Effect of multiple recycling on thermo-mechanical and rheological behaviour of PP/EPDM thermoplastic vulcanizates

Driven by the need to design sustainable polymeric materials that remain superior mechanical properties after recycling, this work is centred on the effect of multiple recycling of thermoplastic vulcanizates (TPVs). Among thermoplastic elastomers, TPVs combine several favourable characteristics such as damping, light weight, ease of processing by means of injection moulding, design �exibility and recyclability. Multiple processing of a commercially available PP/EPDM TPV by shredding and injection moulding was followed by analytical investigations on rheological and thermo-mechanical properties using melt rheological measurements, dynamic mechanical analysis, differential scanning calorimetry analysis and mechanical tests. The results show that key performance parameters of the TPV such as Young’s modulus, stress at 100% strain as well as stress and strain at break only change slightly. Stress at 100% strain can be used as a quality indicator as it decreases linearly with every recycling step. This study opens new opportunities to increase the content of recycled PP/EPDM TPV and even manufacture technical components with 100% recycled material.


Introduction
In recent years, the use of recycled materials has become one of the main socio-economical concerns.In the eld of recyclable elastomeric materials, thermoplastic vulcanizates (TPVs) are attracting a growing interest due to their light weight, ease and cost-effective processing and design exibility [1] [2] [3].The concept of TPV was rst proposed by Gessler and Haslett in 1962 [4].TPVs are obtained by mechanical mixing of rubbers and semi-crystalline thermoplastics followed by dynamic vulcanization under shear loading to its characteristic morphology [5] [6].This effect is known as phase inversion where selectively crosslinked rubber gets encapsulated by the thermoplastic due to the increase of the rubber phase viscosity.Phase inversion leads to a ne dispersion of rubber particles in the continuous semi-crystalline thermoplastic matrix giving TPVs unique material properties compared to conventional thermoplastic elastomers (TPEs) [7] [8], such as oil resistance, fatigue resistance and heat deformation among others [9].
Most of the commercial TPVs are based on synthetic rubbers such as ethylene-propylene-diene monomer (EPDM), ethylene propylene rubber (EPR), and butadiene acrylonitrile rubber (NBR).The thermoplastic matrix of polyole n TPVs mostly consists of polypropylene (PP) or polyethylene (PE).Most non-polyole n TPVs are based on polyamide (PA).Apart from their commercial importance for various applications PP/EPDM has unique physical and mechanical properties due to its phase compatibility [10] [11].By dynamic vulcanization of PP/EPDM tensile stress and elongation at break are signi cantly increased compared to conventional TPEs [12].Dynamic vulcanization of PP with EPDM signi cantly reduces PPs low-temperature brittleness opening several industrial applications for this general purpose polymer.From a sustainability point of view, PP is favored to its bene cial carbon footprint compared to other thermoplastics like PA. PP and EPDM both can be synthesized from biobased recourses.
There is a wide interest in recyclability of PP and much scienti c work has been published.Multiple authors have investigated rheological properties during multiple processing [13] [14], as well as thermo-mechanical parameter changes [15].[16] [17].Work on recyclability of TPVs has been published [18] [19] [20], some authors concentrating on the use of ground tire rubber in PP/EPDM [21] [22].Few research is on recycling of PP/EPDM, but promising results have been found by Wang et al. [23] who found that repetitive processing did not cause much loss in the mechanical properties of EPDM/PP TPVs if βnucleating agent has been added.
The aim of the present work is to obtain a more fundamental understanding of the effect of recycling on the thermo-mechanical and rheological properties of PP/EPDM TPVs produced by injection molding.The authors want to gure out to what maximum amount recycled TPV can be added to virgin granulate and if even multiple recycling is possible.Therefore it is useful to nd out which thermo-mechanical parameters change the most and which parameter can be used as a quality indicator for the material's performance.This will help to ensure a competitive and con dent design and application of recycled TPV materials.

