Effect of Chemical and Physical Treatments on Mechanical Properties of Diss Fibers-Based Biocomposites Materials

ABSTRACT This work highlights the exploitation of fibers from a promising plant, commonly called Diss, in polymer matrix reinforcement. In this context, various untreated and treated (with acetic acid, silane, NaOH, thermal) Diss fibers were used to reinforce a thermoplastic matrix, polypropylene (PP). The effect of fiber treatments was evaluated at a ratio of 10% (by mass) of fiber, using mechanical, microstructural and physical characterization. A fiber content optimization study was carried out after choosing the fiber treatment by varying the fiber content from 10% to 40%. The microstructure observations showed a multi-layer structure for the biocomposites studied. The particular morphology of Diss fibers, with the presence of thorns, seems to improve the mechanical adhesion between the two phases. For 10% fiber content composites, the incorporation of the different fibers in the PP matrix led to improvements in most mechanical properties. The biocomposite reinforced with NaOH treated fibers (NPP-10%) showed the best mechanical performance with an improvement of 30%, 22% and 9% in tensile modulus, bending modulus and bending stress, respectively. For the fiber content variation, the NPP was retained, and an enhancement in the tensile modulus was found when the fiber content was below 20%.


Introduction
The use of biocomposites is increasingly adopted in industry as an alternative for conventional composites . The term "biocomposites" refers to polymer composites made from a biosourced or petroleum-based matrix reinforced with plant fibers ). The advantages of using such materials are: ecological, by substituting synthetic fibers produced by energyintensive processes and nonrenewable resources with renewable, degradable and less energy-intensive fibers; economic, because the production cost of natural fibers is much lower than their synthetic fibers extracted by a combined chemical and enzymatic method. These fibers exhibited tensile strength, Young's modulus and failure strain of 19.11 ± 9.01 MPa, 3.05 ± 0.41 GPa and 1.95% ± 0.36% respectively. Recently, Nouri et al. (2019) found a tensile strength and Young's modulus of 273 ± 36 MPa and 11.46 ± 2.2 GPa, respectively, using a mechanical extraction method based on the Diss plant morphology. The treatments' effect was also studied. The Young's modulus can increase significantly after different treatments (with acetic acid, silane, NaOH, thermal). For mechanical properties of Diss based biocomposites, (Sarasini et al. 2019) found that the incorporation of Diss fibers into a polypropylene (PP) matrix did not show any improvement in the mechanical tensile properties compared to the matrix alone. However, the use of a coupling agent (maleic anhydride (MA-g-PP)) resulted in a significant improvement in all these properties with a ratio of no more than 25% for composites with a fiber content of 15% by mass. On the other hand, Touati et al. (2018Touati et al. ( , 2021 found a decrease in the overall mechanical tensile properties of PP (recycled/non-recycled)/polyethylene composite with compatibilizer and 5% in weight (wt%) of Diss fibers compared to the matrix alone. Moreover, Diss fibers have a particular morphology compared to other plant fibers; Nouri et al. (2019) found the presence of thorns on the surface of a large part of the manually extracted fibers. These thorns do not seem to be affected by all the chemical treatments carried out in this study. The authors concluded that this morphology could improve the adhesion between fibers and the matrix, especially when using hydrophobic matrices such as PP.
Due to the number of limited works dealing with the integration of Diss fibers in polymer composites in the literature and the low ratio of fiber introduced, this work aims to present a twoscale approach (micro/macro) in order to develop a new biocomposite material based on Diss fibers. The extraction method developed by Nouri et al. (2019) was used to obtain fibers with good mechanical performance. The effects of treatments and the fiber loading rate were investigated by using fibers that have been extensively characterized in a previous work (Nouri et al. 2019(Nouri et al. , 2021. Particular attention was given to the state of the Diss fiber/PP interface by means of scanning electron microscope (SEM), for which the particular morphology effect of Diss fibers has been identified.

