mising for a decreasing energy consumption. However, the increase in energy efficiency triggers the conflict between rolling resistance, wet grip and wear resistance. [1] This study investigates the influence of various unsaturated and non-polar rubbers on their demolding characteristics. Therefore, C ONTINENTAL provide eight model compounds with different polymers, glass transition temperatures (T g ) and polymer functionalizations. Polymer functionalizations and recipes A radical or an anionic polymerization can functionalize SBR. The radical polymerization produces emulsion SBR (E-SBR) and only every third monomer will be functionalized. The resulting molecular weight distribution is broad. The low molecular weight elements in the molecular weight distribution have a positive effect on Aus Wissenschaft und Forschung 58 Tribologie + Schmierungstechnik · 68. Jahrgang · 3-4/ 2021 DOI 10.24053/ TuS-2021-0021 Introduction Fuel consumption reduction is a development focus in automotive industry. An optimization or exchange of the used material can reduce the total weight of a car and together with further developments as well as adjustments in drive solutions, the energy efficiency can increase significantly. Tires contribute with around ten percent to the total fuel consumption. That’s why functionalized polymers for instance styrene-butadiene (SBR) are very pro- Influence of polymer functionalization on adhesion, friction and mold fouling Stefanie Haupt, Mathias Kröger, Jörn Krüger* Dieser Beitrag wurde im Rahmen der 62. Tribologie-Fachtagung 2021 der Gesellschaft für Tribologie (GfT) eingereicht. Einfluss der Polymerfunktionalisierung auf Adhäsion, Reibung und Formverschmutzung Am IMKF wurden acht verschiedene Elastomermischungen mit differierender Polymerzusammensetzung auf ihre Entformungscharakteristik und die Formverschmutzung untersucht. Für jede Mischung sind zehn Entformungen durchgeführt worden und die Formverschmutzung wurde fototechnisch nach zehn Entformungen festgehalten. Die Polymerzusammensetzung scheint bei den verwendeten Polymeren keinen signifikanten Einfluss zu haben. Jedoch haben Mischungsbestandteile einen deutlichen Einfluss auf die Reibcharakteristik beim Entformen des Vulkanisates. Eine sehr zeitig beginnende Vernetzung hat einen initiierenden Einfluss auf die Formverschmutzung. Die auftretenden messbaren Schwankungen der maximalen Entformungskraft könnten mit der inneren Reibung des Polymers mit dem Füllstoff korrelieren. Schlüsselwörter Demolding-Prüfstand, Entformungskräfte, Polymere, Formverschmutzung, Seitenwandreibung At the IMKF, eight different elastomer compounds with different polymer compositions were examined for their demolding characteristics and mold fouling. Ten demolding cycles were carried out for each compound and the fouling of the mold was photo-technical recorded after ten demolding cycles. The used polymer composition does not seem to have any significant influence. However, the components of the compound have a significant influence on the friction characteristics while removing of the vulcanizate from the mold. Crosslinking that begins very early has an initiating influence on mold fouling. The measurable fluctuations in the maximum demolding force that occur could correlate with the internal friction of the polymer with the filler. Keywords Demolding test rig, demolding force, polymer, mold fouling, side wall friction Kurzfassung Abstract * Prof. Dr. Matthias Kröger Orcid-ID: https: / / orcid.org/ 0000-0002-4132-8323 M.Sc. Stefanie Haupt Technische Universität Bergakademie Freiberg Institut für Maschinenelemente, Konstruktion und Fertigung Agricolastraße 1, 09599 Freiberg Dr. Jörn Krüger Material & Process Development & Industrialization Tires Commercial Tire Technology, Continental Reifen Deutschland GmbH, Jädekamp 30, 30419 Hannover TuS_3_4_2021.qxp_TuS_Muster_2021 03.09.21 13: 28 Seite 58 processing. A negative effect are the energy losses that occur due to the large number of molecular