Materials
A commercial PP/EPDM TPV from Celanese Corporation (Dallas, Texas) has been used.It is a thermoplastic vulcanizate called Santoprene™ 103 − 40.The polymer has a Shore D hardness of 41 and a tensile stress at 100% elongation of 9 MPa (perpendicular to ow, orientation 90°).The manufacturer suggests to add a maximum regrind of 20% to the original material.Selected material properties from the manufacturer's data sheet are presented in Table 1.2.2 Methods

Scanning electron microscopy (SEM)
A morphological study of the samples was carried out using a eld-emission scanning electron microscope.The samples were brittle fractured in liquid nitrogen prior to SEM observation.The cryofractured surfaces were sputter coated with a gold layer prior to microscopy.

Injection moulding of samples
TPV plates of 90 mm x 90 mm x 2 mm were fabricated by injection moulding on a DEMAG IntElect 100/470 − 340 (Sumitomo Demag Plastics Machinery GmbH, Shinagawa, Tokyo) at Leartiker.Material manufacturer's recommendation for injection speed is "fast".Injection speed has been set to 120 cm³/s according to an injection pressure around 400 bar.The used injection speed is higher than the recommended one to ensure maximum stress by multiple processing of the TPV.The processing values are compared with the material's manufacturer in Table 2.

Oscillatory rheological characterization
Oscillatory rheological analyses were carried out using a Thermo Haake Mars III oscillatory rheometer (Thermo Fisher Scienti c Inc.).Characterization of the samples was carried out at 200°C with 20 mm parallel plates.Distance between the plates was set to 0.5 mm.
Tests were carried out in both strain and frequency sweeps.Strain sweeps were performed from γ = 0.1 to 100% at a constant frequency of f = 1 Hz (2π rad s − 1 ).Frequency sweeps were performed from f = 0.1 to 100 Hz at a constant strain of γ = 0.1%.

Quasi-static mechanical behaviour
The quasi-static mechanical properties of PP/EPDM TPV samples, die cut into ISO 37 F type 2 dumbbellshaped bars (75 mm × 4 mm × 2 mm) following transverse (90° orientation) and longitudinal (0°o rientation) directions with respect to the ow (Fig. 1), where measured by performing monotonic and cyclic tensile tests.The position on the bars in the plate has been speci ed by four codes (recycling step, processing, panel-#, position on panel).Monotonic tensile test properties were measured on an Instron test machine at RWU by analysing crosshead traverse path at 5 mm min − 1 speed according to DIN EN ISO 527-1.Cyclic tensile tests were measured on a MTS Insight 100 kN test machine by videoextensometer at room temperature performed at a strain rate of 100% min − 1 .For the cyclic tensile testing, six strain levels were applied (10%, 20%, 30%, 50%, 70% and 100%) and ve cycles of loading-unloading were applied for each level of strain.A new loading stroke was started once a zero stress state was obtained in the tensile sample.The residual strain at zero tensile stress was taken as the permanent deformation of the samples.The hysteresis loss of the samples was measured by estimating the ratio between the areas of loading-unloading cycles and the areas below the loading curves.Quasi-static mechanical behaviour reported in this work is measured at room temperature.Measurement data are not averaged unless otherwise speci ed.The values reported are the engineering stress and strain.

Dynamic mechanical behaviour
The dynamic mechanical behaviour of the PP/EPDM TPV samples cut into ISO 37 F type 2 dumbbellshaped bars upon following transverse and longitudinal directions were evaluated on a Metravib DMA + 300 in tension mode.The temperature sweep tests were performed by varying the temperature from − 70 ºC to 80 ºC at a heating rate of 3°C min − 1 with an oscillating frequency of 10 Hz.A small dynamic γ displacement of 5 µm was applied in all the tests to be inside the linear viscoelastic region.A static force of 5 N was applied to prevent sample buckling.

Morphological investigation
The cryogenically fractured surfaces of PP/EPDM TPVs are analyzed by SEM as shown in Fig. 2. As explained in [24] a coarse fracture surface with incorporation of wrinkles indicates ductile fracture which implies good interfacial adhesion between PP and EPDM.This is the case for samples after 1st injection molding.After 10 times recycling the fracture surface is still coarse, but areas of smooth fractures grow.
In the SEM micrograph shown in Fig. 3 the areas surrounded by dark gray lines correspond to the dispersed EPDM particles embedded into the Polymer, e.g.PP.EPDM enclosures have a tendency to orient along ow direction, agglomerations of EPDM rubber particles are not visible.Both play an important role in the nal mechanical properties.