Diss fibers extraction
The Diss plant leaves were harvested by hand in the north of Algeria at the end of their maturity (in 2018). Subsequently, the untreated technical fibers of Diss (UTFD) were extracted manually according to the method of Nouri et al. (2019). Finally, four treatments were applied to the UTFDs as follows (Nouri et al. 2019): • NaOH treatment (NTFD): the UTFD were submerged in an aqueous solution containing 5% NaOH at room temperature during 5 h. Afterward the technical fibers of Diss (TFD) treated with NaOH were first cleaned by immersing them in distilled water for 24 h. Then, they were submerged in a distilled water solution containing 2% of acetic acid in order to adjust the pH to 7. Finally, the NTFD were washed several times with tap water and dried in the oven at 60°C for 3 h. • Silane treatment (STFD): Octyltriethoxysilane (2%) was dissolved in a mixture of distilled water/ ethanol with a volume ratio of (0.40/0.60). This solution was adjusted using acetic acid until the pH value was equal to 4. Then the solution was stirred for 2 h. Afterward, the UTFD were submerged in the solution for 2 h at room temperature. Finally, the Silane-treated TFD were cleaned with tap water and dried in the oven at 60°C for 3 h. • Acetic acid treatment (ATFD): the UTFD were immersed in an acetic acid solution at room temperature during 90 min. Then, the acetylated TFD were treated with an ethyl acetate solution containing two drops of sulfuric acid to remove excess of acetic acid. Finally, the ATFD were cleaned with tap water and dried in the oven at 40°C for 24 h.
• Thermal treatment (TTFD): the thermal treatment of UTFD was carried out in an oven at a temperature of 140°C during 14 h. Then, the thermally treated TFD were cooled down at room temperature.

Biocomposite moulding
The thermo-compression molding method was used to make plates of different Diss fiber/PP biocomposites. PP sheets of 1 mm thickness were supplied by DVPLAST. A multilayer, composed of three PP sheets and two short Diss fiber mats (randomly dispersed), as shown in Figure 1a, was introduced into the mould/cover system (Figure 1a). Then the assembly was placed in a hydraulic press with heating plates. The mold and press were preheated at a temperature of 180°C. This temperature remained constant throughout the operation to preserve the fibers which degrade at 200°C and above. The pressure for the manufacture of the different biocomposites plates (Diss fibers/PP) and PP (Table 1) was applied following the protocol presented in Figure 1b: the material was heated for 3 min without pressure, then the pressure was applied gradually from 2 bar to 9 bar, finally, the mold was removed and cooled in water.
These plates have dimensions of 24 × 24 cm2 with a thickness that depends on the amount of material used and varies between 2.5 and 4 mm.

Scanning electron microscope (SEM)
The different observation planes of the biocomposites were examined using a JEOL 6060 SEM. The samples were prepared by coating. The observation surface was positioned parallel to the observation plane and then coated in a cold-coating resin (versocit-2). Finally, the observation surface was polished with a polishing machine (Jean Wirtz FT 250). On the other hand, the fracture surfaces were also examined after the tensile tests without coating.

Density
The density of the biocomposites and the PP matrix was determined by the hydrostatic weighing method. This method consists of weighing the sample in two times: once in air (A) and once in a liquid (B) with a known density (ρ l ). For each biocomposite, three samples of 10 × 10 mm2 were extracted from the plate at different locations and dried for 48 h in an oven at 60°C. Ethanol was chosen as the liquid due to its low density compared to biocomposites. A Sartorius balance (1/1000) with a suspension measuring device was used for weighing. The density of ethanol was given by the provider and that of the biocomposites (ρ c ) was calculated by the following equation according to the international standard ISO 1183-1:

Tensile tests
Tensile tests on the biocomposites were carried out by using Instron 3356 and 5566 machines equipped with a 500N and a 1000N load cell, respectively. All tests were performed under controlled laboratory conditions, 23°C and 55%-65% relative humidity (RH). Five specimens of 1B type were obtained from each plate by a punch in accordance with ISO 527-4 (1997). First of all, in order to determine the Young's modulus, the 500N load cell was used and the specimens were loaded to small deformations less than 0.4%. In view of the sliding problem between the jaws and the specimens, an INSTRON extensometer with a gauge length of 50 mm was used. The deformation rate was set at 0.5 mm/min according to the recommendation of (NF EN ISO 527-4 1997). Second of all, the same specimens were loaded until failure using a 1000N load cell with a loading rate of 5 mm/min.

Bending tests
The INSTRON model 3366 machine with a 500N load cell was used to perform three-point bending tests. Tests were carried out under controlled laboratory conditions (23°C and 55%-65% RH) with a loading rate of 2 mm/min and a span of 64 mm according to (NF EN ISO 14125 1998).

