Tribological Behaviors of Natural Fiber Reinforced Nanofilled Polyurea Matrix Composites at High Temperatures

ABSTRACT The properties of natural fibers such as poor ultraviolet (UV) resistance and hydrophilicity that reduce the durability of composites limit their usage as reinforcements in polymer matrix composites. To compensate this, new types of composites are proposed in the present study. The tribological behaviors of multiphase composites with polyaspartic polyurea matrix produced by vacuum assisted resin transfer molding (VARTM) technique with jute reinforced, filled by graphene nanoplatelets (GNPs) and titanium dioxide (TiO2) nano-fillers at high temperatures are investigated. The wear tests were performed under 10, 20, and 30 N loads, at 60% humidity and temperatures of 100°C, 125°C, and 150°C using a heat moduled pin-on-disc tribotester with an AISI 4140 counter-surface disc having an average surface roughness value of 1.033 µm. The tests were carried out under dry friction conditions along 6000 m sliding distance at a sliding speed of 2 m/s. The results show that the wear damage types changed according to the nano-filler type. Wear loss and specific wear rate increased with increasing temperature and load in almost all types of composites. Using nano-filler caused irregularities in the change of the friction coefficient.


Introduction
Today, polymer matrix composites have become indispensable elements of the manufacturing industry. Especially, due to the widespread use of fiber-reinforced polymer matrix composites (FRPCs), which can be produced in large masses, the environmental problems that composites will create after the end of their life cycle have become an issue that needs to be considered. The pandemic period announced in 2020, which we are still in, has once again demonstrated the importance of the concept of "sustainability" in production. Two options come to the fore for sustainability in FRPCs: Materials can be partially recycled by methods such as mechanical recovery, pyrolysis, solvolysis, cement kiln. However, these methods often concentrate on recycling carbon fiber, which is known that PUr has high resistance to reaction and wear (Samiee, Amirkhizi, and Nemat-Nasser 2013), few specific studies have been done on the tribology of PUr composites (Hu et al. 2009;Wang et al. 2014).
In this study, natural fiber reinforced aliphatic polyaspartic PUr matrix composites, unfilled and containing TiO 2 and GNPs nano-fillers in different proportions (0.5,1,1.5 wt%), produced using VARTM, and their tribological behaviors under different loads and at different temperatures were investigated.

Material
In the study, 11 × 12 inches plain woven jute fabric with 313 g/m 2 areal density and 1.53 wt% oil content, which is in the class of natural fibers and procured from Ganges Jute Private Limited, is used as a reinforcement material. As the matrix material, White Chem PA 1005 polyaspartic PUr, which was procured from AKDE Performance Chemicals is used. The properties of the amine and hexamethylene diisocyanate (HDI) components of the polyaspartic PUr are presented in Table 1. The mechanical and physical properties of polyaspartic PUr are presented in Table 2.
As nano-filler material; anatase TiO 2 and GNPs procured from Nanografi Turkey are used. It is known that using TiO 2 nano filler increases wear resistance. It has been observed that the use of TiO 2 nanofillers in short fiber reinforced epoxy composites reduces the friction coefficient and wear rate (Chang and Friedrich 2010;Chang and Zhang 2006;Zhang et al. 2004Zhang et al. , 2016. Nano TiO 2 , which has become widespread in environment friendly applications due to its photocatalyst effect in recent years, also has high UV resistance. The hydrophilic character of natural fibers weakens their attachment to the matrix. TiO 2 behaves super hydrophilic under UV light (Kavitha et al. 2020). In this way, it is anticipated that the disadvantage of the hydrophilicity of the natural fiber can be partially eliminated by making the modified matrix hydrophilic. In other words, it is intended to strengthen the fibermatrix interface adhesion. GNPs also have a reducing effect on friction coefficient and specific wear rate of polymer composites, and their lubricating effect and particularly high thermal conductivity make GNPs a good alternative as nano-fillers in sliding wear applications (Kazemi-Khasragh et al. 2020;Kazemi-khasragh et al. 2019;Kishore, Amrita, and Kamesh 2021). GNPs nano-filler was preferred based on the assumption that the lubricating effect will decrease the friction force during sliding and its high thermal conductivity will increase the wear resistance by ensuring that the heat on the contact surface is transferred to the composite to form a homogeneous distribution. Technical properties of TiO 2 ve GNPs are presented in Table 3.