end groups. The existing styrene and vinyl content can vary. In this study, the E-SBR isn’t functionalized but the T g is modified by an increased amount, because of the increased use of RAE oil (Residual aromatic extract) (Table 1). RAE is an oil with a low proportion of polycyclic aromatics. [1, 2] The anionic polymerization produces solution SBR (S-SBR) and almost 100 % of the monomers are converted. The range of styrene and vinyl content is also significantly larger. The molecular weight distribution can be controlled and the chain ends can be functionalized to optimize the polymer-filler-interaction. One or both end groups can be functionalized. [1] Depending on the end groups, they interact with the silica, carbon black or react chemically with each other. Due to the functionalization, the interaction between polymer and filler is strengthened and an improved dispersion of the silica can be measured. The smaller flexibility and larger dispersion lead to an improved balance between wet grip and rolling resistance. The investigations of Y AMADA show that functional ends at both polymer ends increase the wear resistance, too. [1] The individual recipe components are listed in Table 1. The measured T g is noted behind the polymers, but the end groups of the individual functionalization are not known. A special nomenclature (Figure 1) was developed for this study in order to be able to identify quickly, which recipe is involved. The composition of the blends can be read from the nomenclature, because it markers the changing parameters in this study. In order to be able to distinguish the compositions, three indices were integrated. The first index indicates how many parts per hundred natural rubber (N) was used. The second index has a “y” or “RAE”. The “y” stands for that the polymer has been functionalized and the “RAE” is used to differentiate the two compounds used with E-SBR. A third index (“y”) can be noted on the BR. It means, that the BR polymer has been functionalized. In addition to S-SBR and NR 44 - 50 phr BR are added to three compounds. Two of the used BRs (BR y and BR 2 ) differ in their functionalization, but have the same T g (-95 °C). BR 1 has the lowest T g (-105 °C) of the polymers and differs from BR y and BR 2 in terms of its cis, trans and vinyl units (Figure 2). According to the literature, the T g of cis-1,4 and trans-1,4 components are similar and the 1,2-vinyl has a different T g . [3] Aus Wissenschaft und Forschung 59 Tribologie + Schmierungstechnik · 68. Jahrgang · 3-4/ 2021 DOI 10.24053/ TuS-2021-0021 N 20 - SS y 1 - SS y 2 N 10 - SS y 1 - BR y N 10 - SS y 2 N 20 - SS- BR 1 N 10 - SS- BR 2 N 100 N 10 - ES N 10 - ES RAE NR (T g =-64°C) 20 10 10 20 10 100 10 10 S-SBR y 1 (T g =-23°C) 10 40 - - - - - - S-SBR y 2 (T g =-64°C) 70 - 90 - - - - - S-SBR (T g =-23°C) - - - 36 40 - - - E-SBR (T g =-54°C) - - - - - - 90 - E-SBR RAE (T g =-50°C) - - - - - - - 123.8 BR y (Tg=-95°C) - 50 - - - - - - BR 1 (Tg=-105°C) - - - 44 - - - - BR 2 (Tg=-95°C) - - - - 50 - - - Silica 95 95 95 95 95 95 95 95 TDAE 45 45 45 45 45 45 45 11.25 6PPD Ozon wax Stearic acid ZnO Silane DPG CBS S Recipe ingredients in phr standard quantity for SEV standard quantity standard quantity standard quantity standard quantity standard quantity standard accelerator quantities Functionalized Polymer (index: y) yes no Table 1: Compounds of the polymer study Figure 1: Nomenclature of the polymer study TuS_3_4_2021.qxp_TuS_Muster_2021 03.09.21 13: 28 Seite 59 increase the number of test cycles and the tests are automated. A collaborative robot (UR5) (5) is used for this with a special gripper. For an uniform sample preparation, the compounds are prepared with a volume punch (45 x 45 mm) and positioned on a coated steel plate (2) by using a positioning Aus Wissenschaft und Forschung 60 Tribologie + Schmierungstechnik · 68. Jahrgang · 3-4/ 2021 DOI 10.24053/ TuS-2021-0021 Experimental investigations and analysis of the measurement data In order to be able to record