Properties of Multiple-Processed Natural Short Fiber Polypropylene and Polylactic Acid Composites: A Comparison

Abstract

Natural fiber composites have gained increasing attention due to sustainability considerations. One often neglected aspect is the potential for the mechanical recycling of such materials. In this work, we compounded injection-molded polypropylene (PP) and polylactic acid (PLA) short cellulose fiber composites with fiber shares up to 40 percent by weight. Both matrix materials were reinforced by the addition of the fibers. We investigated a trifold full recycling process, where we subjected the materials produced in the first place to compounding, injection molding, testing, and shredding, and then repeated the process. Although the materials’ properties assigned to degradation were found to decrease with progressive recycling, attractive mechanical properties could be preserved even after the third reprocessing cycle.

1. Introduction

In transport industries like the automotive or aerospace industries, thermoplastic composites have long been standard and have become indispensable due to their light-weight properties, with simultaneous unbeatable mechanical capabilities. The commercial uses of bio-based materials in the plastics economy has increased due to the rising awareness of environmental issues [1]. Using natural fibers as reinforcements in polymers can be an alternative to man-made fibers like aramid, glass, and carbon fibers. Several types of natural fibers and production techniques are applied and yield interesting materials for industrial applications [2,3]. In addition to their eco-friendliness, natural fibers offer several other advantages: low density, low cost and broad availability, non-toxicity, and high specific strength and modulus [4,5,6]. Natural fibers are considered to be recyclable, their production has low energy consumption, they are CO₂-neutral due to their natural origin in plants, and, overall, are they are biodegradable [7].

Despite these advantages, natural fiber composites with synthetic polymers may not reach the strength of composites with glass fibers, for example. The fiber–polymer interface adhesion is poor due to incompatible surface chemistries between the fibers and matrix and a lack of wetting in the fibers and fiber–matrix-interaction [4]. Other drawbacks in in the application of natural fibers in polymer composites is their water uptake and low thermal stability [7,8].

Although natural fibers are often applied in the production of continuous fiber-reinforced composites using thermoset or thermoplastic matrices [4,9], our research was made in the field of short fiber-reinforced thermoplastic composites, as for such composites we have a higher degree of freedom in processing, e.g., injection molding and extrusion, being the two most widespread processing technologies, can be applied.

Bio-composites are, according to the definition by Mohanty et al. [10], composites where a minimum of one of the components has to be from a natural origin [11]. Following this definition, both PP-cellulose and PLA-cellulose composites investigated in our research are bio-composites, but those with PLA seem to have the higher positive impact in terms of sustainability. PLA is not only of biological origin [12], but it is also considered to be bio-degradable under certain conditions [13,14,15]. However, bio-degradation is not necessarily the best end-of-life scenario, as the materials dwindle away when composted and become lost in a circular economy. To keep a material in the circle, it has to be recyclable.

Several authors have dealt with the issue of preparing and comparing injection or compression molded bio-composites with various types and various shares of cellulosic fibers [16,17]. Bio-composites, either with PLA or PP matrix with microcrystalline cellulose [18], wood flour [19], or the combination of both [20], result in bio-composites with suitable properties. Overall, the findings of many studies were that natural fibers were able to substitute glass fiber-reinforced PP in some applications [7] and mechanical properties of PLA–natural fiber composites were found to be able to compete with those of conventionally used thermoplastic composites [21].

Different types of PP and PLA were mixed with a set of various lignocellulosic fibers, and among others also wood flour, in a mixer and compression molded into plates of 1 mm thickness. For some composites with PP, maleic anhydride was used as coupling agent. Micromechanical investigations revealed that fiber fracture is the most dominating effect causing composite failure, unless the fiber and matrix have good adhesion. The overall strength of the composites is determined by the matrix and fiber strength, hence only stronger fibers could further increase composite performance [22]. Comparing PLA with PP natural fiber composites prepared via compression molding revealed that the mechanical properties of the PLA composites were able to compete with those using PP as matrix, and the molecular weight of PLA was preserved to a large extent during compounding [21]. Challenges in replacing PP with PLA are also the chemical surface composition of natural fibers [23]. Although interactions between cellulosic hydroxyl groups and PLA carbonyl groups could contribute to the fiber–matrix connection via hydrogen bonding [24], improvements of interfacial properties between the fiber and matrix could generally be achieved by surface modification or the use of compatibilizers [3].