DSC analysis
To assess changes in the crystallinity of the PP due to multiple recycling steps, DSC measurements were performed.Table 3 shows the melting temperature (T m ), the crystallization temperature (T c ) as well as the normalized enthalpy of melting and cooling per gram of sample (ΔH m ) for the as-received pellets and samples cut from the recycled PP/EPDM TPV plates.Second heating and rst cooling results imply that all recycling steps produced by shredding and subsequent injection molding demonstrate about the same melting and crystallization temperature as their corresponding pellets, according to the results of [25].Melting and cooling enthalpy ΔH m and ΔH c also change marginally, so we can conclude that the lamellae crystal thickness of PP are similar to those in the virgin (0x recycling) PP/EPDM TPV [26].The rheological properties of polymers strongly depend on their morphology and in the case of polymer blends on the interaction between the components [27].All blends show typical thermoplastic non-Newtonian shear thinning behaviour, simplifying material ow and reducing heat generation during processing.Signi cant reduction of viscosity due to recycling would lead to pressure drop and tendency for material ash during injection moulding.Changes in the EPDM will especially change viscosity behaviour in this shear thinning region ( > 100 s − 1 ).A reduction of molecular weight of PP matrix due to recycling should preferably be measured in the Newtonian low shear region ( < 10 s −1 ) where viscosity is constant [28].This will be investigated by dynamic rheometer measurements.
Shear thinning behaviour is typical for injection moulding.The viscosity can be approximated by a simple Power-Law model according to Ostwald. ) and data given in Brydson's Plastics Materials book [29].The values are also typical for an un lled PP in the shear thinning region.The reduction of viscosity due to recycling is only marginal, but can be seen in a slight drop of the exponent from − 0.732 to -0.740.A similar decrease of viscosity in the shear thinning region due to recycling is reported in the literature for un lled PP [14].The loss factor (tan δ), de ned as the ratio of the loss modulus to the storage modulus, clearly shows strong in uence of recycling, especially at low frequencies f < 1 Hz (Fig. 8).The effect is already present after the rst injection molding and more pronounced than the drop of viscosity or shear modulus.
3.4 Quasi-static mechanical behaviour 3.4.1 Monotonic stress-strain behavior: orientation effect The mechanical behavior of TPVs is described by their stress-strain response.Figure 9 presents the longitudinal and transverse monotonic stress-strain curves of TPV samples with different orientation.
The longitudinal samples were located in the middle and side of the injection molded plaques, the transversal ones were located near the gate and at the end of the lling path.
When comparing the stress-strain curves of the PP/EPDM TPV orientations, it can be observed that the stress-strain behavior can be divided in three phases: 1st phase with ε = 0% -10% with elastic-like behavior and increasing stress up to ~ 8 MPa; 2nd phase at ε = 10% -200% with a distinct stress plateau ~ 8 MPa; 3rd phase for ε > 200% with increasing stress until rupture.The three phases can be characterized by the elastic modulus, stress at 100% strain and elongation at break.
Orientation on panel nearly has no in uence on the 1st phase of elastic-like behavior.The changes are at the 2nd phase: The stress at 100% strain and stress plateau is signi cantly higher for samples oriented parallel to ow.There is also a signi cant in uence of the sample orientation regarding elongation and stress at break.Perpendicular orientation has ~ 100 MPa higher strain values at break compared to longitudinal ones.
Position on the panel has only little in uence on the stress-strain behavior.The stress plateau is ~ 5% higher for longitudinal samples when they are oriented in the middle (Pos.C).The transversal samples show no signi cant differences according to the distance to gate.Regarding elongation at break, the tests reveal no signi cant in uence of the position on panel, independently of orientation.
This demonstrates the in uence of the microstructural orientation.Beside the polypropylene crystals and chain orientation, the alignment of the EPDM particles within the PP matrix is important as observed by TEM measurements [30] [31].During injection molding EPDM particles are stretched to ellipsoids which has a signi cant impact on the stress-strain properties.As EPDM has a higher Young's modulus than PP, tensile strain along the longitudinal axis of the ellipsoidal EPDM results in higher stresses and reduced elongation at break; correspondingly strains perpendicular to the longitudinal axes result in lower stresses but higher elongation at break.This observation is comparable to glass ber reinforced plastics -the effect of EPDM orientation goes along with ber orientation.