Mechanical properties of biocomposites
Tensile properties. Figure 2 shows the mechanical properties (tensile strength (σ t ), tensile modulus (E)) of PP plate and the different 10% fiber content biocomposites plates. The PP plate has σ t and E values of 31.01 ± 0.37 MPa and 1.39 ± 0.08 GPa, respectively. The last one increases significantly after 10% of loading in UTFD. This phenomenon becomes more noticeable with treated fibers. NPP-10% shows the highest modulus, followed by TPP-10%, SPP-10%, APP-10% and UPP-10% with an improvement of 30%, 23%, 17%, 16% and 12.50% respectively. Nevertheless, a decrease in σ t is observed for these composites. This loss of resistance is more considerable for UPP-10% with a reduction of 21% compared to PP. The treated fiber-reinforced composites showed a more or less similar reduction, with 16%, 16.40%, 12.70% and 10%, respectively, for APP-10%, NPP-10%, TPP-10% and SPP-10%. Similar results are often reported in the literature with regard to the tensile strength reduction for plant fiber/PP biocomposites (Pérez-Fonseca et al. 2014;Stark 1999).
The biocomposites seem to be influenced by the intrinsic properties of fibers. Indeed, according to the work of Nouri et al. (2019), NTFD and STFD show the best tensile modulus values for the Diss fibers studied, followed by ATFD and TTFD, whose difference is not significant according to the uncertainties of the results, and then UTFDs. For biocomposites, NPP-10% and UPP-10% are in first and last position respectively, similarly to what was observed for fibers. Contrary to what was expected, the STFDs do not show an improvement in the chemical adhesion in biocomposites. The choice of the silane molecule (3-Aminopropyl triethoxysilane) was made because of the compatibility of its amine functional groups with those of the PP matrix. Thi (2011) found that such a coupling agent led to the highest Young's modulus among the studied bamboo fibers/PP biocomposites, even when compared to biocomposites with MA-g-PP. The use of a coupling agent, such as the latter, may be a solution to further improve the mechanical performance of this biocomposite, particularly with regard to tensile stress.
Bending properties. Figure 3 shows the bending properties of PP plate and the different 10% fiber content biocomposites. The PP plate has a bending strength and bending modulus of 34.13 ± 1.95 MPa and 1.28 ± 0.02 GPa, respectively. These results are of the same order of magnitude as those obtained from the tensile test. This is due to the homogeneous nature of PP. The bending properties exhibited by the biocomposites are completely different from those in tensile. APP-10%, NPP-10%, TPP-10%, SPP-10% and UPP-10% show an increase in modulus of 22%, 20%, 13.50%, 10.50% and 7% respectively compared to the PP matrix. These results are discussed below. By taking into account measurement uncertainties in bending stress, NPP-10% shows a significant improvement of about 9% compared to PP, while UPP-10% shows a decrease of 17%. For the other composites, the results are more or less similar to those for PP plate. These results, discussed in the next part, are still significantly higher than those obtained in tension. Figure 4 shows the SEM observations made on different sections of the APP-10% sample. It was observed on the cross-sectional views (Fig .4.1a) that these biocomposites have a layered structure with 3-layer: two layers of skin on the outer parts, matrix only, and a thicker middle layer (active layer) with ATFDs distributed in the PP matrix. Within this, we can also distinguish three sub-layers, two of which have a high fiber content, separated by a thin sub-layer composed only of PP. This multi-layer architecture of the composite is a consequence of the manufacturing process used. It was chosen in order to create a protective layer for the plant fibers (skin layer), especially when the material is used under changing hygrothermal conditions, which makes it possible to use this biocomposite in these conditions without any additional treatment of its outer surface. At this enlargement, no porosity was observed and fibers appear to be well surrounded by the matrix. These observations can explain the mechanical behavior of these composites in bending in particular for the modulus. Indeed, the fibers appear to be requested by only a part of the maximum applied stress, as the position of the active layer is in the middle of the biocomposite and the maximum tensile stress applied in the lower limit (lower outer part) which is mainly composed of PP. Several authors have reported similar behavior where the tensile and bending modulus are different (Le Duigou and Castro 2015).