The effect of thermal degradation on PP–cellulose composites was investigated, and it was found that processing temperature had an influence on mechanical properties and fiber length degradation. Darkening of composites took place, but did not go hand in hand with decreasing properties [25]. Bio-composites using PLA and cellulose in a content up to 40 wt%, as we did in our study too, were mixed in a blender at various temperatures and mixing times, recording the torque. Although the degradation of PLA was observed, the addition of fibers increased melting stability as well as the final mechanical properties. No concerns were raised regarding a possible recycling of these materials [26].

Thermoplastic composite recycling in general poses several challenges, not just the ones with natural constituents [27]. In mechanical recycling of short-fiber thermoplastic composites, the materials are shredded and re-injection molded. However, a performance reduction in those recycled parts can be attributed to a decrease in fiber length, polymer degradation, and degradation of the interface between polymer and fiber [28]. Natural fiber composites do not have an advantage over composites with glass or carbon fibers with regard to fiber degradation during recycling. In addition to the general mechanical fiber breaking, thermal stability may also play a role in natural fibers [29].

The degradation of the polymer matrix during processing is attributed to several factors. Each cycle of reprocessing induces thermo-mechanical and thermo-oxidative stresses to the polymer melt and leads to chain scissions and, hence, lower molecular weight polymers [30]. Recycling PP up to five times was shown to influence viscosity, molecular weight, and the mechanical as well as thermal properties [30,31]. Also, the behavior of pure PLA in mechanical recycling is well investigated. Temperature, mechanical shear, and oxygen lead to degradation during processing [32,33]. Splitting the aliphatic polyester backbone during processing leads to shorter polymer chains accompanied by a decline in properties [11]. Reprocessing pure PLA ten times showed the most effect on impact strength, decreasing it by about 20% throughout recycling, and showed a strong increase in MFR, indicating a chain scission [11,34]. In the case of PLA versus PP, moisture also plays a significant role in the degradation mechanisms (hydrolysis of the ester bonds) [35,36].

Evens et al. investigated PP reinforced with 20 wt% of glass, carbon, and natural fibers subjected to ten recycling steps, where the injection-molded materials were grinded and injection molded again. Mechanical properties were investigated after each cycle and revealed that the natural flax fibers used improved mechanical properties of PP just a little, but over-recycling the properties remained almost constant, whereas glass fiber-reinforced composites showed the least recyclability by reprocessing [37].

PLA-cellulose composites with fiber contents of between 20 and 50 wt% were recycled via injection-molding them for six rounds by Åkesson et al. [38]. Whereas the materials with the lower fiber share exhibited acceptable degradation in mechanical properties (tensile strength decreased about a third from cycle 1 to cycle 6), the compounds with higher fiber loading drastically lost not only mechanical strength, but also a shift in Tg exhibited a degradation of the polymer.

Regenerated cellulose fibers of various types were used to reinforce PLA composites with a fiber content of 30 wt% by Graupner et al. [39]. Multiple processing of the composites revealed that trifold processing does not only damage fiber length, but also the matrix. The researchers also stated that higher moisture in the fibers impact the PLA matrix and hence the mechanical properties [39].

Bio-composites with a PP or PLA matrix containing short cellulose fibers are certainly a field of research attracting a lot of interest, just as recycling of composites is increasingly becoming an issue. Our interest in this research arose from the combination of both preparing bio-composites with varying amounts of cellulose fibers and additionally recycling them three times to describe their degradation behavior. Therefore, the aim of this work was to investigate the reinforcement effects of cellulose fibers in PP and PLA and evaluate their performance in up to three reprocessing cycles.