Monotonic stress-strain behavior: effect of multiple recycling on orientation effect
Monotonic stress-strain behavior has been analyzed after selected recycling steps.Stress level at 100% strain and strain at break have been picked out to represent changes due to recycling.After multiple recycling, the effect of sample orientation on stress-strain behavior is still strong as illustrated in Fig. 10 after 10th injection molding (9x recycling).Tensile test specimen oriented parallel to ow still have a signi cantly higher stress level at 100% strain compared to transversely oriented ones and strain at break is reduced.
3.4.4Monotonic stress-strain behavior: effect of multiple recycling on stress at 100% strain Figure 11 shows stress-strain dependence for different recycling steps for positions B and E on the panel, both longitudinally oriented.Changes due to recycling are small, especially stress and strain at break are not signi cantly changing by recycling.Stress at 100% strain is slightly decreasing due to recycling.
Transversally orientated specimen also do not show a strong in uence of recycling on stress and strain at break, shown for positions C and D (Fig. 12).But a tendency for higher stresses and strains due to recycling can be observed.The reason for the bigger in uence of recycling on strain at break can be found in the orientation of the EPDM droplets.They reduce strain at break for longitudinal orientation, but not for transversal orientation.Therefore changes in the PP matrix due to recycling can be seen more clearly when EPDM droplets are oriented transverse to strain direction.Stress at 100% strain is again slightly decreasing due to recycling, as already observed for longitudinal orientation.
Having a closer look on stress at 100% strain, an interesting effect of recycling can be observed (Fig. 13).
Stress plateau values decrease linearly by ~ 50 kPa (~ 0.5% of 10 MPa) with every recycling step.Stress levels at 100% strain are ~ 1.5 MPa lower for transversal orientation, but stress decreases by recycling with the same rate.This decrease of stress at low strain levels of ~ 100% implies again that the recycling rst of all changes the PP matrix.