Analysis of the microstructure
When the resolution is high, still on the same plane ( fig.4.1b), a debonding between fibers and the matrix is clearly observable, indicating a poor chemical adhesion between them. The thickness of this interfacial region seems to be relative to the fibers size. These findings are often raised in the literature when using such a matrix with plant fibers (Chen, Guo, and Mi 1998;Law and Ishak 2011;Le Duigou and Castro 2015). The observation in plan of the active layer was also reported ( fig.4.2). For an enlargement of 18 times ( fig.4.2a) the fibers appear to be well surrounded by the matrix and dispersed in a homogeneous and random way without cracks or porosity within the matrix. When the enlargement is high ( fig.4.2b), debonding is clearly observable. These observations could justify the decrease in the tensile stress of biocomposites due to the presence of stress concentration zones. It should be noted that the fibers are randomly oriented, which can amplify this phenomenon, particularly when the fiber is oriented perpendicular to the direction of the applied load.
We can also observe the presence of thorns in fibers ( fig.4.2b) which seem to provide more mechanical adhesion with the matrix. These thorns are frequently present in the Diss fibers after  Figure 4.3 gives more information on their geometry. These thorns provide the fibers a good mechanical adhesion with the PP matrix, which could explain the improvement of the Young's modulus in the case of tensile tests. The findings raised in this section were also observed for the other 10% fibers content biocomposites. Figure 5 shows the SEM images of the fracture surfaces of the different biocomposites. The following findings were raised for all types of composites: (i) a debonding between fibers and the matrix with the presence of fiber imprints when fibers are oriented perpendicular to the loading direction; (ii) a break in the fibers oriented parallel to the loading direction; (iii) a combination of the two above failure modes when fibers are oriented between these two directions. However, no traces of PP were observed on the pulled fibers, indicating once again the poor chemical bonding between fibers and the matrix.
In fact, during the tensile loading of the specimen, sections containing more fibers oriented vertically to the loading direction are the most vulnerable to failure. Images (a) and (b) in Figure 5 show two fracture surfaces of a specimen loaded in traction where we can observe the presence of this  kind of fibers on one part and its imprint on the other part. This leaves the rest of fibers and the matrix to withstand the loads in this section. At higher stresses close to the failure of PP, the matrix in this section starts to deform in a ductile manner until it is tearing (see Figure 5c) leaving the rest of fibers to resist the loading. These fibers do not persist for a long time and adhesive (fibers pull-out) or cohesive (breaking of the fibers) failures occur depending on their adhesion to the matrix (Figure 5 (d) and (e)). This adhesion, which can only be mechanical, is governed mainly by the presence of thorns. Figure 5d shows a cohesive break in the fiber due to good adhesion of the thorns to the matrix. We can also see the imprint of the thorns on the matrix. Therefore, the conclusion is that the decrease in the composites tensile strength is mainly due to the presence of stress concentration areas caused by poor chemical adhesion. Whereas the presence of the thorns ensures the transfer of loads between fibers and the matrix and leads to an improvement of the elasticity (Young's modulus).

Analysis of the density
The PP matrix density has been measured at 0.91 ± 0.001 g/cm 3 . This value is in the same order of magnitude as that of UTFD, measured at 0.93 ± 0.01 g/cm 3 in the work of Nouri et al. (2019). Figure 6 shows the density results for the different 10% fiber content biocomposites. APP-10%, TPP-10%, SPP-10% and NPP-10% show more or less similar values higher than UPP-10%. It seems that the removal of some non-cellulosic components from fibers during their treatment has led to an increase in their density.
Indeed, Nouri et al. (2019) concluded that treatment with NaOH significantly degrades the noncellulosic components of Diss fibers, followed by silane and acetic acid treatments. The heat treatment does not significantly change the chemical composition of fibers. These degradations should be accompanied by an increase in the density of these fibers as follows: NTFD > STFD ≈ ATFD > TTFD > UTFD. The evolution of the density of the different biocomposites follows this rank with the exception of the TPP-10% plate. The latter showed a density equivalent to that of NPP-10%. Two hypotheses are possible to interpret these results for TPP-10%: samples are taken from local fiber concentration, or rearrangement of the crystalline structure during heat treatment of fibers resulted in a reduction in fiber volume.