a large number of cycles of demolding characteristics, the self-developed test rig (Figure 3) has been continuously expanded since 2017. The aim of the further development was to significantly Figure 2: Schematic illustration of the repeat units [3] Figure 4: Current test procedure for the measurement of the demolding characteristics Figure 3: Demolding test rig: (1) storage, (2) coated steel plate, (3) stopper at the storage, (4) slide rail of storage, (5) UR 5 with gripper, (6) rack, (7) base of the UR 5, (8) displacement sensor, (9) hydraulic cylinder, (10) force sensor, (11) extension, (12) punch with vulcanization mold and heating unit, (13) slide rail, (14) furnace with heating unit, (15) slide rail with slider TuS_3_4_2021.qxp_TuS_Muster_2021 03.09.21 13: 28 Seite 60 aid. The vulcanization mold is being sandblasted, before a new test series is started and can be assembled immediately in the heated test rig (160 °C). This adjustment by temperature assembling reduces the set-up time by over 70 %. The prepared samples are sorted into the storage (Figure 4) and are taken by the robot. The gripper detects, whether there is a sample in the storage and feeds it into the slider. By using the gripper the slider is closed and the compression molding sequence is initiated, when the vulcanizing mold temperature (160 °C) is reached. In the closing sequence, the hydraulic unit (8) applies a vulcanization pressure of 16 bar and the mold itself cools down approx. 10 degrees due to the cold rubber sample. After the compound specific vulcanization time the demolding sequence is started by an elastomer crosslinking of 90 %. Afterwards the robot opens the slider and activates the halogen lamp and the camera with borescope. The gripper moves the camera in the furnace and afterward it removes the vulcanized sample. For the calculation of the vulcanization time C ONTINENTAL provides the individual Moving-Die rheometer (MDR) curves for each compound (Figure 5 & 6). The vulcanization time is designed to have 90 % crosslinking at the critical point. For this purpose, core temperature curves with a special mold and various compounds have been carried out. Both measurements are combined using the Arrhenius equation and give the necessary equivalent vulcanization time, which must be set on the demolding test rig. The cross-linking characteristics of seven used compounds are almost similar (Figure 6). The compound with RAE (N 10 -ES RAE ) shows the first crosslinking reactions very early. The compound with 100 phr NR differs significantly from the crosslinking characteristics of the other compounds. Due to the rapid crosslinking, the vulcanization time for this compound is shorter. In this study, ten samples were prepared and vulcanized for each compound and their demolding behavior was analyzed. All 80 demolding cycles were carried out without problems and surface defects. However, noticeable is that the N 10 -ES RAE compound shows sidewall friction (Figure 7). The sidewall friction increases with the number of vulcanizates. Aus Wissenschaft und Forschung 61 Tribologie + Schmierungstechnik · 68. Jahrgang · 3-4/ 2021 DOI 10.24053/ TuS-2021-0021 Figure 5: Left: MDR data; Middle: Core temperature curve Right: Comparison of the cross-linking characteristic at 160 °C for one compound Figure 6: Crosslinking characteristics measured with MDR: Left: with functionalized polymer chains; Middle and right: without functionalized polymer chains TuS_3_4_2021.qxp_TuS_Muster_2021 03.09.21 13: 28 Seite 61 Significantly stronger visually determinable fouling effects could be recorded photo-technically after ten demolding cycles (Figure 8). The other compounds show only minimal mold fouling. In order to compare the demolding forces of the eight compounds with each other, the global maximum of each cycle is determined. The maximum demolding forces are shown in Figure 9 and are split up to visualize the differences between functionalized and non-functionalized compounds. The error