2. Materials and Methods

2.1. Materials

We used polypropylene (PP) HE125MO from Borealis, Vienna, Austria, a homopolymer intended for injection molding, with an MFR of 12 g/10 min (230 °C/2.16 kg) according to the data sheet, as polylactic acid (PLA) IngeoTM Biopolymer 4032D from NatureWorks, Plymouth, MA, USA, was used. The MFR of this film-grade PLA was determined by measurements to be 6 g/10 min (190 °C/2.16 kg). The natural fibers we used were cellulose fibers Arbocel^® Grade PWC 500 purchased from J. Rettenmaier & Söhne GmbH+Co KG, Rosenberg, Germany, with a cellulose content of about 90% and an average fiber length of 500 µm, an average fiber thickness of 35 µm, and a bulk density of 70–100 g/L, according to the data sheet. For the compounds with PP as the matrix material, a coupling agent (CA) grade Scona TPPP 8112 FA from Byk, Wesel, Germany, was used in a quantity of 10 wt% of the natural fiber content, subtracted from the amount of PP. Due to the potential interactions of the PLA with the hydroxyl groups of the cellulose and the therefore assumed compatibility, no compatibilizer was used. All of the materials were used as received.

2.2. Processing

Before compounding, the polymers were dried in a dry air-dryer, Luxor 120 from Motan Group, Konstanz, Germany, at 90 °C for PP (1 h) and 60 °C for PLA (4 h), respectively. The compounds were produced in a co-rotating twin-screw extruder (TSE 24 MC, ThermoFisher Scientific, Waltham, MA, USA) with a throughput of 10 kg/h. The polymer granules, as well as the coupling agent for the PP compounds as dry blends with PP, were added at the very beginning of the extruder. The fibers were dosed to the molten polymer with a twin-screw side-feeder in zone 8 (residual 8 L/D from the 40 L/D overall length) to the respective percentage. The temperatures for the PP–matrix compounds were set to 180–200 °C from the intake to the die, and to 180–210 °C for the PLA–matrix compounds. The compounds were produced as strands with a three-hole die, cooled in a water bath, and chopped to granules of about 3 mm in length. The granules containing PLA as a matrix polymer were directly dried and crystallized after being chopped to strands at 80 °C in a dry air-dryer.

Universal test specimens (specimen geometry 1A according to ISO-527 [40]) were produced from the dried granules (90 °C for a minimum of 2 h (PP compounds) or at 60 °C for a minimum of four hours (PLA compounds)) with a Victory 80, Engel, Schwertberg, Austria injection molding machine, all compounds of both matrix materials with a mass temperature of about 200 °C.

The sprues, runners, and test specimen not used for the mechanical tests were shredded with a cutting mill (MAS1, Wittmann, Vienna, Austria) using a 5 mm screen. These regrinds served as input material for the following reprocessing cycles, as shown in the processing scheme in Figure 1. The reprocessing of the compounds was conducted in the same way, whereby all of the material was filled at the very front to the extruder. Overall, the whole process of compounding, injection-molding, and subsequent shredding was performed three times, ultimately resulting in four materials per compound. An overview of the materials produced for this study is given in

Table 1. Materials overview: Four compounds with an intended share of 0, 10, 30, and 40 wt% of cellulose fibers were produced with both matrix materials, PP and PLA. The compounds were reprocessed three times each, resulting in a total amount of 24 materials tested. “0” in the column reprocessing means the initial processing established the compound.

Matrix (wt%)		Coupling Agent (wt%)	Cellulose Fiber (wt%)	Reprocessing
PP	100	0	0	0	1	2	3
	89	1	10	0	1	2	3
	67	3	30	0	1	2	3
	56	4	40	0	1	2	3
PLA	100	-	0	0	1	2	3
	90	-	10	0	1	2	3
	70	-	30	0	1	2	3
	60	-	40	0	1	2	3

.

2.3. Materials Testing

As gravimetric dosing of a fluffy lightweight fiber with a high bulk density is challenging, we determined the actual fiber content afterwards via calculating it from the estimated fiber density. As we know the fiber consists of more than 90% of cellulose and is very lightweight and hollow, we assumed the fiber was completely compacted during injection molding and set the fiber density to 1.5 g/cm³, i.e., slightly below the density of pure cellulose, which is 1.59 g/cm³ [41]. The calculations of the fiber volume share were conducted according to Equation (1).

$$V_f={\displaystyle \frac{{\rho }_c-{\rho }_{\mathrm{m}}}{{\rho }_{\mathrm{f}}-{\rho }_{\mathrm{m}}}}$$

(1)

where V_f is the fiber volume share in vol%, ρ_c is the density of the composite (measured), ρ_m is the density of the matrix (measured), and ρ_f is the density of the fiber (estimated).