Composite manufacturing
Firstly, 230 × 330 mm pieces were cut from the jute fabric roll in the same direction. Then, washed with tap water to remove dust, oil, wax from the cut fabric. Washing away the impurities enhances fiber-matrix interface adhesion but also removes the hard folds of the fabric to maintain easy formability. Washed fabric sheets were dried at room temperature for 1 week. Concurrently, resins for the VARTM process were prepared according to the parameters given in Table 4, with the help of an IKA RW 20 digital mixer. Since it is known that the viscosity of polyaspartic PUr material increases very rapidly, attention has been paid to keep amine and HDI mixing times particularly short and mixing speed particularly low.
As can be seen from Figures 1, 4 layers of jute fabric were placed in the vacuum package and VARTM process was initiated. In the production of unfilled and TiO 2 filled composites, a normal sequential setup is used as in Figure 1a, while in the production of GNPs reinforced composites, double-sided peel ply and distribution media ( Figure 1b) are used. In this way, the problem of poor wetting of the fibers with the matrix material due to the increase in viscosity and agglomerations were reduced and better dispersion was achieved.
After the VARTM process was completed, the vacuum pack was left to cure at room temperature for 24 h, and the composite plates produced after curing were removed from the vacuum pack. Wear test samples of ⊘10 mm were taken from the produced unfilled and 0.5, 1, 1.5 wt% nano-filled composite plates (Figure 2a) using EBH CNC laser (Figure 2b). In order to attach the composite specimens to the test device, composites were adhered to the steel pins with BETA Mitreapel instant adhesive and made ready for the test (Figure 2c).

Tribological tests
While determining the test type and test parameters, preference is made based on; the ability to show the durability limits of the material, compatibility and comparability with the literature as much as possible. Wear tests were carried out using the heat module pin-on-disc type Turkyus POD & HT & WT tribotest device shown in Figure 3. An AISI 4140 steel disc of 58 HRC surface hardness, 20 mm thick, ⊘80 mm is used as the counter surface. The decision to utilize AISI 4140 alloy steel as a counter surface, aside from its extensive use in the aviation and automotive industries, is based on such properties as convenience for surface hardening process and high wear resistance. Since it is aimed to determine the limits of the produced composites by testing under more difficult conditions, a high surface roughness was preferred. The surface roughness of the disc was measured at 0.8 mm sampling length and 4 mm evaluation length, with a Time TR 3200 roughness meter and the roughness curves were obtained using the Time 3200-EN (1.0) software. The average surface roughness value was found as Ra = 1.033 µm and the roughness curve closest to the average value is presented in Figure 4.
Tribotests were conducted in laboratory conditions with an ambient temperature of 20-23°C and a humidity around 60%. PUr is a material that is generally used for waterproofing in the construction industry and the majority of the ambient humidity is removed from the polymer at the selected high tribotest temperatures, no humidity adjustment was required in the test environment. Tribotests were  In determining the test temperatures, the temperature resistance limit of natural fibers described in the first chapter was taken into account. At the beginning and at the end of each test, the samples were weighed with the Necklife FA 2004 precision scale with 0.1 mg sensitivity and the wear loss was calculated by calculating the difference between the values. Specific wear rates, which are a reliable parameter in comparing wear resistance, are calculated by Equation (1).  It is given here as; W: specific wear rate (mm 3 /Nm), Δm: wear loss (g), ρ: density (g/mm 3 ), N: load (N), L: sliding distance (m).
Before each test, the disc counter surface was cleaned with acetone. During the tests, thermal images of frictional surfaces were taken with Testo 875-1i thermal camera at 10-min intervals ( Figure 3) and the temperature value of the highest temperature point in the interface area was taken with the help of Testo IR Soft software. The friction coefficient (µ) values were recorded at a frequency of 4 Hz during the tests with the Turkyus software. For each test parameter, three samples were tested and the average values of the test results were taken. At the end of the tribotests, scanning electron microscope (SEM) and energy dissipative X-ray (EDX) analyzes were performed on composite samples prepared with gold palladium powders and on the opposite surface of the disc using a 20 kV capacity Leo Evo 40 SEM device.
The chemical properties of the jute reinforced unfilled, TiO 2 and GNPs filled PUr composites prepared within the scope of the study were characterized by Fourier-transform infrared (FTIR) spectroscopy. Perkin Elmer Two model FTIR spectrophotometer was used for FTIR analysis. Spectrums were obtained in the range of 400-4000 cm −1 and with a sensitivity of 4 cm −1 . Homogeneous powder samples were prepared from the composites and their components by powder etching method and FTIR analyzes were performed in ATR mode.
The softening temperatures and thermal properties of the samples were determined by differential scanning calorimetry (DSC) analysis. DSC analyzes were carried out using Shimadzu DSC-60 device, in a static air atmosphere with a heating rate of 10°C/min, with 5 mg samples each in Al pan in the range of about 20-500°C. Alumina (Al 2 O 3 ) was used as reference material in the analyzes.