Cyclic stress-strain behavior
When trying to replace traditional thermoset rubbers with TPVs, one of the required key properties is the elastic recovery ability.The elastic recovery ability is a key property for anti-vibration products, for which a value as high as possible is sought.The elastic recovery ability can be characterized by measuring permanent deformation in cyclic stress-strain tests.Low permanent deformation correspond to a good elastic recovery ability [32].
The cyclic stress-strain level of the longitudinal PP/EPDM TPV is increased sequentially from 10%, 20%, 30%, 50%, 70% and 100% as shown in Fig. 14 (a).A stabilized state, characterized by a constant stress amplitude and constant hysteresis loop, was achieved after four cycles.As a result, the fth cycle was considered as the stationary state.As shown in Fig. 14 (b), the PP/EPDM TPV shows a large hysteresis loss in the rst loading-unloading cycle, together with a pronounced loss of hysteresis when moving from the rst to the fth stabilized cycle, suggesting a signi cant Mullins softening effect.The Mullins softening phenomena is schematically shown in Fig. 14 (b) [33].Several physical interpretations exist for the Mullins softening behavior, however a general agreement for the cause of this effect at the microscopic level is still absent.
Figure 15 presents the permanent deformation of the 1st and 5th cyclic deformation after (a) 1st injection molding (0x Rec.),(b) rst recycling step (1x Rec.) and (c) 10th recycling step (10x Rec.).Recycled, as well as original TPVs show a more or less linear increase of permanent deformation at 1st cyclic deformation.
5th cyclic deformation shows a constantly rising increase of permanent deformation with higher strains.
The permanent deformation is slightly higher after 10 times recycling, what may be attributed to the degradation and fracture of the PP due to the recycling.
Figure 16 shows the corresponding increment of permanent deformation due to recycling after (a) 1st cyclic deformation and (b) 5th cyclic deformation.The increments show the differences between the recycled TPV in comparison with the original one.So the increments after ten recycling steps show the summed permanent deformation of 10 recycling steps.The permanent deformation increments are highest around 20% strain.If the strain is further increased, the increment is more or less constant.For the 5th cyclic deformation, most of the permanent deformation growth is due to the 1st recycling.The increase doesn't rise much further by the nine additional recycling steps.Only the 1st cyclic deformation shows additional changes around 20% strain.
Cyclic deformation and microstructural changes of PP/EPDM TPVs are rarely correlated by models in the literature [34] [35] [36] [37].Soliman et al. reported the deformation behavior of a PP/ EPDM TPV compound cured with a phenolic resin [38].They performed combined infrared spectroscopy and tensile stress-strain tests in order to measure the orientation of the rubber and thermoplastic phases during the stretching of the sample.They found that the whole rubber phase was stretched, while only a small portion of the thermoplastic phase was stretched.According to them, the plastic deformation of the thermoplastic phase is concentrated at the boundary between the rubber droplets.During the unloading process the previously deformed thermoplastic portion is pulled back to some extent owing to the elastic recovery capacity of the rubber.A schematic representation of the Soliman model is depicted in Fig. 17.
Based on the Soliman model, the cyclic deformation behavior of TPV samples may be mainly driven by the yielding, buckling and bending of the thin PP matrix ligaments and the elastic recovery ability of the rubber droplets.At low strains up to 10% both the PP matrix phase and the rubber droplets may deform elastically and so the permanent deformation (without recycling) is quite minimal at the initial strain levels.With increasing the strain level, the semicrystalline PP matrix will start to yield in the regions where matrix ligaments are thinnest.Upon unloading, the elastic forces of the stretched interconnected rubber network are able to pull back the plastically deformed thin ligaments by either bending or buckling.The loss of stiffness during the reloading step may be attributed to the damage generated by the bending or buckling of the thin plastic ligaments during the unloading.The thicker thermoplastic ligaments of the matrix will interconnect the rubber droplets, forming an elastic interconnected network structure of rubber droplets and slightly deformed thicker zones of the PP matrix.When the strain level is increased the thin matrix ligaments will continue to yield in order to continue to the elastic deformation of the rubber droplets.
Analyzing the effect of recycling, original PP/EPDM demonstrate lower permanent deformation compared to recycled parts (Fig. 16).Until 10% strain level, the effect is marginal.Original and recycled TPV compounds present similar permanent deformation as long as we stay in the elastic region.PP phase crystallinity degree and crystalline lamellae thickness are important for elastic recovery [39].As found out by DSC analysis, crystallinity is not affected by recycling and so the elastic recovery at small strains more or less stays the same, independently of recycling steps.If the strain is increased to 20 %, Mllins softening occurs and the TPV samples present an increase of the permanent deformation by recycling.1st recycling already weakens this elastic interconnected network structure between rubber droplets and PP matrix.
With subsequent recycling steps this rise slowly goes on.It may be attributed to the damage generated by recycling that affects not only thin PP ligaments, but also thicker ones that interconnect the rubber droplets.These results are consistent with DMA results, that indicate network interactions are stronger in the original TPV compound, which means that they are able to dissipate more strain energy in terms of viscous losses [40] [41] compared to the recycled TPV.
Above 30% strain level, the increment of permanent deformation by recycling goes down and stays more or less constant.No more tremendous weakening of the matrix ligaments seem to occur and recovery ability is less affected by recycling.
In this work, authors propose that worsening of elastic recovery capacity by damaging the PP is the main reason for the early rise of permanent deformation at 20% strain level after 1st recycling.