Density was measured using pieces of the universal test specimen by weighing them in air and immersed in ethanol using a density kit YDK01 on an AX224 balance from Sartorius, Germany, with five replicates per sample according to ISO-1183 [42].

Before mechanical testing, the specimens were stored in a climate chamber (KBF 720, Binder, Tuttlingen, Germany) at standard climate (23 °C, 50% r.h.) for at least 24 h to allow for acclimation.

Tensile tests were carried out according to ISO-527 [40] using five replicates on a 20 kN universal test machine Z020 from Zwick–Roell, Germany. The testing speed was set to 1 mm/min for the determination of the E-modulus in the first 0.25% of elongation in the test specimen; then, the speed was enhanced to 5 mm/min until specimen failure. The elastic modulus was determined to be between 0.05 and 0.25% of the strain by approximating the slope (secant modulus).

Notched Charpy impact strength was tested according to ISO-179 [43] on a pendulum impact tester (5113.300, Zwick–Roell, Ulm, Germany) with ten replicates. Test specimens with a size of 80 × 10 × 4 mm were used, cut out from a universal test specimen. The V-notch was cut on a precision notch saw (Mutronic Diadisc 4200, Rieden am Forggensee, Germany) as type A with a radius of 0.25 mm, an angle of 45 ± 1°, and a depth of 2 mm according to ISO-179.

The melt volume rate was determined using a test specimen milled on the cutting mill to less than 5 mm in size on a flow test device (Aflow BMF-005 for PLA matrix compounds, and 4106 for PP matrix compounds, both Zwick-Roell, Ulm, Germany). The materials were dried in a vacuum oven (VD 115, Binder GmbH, Tuttlingen, Germany) for at least 4 h with a temperature of 90 °C (PP compounds) or 60 °C (PLA compounds) before measurements were performed in double determination. The measurement settings were 230 °C and a piston weight of 2.16 kg for the PP and PP compounds, and 190 °C and a piston weight of 2.16 kg for the PLA and PLA compounds.

Oxidation induction times of PP and PP compounds were measured in accordance with ISO-11357 [44] on a DSC 3 (Mettler–Toledo GmbH, Columbus, OH, USA). Approximately 10 mg of samples were weighed in a 40 µL Al-Standard crucible and heated from 50 to 190 °C with a heating rate of 20 K/min and a purge flow of 50.0 mL/min N2, keeping 190 °C for 3 min. The measurements were started with the switch to 50.0 mL/min O2 purge flow.

Intrinsic viscosity (IV) of PLA and PLA compounds was measured according to ISO 1628-6 [45] with a 0c capillary using chloroform (>99% stabilized, GPR RECTAPUR^®, VWR, Radnor, PA, USA) as solvent. The fibers of the PLA–cellulose compounds were filtered before measurements through a filter paper (Whatman size X).

Color change in the composites over the various cycles was determined by using a color- and gloss-meter Spectro-guide sphere gloss from Byk, Wesel, Germany. Two test specimens were used, and five measurements were taken of each, which totals ten individual measurements.

Microscope images were taken with a stere microscope Zeiss Stemi 2000-C, Zeiss AG, Oberkochen, Germany. Scanning electron microscopy (SEM) pictures were taken on the gold-sputtered fracture surface of a cryo-broken test specimen with an MIRA3 from TESCAN, Brno, Czech Republic.

3. Results and Discussion

3.1. Reinforcing Properties of Cellulose Fibers

The density of pure PP is 0.903 g/cm³ lower than the density of PLA, which is 1.247 g/cm³. Adding 10 wt% fibers to the matrix leads in a higher fiber volume share in PLA than PP. The calculated fiber volume shares are given in

Table 2. Gravimetrically dosed quantities of cellulose fibers to PP and PLA in extrusion vs. calculated quantities in percent by volume.

Matrix Material	Cellulose Fiber Share
	Dosed (wt%)	Calculated (vol%)
PP	0	0
	10	6.53
	30	19.43
	40	26.80
PLA	0	0
	10	12.25
	30	30.04
	40	35.97

.