Wear tests
In wear tests performed under different loads and high temperatures, three different parameters are tracked; friction coefficient, wear loss and interface temperature. Wear tests are carried out under 10, 20, and 30 N loads of unfilled composite specimens and specimens containing different proportions of nano-fillers. The average friction coefficient values after the running-in phase of the tests are presented in Figure 5.
As can be seen in Figure 5, µ values of unfilled composite samples have a linear relationship with temperature in which µ values decrease as the temperature increases. In nano-filled composites, no regular relationship between temperature and µ is observed, but generally µ values decrease with increasing temperature. When the behavior of µ with respect to load is observed in unfilled and TiO 2 filled samples, it is seen that generally µ values decrease as the load increases under the same temperature conditions. The decrease in friction force as the load increases while the polymers are in friction contact with steel counter surfaces indicates the elastic deformation of the polymer surface asperities, whereas the increase in the friction force indicates the plastic deformation of the polymer surface asperities (Sinha and Briscoe 2009). Although it was expected that TiO 2 nanofilling would directly decrease the µ values, more irregular results have been encountered. Jiang and Zhang (2013) stated that particle fillers can have positive and negative effects depending on the conditions. As a negative effect, they stated that especially clustered particles can act as an abrasive body and increase wear in sliding, and the continuity of the matrix may decrease due to broken particles. In addition, TiO 2 's super hydrophilic property had a weak effect since it was only under the influence of UV light filtered from the atmosphere. On the other hand, in samples filled with GNPs, a regular relationship between the µ value with the applied load is not observed. Although it is expected that the high lubricity and thermal conductivity of GNPs would reduce the friction coefficient, viscosity enhancement and agglomeration tendency causes more errors in production and weakens the fiber-matrix interface adhesion Erdoğdu, Korkmaz, and Temiz 2021). This state of disorder can be explained by the presence of such unique characteristics of GNPs and by the predominance of one feature over others at certain time intervals. The usage rates of the nano-filler materials used are also decisive at this point. Kazemi-khasragh et al. (2019) studied the sliding wear of GNPs nano-filled basalt fibers-reinforced epoxy composites at different ratios (0, 0.1, 0.2, 0.3, 0.4, and 0.5 wt%). They reported that GNPs fillers improved the wear properties from unfilled to 0.3 wt%, and after 0.3 wt%, they weakened these properties due to GNPs agglomeration. Having a high surface roughness value of the counter surface limited the film formation and reduced the positive effects of nano-fillers. It has been observed that there is no regular relationship between nano-filler ratio and µ in nano-filled composites.
During the wear tests, interface temperatures are monitored utilizing a thermal camera. Heat given by the heat module resistor together with the heat energy resulting from the friction, causes temperature increase on the frictional surface between the composite samples and the steel disc. The interface temperature values followed during the tests carried out under 10 N, 20 N, and 30 N loads are presented in Figures 6-8, respectively.
Increasing both the sliding distance and the temperature in all samples naturally increased the interface temperature (Figures 6-8). Maximum interface temperature values and maximum interface temperature difference values are seen in experiments conducted under 30 N load and 150 °C temperature conditions. While the interface temperature values in GNPs filled composite samples generally show a high tendency to increase up to about 150 °C, the tendency of the interface temperature values to increase at values higher than 150 °C 'has somewhat decreased (Figures 6c-8c). The reason behind this is the significant decrease of friction coefficient value as the temperature rises after this interface temperature value. The highest interface temperature is measured as 165.1 °C on unfilled composite samples under 30 N load and 150 °C temperature. Wear losses of unfilled and nano-filled composite samples are given in Table 5.
Since the tests are carried out at 100 °C and higher temperatures, it should be taken into account that the wear loss values are actually lower than the values given in Table 5, considering that there is additional loss due to moisture. In this case, it can be said that the wear resistance of PUr matrix composites is generally high. In general, wear loss increases with increasing load and temperature. Under the most extreme conditions (30 N load and 150 °C), it is understood that the lowest wear loss is in samples filled with 0.5 wt% GNPs. At the same temperature values, the highest wear rates are generally observed under 10 N load (Table 6). Under the same temperature conditions, the increase in the load caused the wear rate to tend to decrease, albeit irregularly. Under the same loads (except 0.5 wt% GNPs samples under 30 N load), wear rate increases with the increase of temperature in all composite samples. However, if it is taken into consideration that the µ value generally decreases with the increase in temperature, it is understood that a film layer is formed between the steel counter-surface and the composite samples. In addition, analyzing the data in Table 6, it can be seen that temperature is more determinant in the  change of wear rate than load. It should also be noted that, in the most extreme conditions (30 N load and 150 °C temperature), highest wear resistance is in samples filled with 0.5 wt% GNPs.