DMA
Figure 18 presents the dependence of tensile storage modulus E', loss modulus E'' and tan δ on the temperature at a constant frequency of 10 Hz for the 1st injection moulded TPV sample ("0x Rec"), after 1st ("1x Rec") and after 10th recycling ("10x Rec").Immiscible blends show distinct glass transitions for each component.In particular, as shown in Fig. 18 (c) PP/EPDM TPVs show a lower temperature peak of tan δ around − 50°C which can be assigned to the T g of the EPDM rubber phase, whereas the tan δ peak at about 0°C is assigned to the PP amorphous phase T g .Therefore, the DMA results con rm the existence of the two-phase separated structure in the original and recycled TPVs.Furthermore it can be seen in Fig. 18 (a) that the drop of storage modulus due to recycling is only pronounced at temperatures below T g of the PP around 0°C.This correspond to the observation that rheometer investigations only show a marginal drop in viscosity due to recycling.Reduced viscosity would have indicated reduced molecular weight facilitating crystallization.The independence of stiffness of unreinforced PP on recycling at room temperature is also reported in [16].This might also be the reason why elongation at break and yield stress of TPV remain unchanged during to recycling as long as EPDM phase is not damaged.
On the other hand, Fig. 18 (c) shows a decrease of tan δ above Tg of EPDM due to recycling, according to the work of Li et al. [17].The corresponding tan δ peak is shifted towards lower temperatures.As tan δ is associated with damping a reduction of damping properties of the TPV at high temperatures can be assumed.
By characterizing the temperature dependence of loss modulus versus storage modulus it is possible to analyze the dynamic response of viscoelastic materials [42].Figure 19 reveals the in uence of temperature on E´ and E''.For the same elastic modulus the recycled TPVs possess a lower loss modulus (less dissipation in the material).This reduction of the loss modulus or mechanical loss observed by DMA may be ascribed to the fact that recycling is causing degradation and fracture of the PP, lowering the molecular weight.

Conclusions
In this article, the effects of multiple simulated recycling of a commercial TPV compound based on PP and EPDM have been experimentally investigated through DSC, DMA melt-rheological analyses and monotonic as well as cyclic tensile tests.
DSC measurements indicate that the lamellae crystal thickness of the PP matrix is not affected by recycling.Monotonic tensile tests show a linear decrease of stress at 100% strain by 0.75% with each recycling step independently from specimen orientation.Oscillatory rheological measurements also show a decrease of viscosity by recycling.Cyclic tensile tests show an increase of permanent deformation, especially after 1st recycling step.Also DMA analysis at temperatures below T g (PP) con rm this result.
DMA analysis additionally show that damping properties investigated by tan δ are reduced by recycling, if

3 4Figure 5
Figure5shows the Power-Law behaviour of the TPV after rst injection moulding ("Cycle 0"), after 1st ("Cycle 1") and after 10th recycling ("Cycle 10").The values correspond with the measurements of Santoprene™ 103 − 40 given by Autodesk Mold ow (measurement by Beaumont Advanced Processing in 2021) and data given in Brydson's Plastics Materials book[29].The values are also typical for an un lled

2
Oscillatory rheological characterizationOscillatory rheological characterization has been used to analyze low shear region and onset of shear thinning with respect to its correlation with simulated recycling.Polymers with reduced molecular weight M w of the PP matrix tend to thin more at lower shear rates than those with high M w .Oscillatory measurements have been used to characterize frequency dependence of viscosity at constant shear strain γ = 0.1%.Figure6shows the behaviour of viscosity during frequency scan and of shear modulus on strain scan.Both are related by .Mean values have been taken for several steps to increase measurement accuracy.Standard deviation between measurements is around 25%.Repeated injection moulding cycles, especially after 10th recycling step, show a clear tendency for decreasing viscosity and shear modulus values.Figure7shows viscosity at γ = 0.1%, f = 1 Hz at different recycling steps.An obvious drop in viscosity due to recycling can be seen.2nd until 9th recycling step have not been analysed.The authors found decreases of viscosity due to 10 times recycling of nearly 50%, comparable to a study by Spicker at al. [28].One has to keep in mind that oscillatory rheological characterizations shown here are situated in the high shear rate (shear thinning) region.Viscosity decrease is usually more pronounced at low shear rates [28].

Figure 7 Mean
Figure 7

Figure 8 tan d dependency
Figure 8

Figure 11 Stress
Figure 11

Figure 12 Stress
Figure 12

Figure 15 Permanent
Figure 15

Table 1
Selected material properties from the manufacturer's data sheet

Table 2
Recommended processing values compared to values set for investigation

Table 3
Consequently we can conclude that structural changes in the EPDM particles, as they would be seen at high shear rates, might only be marginal.