Adding cellulose fibers to PP– and PLA–matrix compounds gives an expected reinforcing effect. Although the fibers used are a rather short grade, with an average length of 0.5 mm, the E-modulus and tensile strength is increased. The enhancement of the E-modulus is almost linearly for both matrix materials over the entire fiber volume range (Figure 2a). Pure PLA has almost twice the E-modulus and tensile strength than pure PP, but the E-modulus is increased by the addition of cellulose fibers for both, with almost the same magnitude. The addition of fiber shares below 10 vol% in PP and, respectively, 15 vol% in PLA does not increase tensile strength (Figure 2b), but higher fiber shares do. However, the increase in tensile strength is steeper for the PP–matrix compounds, starting at a lower level. It can be concluded that the cellulose fibers used have a higher reinforcing effect on PP in terms of increasing strength than on PLA, and higher fiber shares in PP can reach a tensile strength of the lower filled PLA–cellulose compounds.

PP is a less-brittle polymer than PLA, and therefore shows a much higher strain at break than unfilled PLA, as depicted in Figure 3a. The addition of cellulose fibers reduces the strain at break for both matrix materials, as the fibers act like defects in the polymeric material and prevent the polymer chains from stretching. Notched Charpy impact strength (Figure 3b) shows another trend: starting from a general low level, the addition of fibers increases the impact strength slightly, as while the impact happens, a higher number of fibers can distribute some energy throughout the matrix, therefore consuming more energy while fracturing than the unreinforced polymers, although the effects here are rather low.

Attractive properties can be achieved after the first processing and initial production of PP– and PLA–cellulose composites. The PLA-based materials reach higher values for the E-modulus and tensile strength than the PP–cellulose composites, as well as lower breaking strain and similar behavior in notched Charpy impact tests. How these two materials would behave in a simulated recycling process is investigated in the next section.

3.2. Multiple Processing of Fiber Composites

To simulate several rounds of recycling, the compound materials were reprocessed (shredded—compounded—injection molded) three times. The mechanical properties after each of the cycles were examined and compared. To keep it simple for the reader, the intended and dosed fiber quantity is used in the following paragraphs.

The first indication of the degradation of the compounds was given by the observed color change after each reprocessing cycle; they all became darker (Figure 4). The change in color of PLA without fibers is the most striking, becoming a brownish color after a clear transparent start. PP also became darker, albeit not as strikingly as PLA. All fiber containing materials became darker with the proceeding processing cycles.

We quantified the darkening by color and gloss measurements. The lightness in absolute values is given in Figure 5. As is visible in the image (Figure 4), pure PP only shows a slight decrease in lightness, and pure PLA becomes significantly darker over reprocessing cycles. The trends in the compounds with cellulose fibers are in reverse. Whereas the compounds with the PP matrix lose more lightness over processing than pure PP, the compounds with PLA darken by a similar value. This already indicates that both the matrix and fibers suffer from multiple processing.

The visual degradation is clearly evident, and its effects on the preservation of mechanical properties were analyzed. Figure 6 shows the E-modulus of PP–matrix and PLA–matrix fiber composites produced and over three reprocessing rounds. The empty symbol shows the pure polymer, and for both types of matrices, the E-modulus remains basically constant over the processing cycles. A similar effect is seen for the low-filled 10 wt% compounds, and in the case of PP compounds, and also for the 30 wt% materials. A significant decrease from cycle to cycle in the E-modulus can be seen in 40 wt% PP composites, an effect that occurs with the PLA materials already from 30 wt% onwards. There is even an obvious reduction in the E-modulus; the effect is low in absolute figures. The drop in the E-modulus for the 40 wt% PP compound is from 4588 ± 38 to 4144 ± 23 MPa, and for the respective PLA compound, it is from 7549 ± 17 to 6676 ± 47, meaning that for both materials, almost 90% (90.3% for PP with 40 wt% fibers and 88.4% for PLA with 40 wt% fibers) of the value is preserved, even after reprocessing the materials three times.