EDX and SEM analysis
Following the wear tests, the material transition from polymer composite to AISI 4140 steel counter surface is investigated. Visually observing the counter surface, very few and poorly adhered polymer residues are found to be evident. In order to examine this in detail, a SEM image is taken on the metal surface after the wear test of the TiO 2 nano-filled sample by making use of the distinctiveness of the Ti element and EDX analysis is performed throughout various regions (Figure 9).
When Figure 9 is examined, even the fragment (Objects 1) detached from the refractory of the heat module is seen in the EDX data taken from different points and regions, no evidence of Ti element was found. This shows that even if the film layer is formed during the test, and even the material is transferred to the metal surface, there is no actual polymer transfer residue on the metal surface after the wear test under high temperature and pressure. The SEM and EDX analysis image of the regions with rare and poorly adhered residues is presented in Figure 10.
When Figure 10 is examined, it is seen that Ti element is found in EDX analysis of polymer residues that are noticeably large and weakly adhered on the disc surface. Although film formation and polymer  transfer to metal surface occurs as a result of using an elastomer matrix such as PUr and high sliding rate and temperatures engaged, it's seen that polymer residues are mostly removed from the metal surface simultaneously with the test. SEM images are taken from the worn surfaces of the composite specimens in order to observe the wear damage and wear mechanisms. SEM images of unfilled samples can be seen in Figure 11. During the wear of multiphased polymer composites with fiber and filler, first the matrix protecting the fiber is worn, then the fiber itself is worn and thinned, then it is broken and separated by breaking (Chang and Zhang 2006). These stages, which include matrix and fiber wear damages due to sliding, are shown in Figure 11. In addition to these damages, sharp-edged pits and cracks in the matrix, indicating the fatigue wear mechanism, are also observed. In Figures 12 and 13, SEM images of the worn surfaces of TiO 2 and GNPs filled composites are shown, respectively.
In Figure 12, damages due to adhesion is seen first in the form of matrix wear due to sliding, then interface wear, fiber cracking, and finally breakage. The occurrence of damage in such a sequence also indicates that the fiber matrix interface adhesion is strong. In Figure 13, similar to Figure 11, matrix cracks and sharp-edged pits are seen, indicating strong adhesion damage as well as fatigue damage. Conducting the tests at a long sliding distance and the relatively weak fiber-matrix interface adhesion were effective in the occurrence of fatigue damage. First, crack zones were formed in the matrix and the cracks deepened due to fatigue and caused large breakages in these areas. By increasing the cohesive strength of polymers, nano-fillers can increase the resistance of composites to abrasive wear and fatigue wear (Sinha and Briscoe 2009). This is also linked to fiber-matrix interface adhesion. TiO 2 nano-filler showed a similar effect, increasing the resistance to fatigue wear, too. On the other hand, in unfilled composites and GNPs nano-filled composites, significant fatigue damages are observed in addition to adhesive wear damages due to the effect of weaker interface adhesion.