Results analysis on the effect of fiber content in NTFD-based biocomposites (NPP)
Following the different results presented in the previous section, NPP composites were chosen to study the effect of fiber content on the mechanical properties of Diss/PP biocomposites. Figure 7 shows SEM images of NPP-20%, NPP-30% and NPP-40%. For each composite, images (a) and (b) are seen in the cross section of the composite, while image (c) was examined in the plane of the active layer. For NPP-20%, the thickness of the active layer has significantly increased at the expense of the skin layers and the inter-fiber space appears to be significantly reduced compared to NPP-10% (Figure 7. (1a) and (1b)). Furthermore, fibers are randomly distributed and well surrounded by the matrix at this enlargement ( Fig.7.1c). Regions of fiber agglomerates can sometimes be observed on the active layer ( Fig.7.1a). At present, no porosity has been detected in the matrix. However, debonding between fibers and the matrix is always noticed at high resolution ( Fig.7.1b). For NPP-30%, the skin thickness seems to be greater than for NPP-20% and at the same time the space between fibers seems to be reduced further (Figure 7 (2a) and (2c)). Microcracks can sometimes be observed due to fiber-to-fiber contact ( Fig.7.2b). For NPP-40%, a significant reduction in the thickness of the skin layer compared to the other composites was observed (Fig.7.3a). Fiber-to-fiber contact seems to become much more important, as well as the presence of microcracks (Fig.7.3b). The latter appear to become larger than those observed on NPP-30%. They are even observable at low enlargement (Figure 7 (3a) and (3b)). In addition, porosity in the matrix was clearly observed (Fig.7.3c). Figure 8 shows the density values of the different NPP. An increase in density as a function of fiber loading rate was observed following a linear law when the latter was below 30%. At 40%, the density has decreased, indicating the presence of significant porosity. This porosity was clearly observed on the SEM images. In addition, delamination regions were also observed on this composite with the naked eye. Indeed, when the loading rate is high, fibers are not sufficiently impregnated by the matrix, which creates fragile regions where the fiber-to-fiber contact is high, leading to the presence of voids and the delamination of the biocomposite.

Mechanical properties of NTFD-based biocomposites (NPP)
Tensile properties. Figure 9 shows the tensile mechanical properties of NPP based composites containing various percentages of Diss fibers. The tensile strength decreases with increasing loading by 6%, 18% and 28% for NPP-20%, NPP-30% and NPP-40%, respectively, compared to NPP-10%. The Young's modulus improves as the loading rate increases. This is the case when the loading rate is less than 20%, above this value, no significant improvement was observed. The decrease in tensile stress could be explained, as discussed below, by the increase in stress concentration regions with increasing fiber loading rates. As regards modulus, we see an improvement when fibers are well impregnated by the matrix at 10% and 20% loading. When the fiber content reaches 30% of loading, the presence of local microcracks is observed (Fig.7.2b). This leads to a locally poor load transfer between fibers and the matrix. At the same time, the addition of fibers is also locally beneficial. Therefore, compensation between these two facts leads to a more or less stable Young's modulus. Chen, Guo, and Mi (1998) observed a similar behavior for tensile modulus when Bamboo/PP composites have a fiber content below 40%. Above this value the modulus starts to decrease. However, other authors have observed an increasing modulus evolution even if the fiber loading rate is higher than 30% (Dhanalakshmi, Ramadevi, and Basavaraju 2017;Pérez-Fonseca et al. 2014;Stark 1999).
Bending properties. Figure 10 summarizes the bending properties of NPP. The maximum bending stress is higher when the fiber content is 10% (37 MPa). At higher fiber contents the stress drops and stays similar for all remaining biocomposites, between 26 MPa and 28 MPa. These values are lower than those for PP. The bending modulus does not appear to change significantly regardless of the fiber content of the biocomposites. However, it is still higher than that of PP. In the literature, we often observe an increase in bending modulus with increasing load rate (Le Duigou and Castro 2015;Pérez-Fonseca et al. 2014). However, in our case, the multilayer structure of the composite, as explained above, seems to affect the evolution of this property. Certainly, the increase in the number of layers could improve the bending modulus of the composite.

Conclusion
In this work application of Diss fibers as reinforcement to polymer composites has been studied. The treatment effect was also examined by three chemical treatments and one thermal treatment.
The different biocomposites showed a layered structure with no porosity in the matrix. However, at the scale of the fiber/matrix interface a lack of chemical adhesion was noticed.
At 10% fiber loading, an improvement in mechanical properties (modulus and stress) in bending and tension was reported, except for tensile stress. NPP-10% showed a better improvement in these properties compared to other composites of the same loading rate, 30%, 22% and 9% for tensile modulus, bending modulus and bending stress, respectively. A combined adhesive and cohesive failure mechanism was observed depending on the orientation of fibers with respect to the loading direction which the presence of the thorns seems to improve the mechanical adhesion.
With increasing loading rate, an increase in tensile modulus was found for NPP composites when the fiber loading rate was below 20%. Above this value, this property seems to remain constant.
Studies will be carried out in the future to improve the chemical adhesion between the Diss fibers and the PP matrix in order to obtain better mechanical properties.