Tensile strength for PP–matrix compounds (Figure 7a) shows the same trend as for the E-modulus: higher fiber shares lead to a greater drop in properties, although the compounds with between 30 and 40 wt% fiber still exhibit higher strength than unfilled PP. The loss of strength is almost linear, as seen across the rounds. Pure PLA also preserved its strength across all of the recycling rounds, and PLA with 10 wt% fibers is halfway there, too (Figure 7b). The situation is different for PLA–matrix compounds with fiber shares of 30 and 40 wt%, respectively. PLA with 30 wt% fibers shows an almost linear trend downwards, while PLA with 40 wt% fibers retains the strength in the first round of reprocessing, but shows a huge drop from round 2 on to half of the starting value after reprocessing cycle 3, indicating a degradation in the matrix, which will further be discussed later in this paper. From reprocessing round 2 onwards, none of the fiber-reinforced materials exceed the tensile strength of pure PLA.

As already seen in round 0, the strain at break for PP–matrix composites exhibits a severe drop due to the addition of fibers to the pure polymer. The fiber composites showed a slight trend in increasing strain at break, while reprocessing can be seen in Figure 8a followed by a drop in round 3. We assume that fiber agglomerates are broken with ongoing processing, resulting in a better wetting. The strain at break (Figure 8b) starting from a generally lower level of PLA and PLA compounds rather remains constant and declines after 2 or 3 reprocessing cycles.

Notched Charpy impact strength measurements show a similar trend for the compounds with both matrix materials (Figure 9). Whereas the pure polymers retain their impact properties after recycling, adding fibers to the matrix leads to a steeper decline, but it was also partly started from a higher level. This is due to the reduction in fiber length. In the beginning, the fibers increased the notched impact strength, and when the fiber length was reduced, the reinforcement effect also was reduced, which is shown by the overall trend of an increasing number of recycling cycles.

Investigations of the compounds after three times of reprocessing reveal differences in the materials with no or low fiber content and the compounds with high fiber shares. From that we conclude that an influence of the amount of fiber is evident, and the fibers seem to have an effect on the deterioration of the mechanical properties. Although some trends are similar for both types of materials, independent of the matrix, some differences, especially in the dimensions of the tendencies, are clear, indicating an influence of the fibers on the different types of polymer matrices.

3.3. Influence of the Fibers and the Matrix on the Property Development in Reprocessing

Breaking down the influences on the fibers and the matrices emphasize the differences between the olefin PP and polyester PLA. Measurements of the melt volume rate (MVR) show significant differences between the compounds with PP and PLA (Figure 10). MVR increases for all of the PP-based materials, least of all for pure PP. This trend is quite linear, whereas MVRs for PLA–cellulose compounds tend to increase exponentially. Pure PLA exhibits a linear increase, just as PP and its compounds do. Nonetheless, for all of the materials filled with fibers, changes in the MFR are observed in the proceeding processing steps. The exponential increase in PLA-cellulose compounds indicate that the addition of fibers led to a degradation in the matrix material. Since PLA as polyester is prone to hydrolytic cleavage in the polymer chains, the fibers may introduce functional groups catalyzing the breakdown of the material. This degradation was also shown above in the mechanical properties, but not as strongly as for MVR, as the fibers still provide good reinforcement, even after three reprocessing cycles.

Another proof of matrix degradation can be seen in the measurements of oxidation induction time (OIT) in PP–matrix compounds and intrinsic viscosity (IV) for PLA-based materials. OIT describes the stability of the material against oxidation. Figure 11a shows that the most significant reduction is seen for pure PP, where OIT is decreased from 7.01 to 3.04 min after three reprocessing cycles. The PP materials containing fibers exhibit a lower OIT in the first place and a reduction after reprocessing is lower. As we reduce the share of polymers with the addition of cellulose fibers, we assume the higher oxidation stability of the cellulose to result in higher OIT values. IV measurements for PLA and PLA compounds (Figure 11b) show another trend and prove the abovementioned presumption of cellulose fibers accelerating the degradation of PLA. The more cellulose fibers are added to the matrix, the stronger is the degradation of the polymer.

The fact that not only the matrix, but also the fibers suffer from degradation is shown by microscopic images taken from extracted fibers (Figure 12). After the actual production of the compounds, the fibers retain their initial fiber length of 500 µm according to the data sheet to a large extent. After each reprocessing cycle, the fibers reduce in length, as is obvious from the images.

