is slightly higher than for α-crystallites, which confirms the lower efficiency of crystallization of metastable phase.  Nevertheless, the calculated average sizes of crystallites ( Figure 8) suggest a phenomenon consisting in the formation of lower crystallites in neat polypropylene. This presupposes the higher number of point nuclei in neat polypropylene than in composites, probably for the reason of the limitation of the chains' movements. This results in the formation of larger crystallite grains on the normal orientation (Li et al. 2015). It only confirms the explanation proposed above in the current manuscript: the brown coal, native and chemically modified, is not an active nucleating agent for iPP. Moreover, the xylite modification has a negative effect on iPP crystallization due to the increase in the coals' hydrophobicity. Figure 9 shows the selected mechanical properties of the analyzed composites. The mechanical characteristic values of both the composite types xylite and modified xylite only slightly differed in comparison. For both composite systems, a decrease in most of the determined values was noted as compared to the polypropylene matrix. These composite parameters do not differ from the previously described materials filled with natural plant material. As noted by Bledzki and Faruk (Bledzki and Faruk 2006), composites with the addition of a powder-type filler have lower mechanical properties when filled above 20%. According to Diene Ndiaye et al. (Ndiaye et al. 2011), tensile strength and tensile strain in WPCs containing different amounts of wood flour do not change significantly when there is up to 25% wood in the composite. At a higher amount of additives, there is a drastic decrease of these properties (Albano et al. 2001). Figure 9a shows the dependence of Young's modulus on the type of sample. The addition of a xylite fraction filler with a grain diameter of 1 to 1.5 mm to the iPP matrix serving as a matrix allows the obtainment of an approximate value of the linear modulus for the extruded iPP granulate without the addition of reinforcement. For samples modified with the KOH solution, an increase in the value of Young's modulus by 150 MPa was noted in comparison to the polymer matrix. Figure 9b shows the relationship between the tensile stress and the sample type. Regardless of whether the additive was native or modified, the tensile stress values of the obtained composites are lower than the values of the extruded iPP without the addition of reinforcement. It was observed that the values for both types of composites are very similar. The impact toughness analysis was also performed for the tested systems (Figure 9c). The obtained values for the PP control sample and for the PP composite with the filler in the form of a xylite fraction of brown coal with a grain diameter of 1 to 1.5 mm are of similar size (PP 3.1 kJ/m 2 , PP_C 3.0 kJ/m 2 ). It can be concluded that the addition of surface xylite does not affect the impact toughness value. The addition of a modified fraction of bunk carbon to the polymer matrix slightly reduces the impact strength of the material. The PP composites with the addition of modified and unmodified brown coal are characterized by a reduced elongation at break (Figure 9d). For both composites, the recorded values are significantly lower compared to the control sample. The following elongations at break values have been reported for composites 4.95% for PP_C and 6.49% for PP_Cm, which are almost half the elongation of a PP dumbbell.

Conclusions
In this paper, fundamental studies of iPP composites with brown coals -xylites -have been performed. Composites containing 25 wt% of native and chemically modified xylite were considered. They were mainly focused on the correlation between structure and properties of the obtained composites. The results confirm that the xylite from brown coal depositions may be successfully used in obtaining thermoplastic composites, and the mechanical properties of these materials are comparable with another types of WPCs.
Moreover, there is a strong correlation with the formation of β-iPP form and processing conditions. The shear stress between mold wall and composite as well as the local shearing on the border of flowing filler and polymer cause the formation of β-iPP in dumbbell volume. Moreover, the occurrence of beta phase is observed in both of composites -the modification of filler with KOH solution does not affect on the formation of β-iPP. It was confirmed that xylite is not an active nucleating agent for iPP. However, the modification of xylite with KOH solution increases its polarity, and it is a unbeneficial effect on the interfacial adhesion. It is confirmed by results of Charpy' impact strength.
In neat polypropylene and in the composites with 25% wt of native xylite, the impact strength is comparable. The impact strength for composites with modified filler decreases, which suggests the worse adhesion between polar additive and nonpolar matrix.