FTIR characterization
FTIR spectrums prepared for structural analyzes of composites are shown in Figures 14 and 15. Looking at the spectra of the TiO 2 filled composites in Figure 14, the Ti-O-Ti stretch in the range of 507-930 cm −1 on the peak of the pure nano TiO 2 structure is seen in Figure 14a. In addition, Ti-O   stretch is evident around 411 cm −1 . The spectrum of the jute fiber structure is shown in Figure 14b. In this spectrum, especially the -OH group -H bond stretching vibrations of the cellulose groups in the jute structure give a wide band peak in the range of 3000-3600 cm −1 . Again, C-H groups in the cellulose structure give aliphatic methyl peak in the range of 2850-2950 cm −1 . C-O-C etheric stretching vibration is observed in the range of approximately 900-1100 cm −1 . Figure 14c shows the FTIR spectrum of pure PUr. In this spectrum, C-H stretching vibrations originating from polymeric main chain are seen in the range of 2810-2960 cm −1 . Especially around 1686 cm −1 , a very strong C=O (carbonyl) peak is observed. Around 1450 cm −1 C-N stretching vibration and around 1371 cm −1 C-C aliphatic stretching vibration is observed. In addition, the C-O band peak around 1220 cm −1 and the C-H band peak around 760 cm −1 confirm the PUr structure. In the spectrums in Figure 14d, e and f, the -OH group -H bond in the range of 3000-3600 cm −1 and the C-O-C etheric stretching vibration around 1025 cm −1 are seen, originating from the fiber groups in the structure. In addition, C=O, C-N and C-C peaks, which are seen around 1682,1451,1370 cm −1 , respectively, confirm the PUr structure. As the amount of TiO 2 in the structure increases, the Ti-O peak intensity around 530 cm −1 increases and this proves that the desired structure is obtained.
Considering the spectrums of the GNPs filled composites in Figure 15, there is no significant peak in the spectrum of pure GNPs in Figure 15a. This is compatible with the structure of GNPs. In Figure 15d, e and f spectrums, peaks belonging to composite structures with increasing GNPs filler content are seen. In all three composite structures, the basic peaks in both jute and pure PUr structures are clearly visible. Especially the band peak in the range of 3000-3600 cm −1 confirms the presence of jute in the composite. In addition, C-N peaks around 1450 cm −1 and C=O peaks around 1680 cm −1 confirm the presence of PUr in the composite. In addition, as the amount of GNPs in the structure increases, the C-C peak intensity around 1371 cm −1 increases and this proves that the desired structure is obtained.

Thermal analysis
The thermal curves of pure PUr, 1 wt% TiO 2 and GNPs filled composites obtained after DSC analysis are shown in Figure 16. When the thermal curve of the PUr structure in Figure 16a is examined, it is seen that the glass transition temperature (Tg) is 98.65°C. This value is in accordance with the literature and confirms the aliphatic structure of PUr. There is also a two-stage thermal degradation process. In the first stage, degradation is seen in the crosslinking units in the PUr structure, while the main PUr structure is stable. At around 263°C, thermal degradation of the PUr structure begins and ends at 440.64°C. When the thermal curve (Figure 16b) of the composite prepared by filling TiO 2 into this structure is examined, it is seen that the Tg value has increased to 131.32°C. In addition, the other two thermal degradations remain stable. The first degradation starts at 188°C and continues up to 265°C. The second degradation starts from 265°C and continues up to 435.37°C. When the thermal curve in Figure 16c is examined, it is seen that the Tg value of the GNPs filled composite is 145.74°C. Considering the thermal degradation, it is seen that the first mass loss is in the range of 178.94°C − 251.18°C. Due to the degradation of GNPs, the area under the peak in this region is larger than that of other samples. The second decomposition exotherm peak, which is caused by the degradation of the polymeric structure, starts at 260.36°C and ends at 441.06°C. When all the data are examined together, it can be said that the composite structures are thermally stable up to about 190°C and both added nanofillers increase the Tg value.

Conclusions
In this study using the VARTM method, woven jute reinforced, polyaspartic PUr matrix, both unfilled and nano-filled composites are produced, and their tribological behavior at high temperatures is investigated. The usability of polyaspartic PUr as a matrix and its producibility with the VARTM method is tested. The effects of TiO 2 and GNPs nanofillers with different filling ratios on tribological properties are investigated. At the end of the study, the following results are collected: Polyaspartic PUr material is used as a matrix in composites and found to be compatible with natural fiber reinforcement and the VARTM technique.
It is observed that the nano-filler used in multiphased polymer composites significantly affects the dominant wear mechanisms. It is also understood that while adhesive and fatigue wear mechanisms are dominant in unfilled and GNPs filled composites, only the adhesive wear mechanism is dominant in TiO 2 filled composites.
The increase in temperature in unfilled composites resulted in a decrease in the friction coefficient. Although the increase in temperature in nano-filled composites also causes a tendency to decrease the friction coefficient in general; the modification of the matrix caused serious differences in the friction mechanisms which in turn, caused irregular changes in the friction coefficient.
Except for 0.5 wt% GNPs filled samples, wear loss and specific wear rate increases with increasing temperature and increasing load in all samples with and without nano filler.