Another major influence on the mechanical properties is also the interfacial properties between the fiber and matrix. We examined the wetting of the fibers with SEM (Figure 13). Individual fibers can be recognized as being effectively wetted by the matrix polymer. We did not identify a change in wetting behavior over the reprocessing cycles from examinations of the fractured surface.

The interfacial shear strength (IFSS) τ_c can be calculated for composites with subcritical fibers using the modified Kelly–Tyson equation (Equation (2)) [46,47]:

$${\sigma }_c={\eta }_O{\tau }_c{\displaystyle \frac{l}{d}}{V}_f+{{\sigma }^{\prime }}_m(1-V_f)$$

(2)

σ_c is the strength of the composite, which we determined by tensile tests, as well as σ’_m, the reduced tensile strength of the matrix at the tensile strain of the composites. The fiber orientation factor η_O was set to 0.75 [47,48]. l is the length of the fibers, which initially was 500 µm, and d is the fiber diameter, being 35 µm according to the fibers’ data sheet. V_f is the volume fraction of the fiber in the respective composite, which we determined using calculation in Equation (2).

Although we have examined the cellulose fibers extracted from the matrix with a microscope (Figure 11), it is difficult to determine the exact length due to the nature of the fibers. They are not as straight as glass or carbon fibers, where the determination of the fiber’s length is rather easy, but are tangled and twisted. Nonetheless, a decrease in fiber length is clearly obvious. For our calculations, we do not aim for exact values, but rather aim for comparisons to assess degradation over multiple processing. For a first estimation of τ_c, we assume a fiber length of 500 µm after the initial compound preparation. As we see from Figure 14, IFSS seems to increase with increasing cellulose content if we calculate it for all of the composites with the initial fiber length. From our point of view, IFSS should not change with increasing fiber fraction, as it describes the interaction at the interface between the fiber and the matrix, and is, therefore, mostly independent of the fiber and matrix fraction. An IFSS of about 5 MPa for PP with cellulose fibers is in good accordance with the literature, where an IFSS of about 6 MPa was reported for sisal PP composites [48,49]. The IFSS of PLA compounds is lower than for PP compounds, indicating a poorer matrix–fiber interaction.

We therefore set τ_c to a constant value, calculated the fiber’s length, and standardized these values to the compounds of round 0. In this way, we received a good indication of how fiber length decreases over reprocessing and also of the correlations between fiber volume share and fiber length. Under the assumption that τ_c is constant, Figure 15 reveals that the degradation in fibers takes place more rapidly for lower than higher fiber shares, most likely due to the higher fiber length reduction in the initial fiber length at higher fiber shares in the composites, applying to both matrix materials over the reprocessing cycles. For PLA with a low fiber content, we see less degradation; the nominal increase in the fiber length can be attributed to some scattering in the evaluation.

Although these calculations are based on a set of assumptions, owing to the nature of the type of fibers used, they show a clear trend in decreasing fiber length over processing cycles and support our thesis that a large part of performance degradation is due to fiber degradation.

4. Conclusions

In our work, we investigated the reinforcement effects of cellulose fibers in PP and PLA. Both matrix materials can efficiently be reinforced with short cellulosic fibers, and mechanical properties such as E-modulus and tensile strength increased with an increase in fiber share.

Reprocessing the compounds trifold shows that although the materials show signs of degradation, the mechanical properties could be preserved to a large extent, especially for lower fiber contents. Nevertheless, reprocessing the compounds three times led to a progressive degradation in fiber length. A decrease in fiber length was also proofed by calculating the standardized fiber length degradation with the modified Kelly–Tyson equation for subcritical fibers. Measurements of the intrinsic viscosity of PLA and oxidation induction time for PP revealed that especially PLA, which is sensitive to hydrolytic chain scission, suffers during reprocessing.

The results of this study reveal that the mechanical recycling of short fiber-reinforced bio-composites with PP and PLA as matrix materials is possible, although fiber length is clearly decreased, but mechanical properties are widely preserved. Future studies should investigate how additivation can preserve fiber length and properties for improved recyclability.