plots on the right side of the graph show the mean value of cycle 5 to 10 and the respective distances to the measured maximum and minimum. Six of the eight compounds fluctuate with their mean demolding force (cycle: 5 to 10) around 2.5 kN. The N 10 -SS y1 -BR y compound is an exception of the functionalized polymer compounds. The mean demolding force is below the other measurements and if the ninth measure- Aus Wissenschaft und Forschung 62 Tribologie + Schmierungstechnik · 68. Jahrgang · 3-4/ 2021 DOI 10.24053/ TuS-2021-0021 After releasing the base area, the skin surface loosens with some delay. This is due to the apparent sidewall adhesion and forms a plateau at 1.8 - 1.9 kN. With each subsequent demolding cycle, the measured plateau is extended. A possible reason for this increase could be the characteristic of chemical cross-linking of the compound. The cross-linking reaction starts very quickly, which then also significantly influences the interface between mold surface (AlMg4.5Mn) and vulcanizate. Figure 7: Demolding characteristics from cycle 5 to 10 of N 10 -ES RAE Figure 8: Pictures of molds after ten cycles Figure 9: Demolding forces of the polymer study TuS_3_4_2021.qxp_TuS_Muster_2021 03.09.21 13: 28 Seite 62 ment with 3.56 kN is not included in the assessment, the mean variation is significantly lower and the mean measured demolding force is reduced by almost 20 % to 2 kN. The N 100 also has a very large variation of the measured data. The compound N 10 -ES has the highest measured mean demolding force. From the fourth measurement on, this compound also has a very low mean variation and is around 3 kN. The mean demolding force, which is higher than the other compounds increased by approx. 20 %. The higher demolding forces are possibly explainable due to the high branching of the E-SBR. [4] The measured demolding forces were subjected to a correlation analysis with other selected physical data. C ONTINENTAL made the other physicals available (Table 2). If a 5 % significance level of p with a 95 % confidence interval is selected for a clear correlation, the correlation analysis does not reveal any static significance between the demolding forces and the analyzed “physicals”. If the confidence interval is reduced to 85 %, the probability of an existing correlation is also reduced. The probability of an error is increasing to 15 %. But the measured Shore hardness at 70 °C, the elongation of the vulcanizate to break, the elasticity as well as the complex storage modulus at 70 °C and low elongation could be of interest for estimations. The adhesion force analyzes by N ITZSCHE shows, for example, that vulcanizates tend to have higher adhesion forces at room temperature in combination with high Shore hardness. [5] In the following, the complex storage modulus (G*) should be considered, which is composed of the real part of the storage modulus (G´) and the imaginary part of the loss modulus (G´´): (Eq. 1) The storage modulus (G´) of filled vulcanizate samples decreases significantly with increasing deformation amplitude (Figure 10). This phenomenon was discovered in 1942 and intensively researched by P AYNE . In the case of sinusoidal deformation, there is a time delay in the ex- 4 5 46 7 466 Aus Wissenschaft und Forschung 63 Tribologie + Schmierungstechnik · 68. Jahrgang · 3-4/ 2021 DOI 10.24053/ TuS-2021-0021 mean Tensile Elongation Break Ener- G* at G* at G* at G* at F E strength at Break gy Density Hardness Resilience ≈ 10 % ≈ 15 % ≈ 25 % ≈ 40 % [kN] [Mpa] [%] [J/ cm³] shore A [%] [kPa] [kPa] [kPa] [kPa] N 10 -ES RAE 2.5 15.2 727 40.2 53 39.8 1694.3 1357.8 1029.7 804.1 N 10 -ES 3.0 12.5 733 34.4 51 36.2 1481.9 1201.6 925.8 735.3 N 100 2.4 13.9 589 30.3 59 41.8 2039.5 1559.3 1122.8 851.1 N 10 -SS-BR 2 2.6 12.5 553 26.6 61 46.1 2057.4 1694.4 1329.4 1074.4 N 20 -SS-BR 1 2.5 13.4 592 31.1 60 43.1 1867.6 1517.3 1171.8 940.5 N 10 -SS y 2 2.5 12.3 501 23.2 57 50.0 1793.9 1525.8 1230.5 1011.0 N 10 -SS y 1 -BR y 2.3 11.6 498 21.4 56 48.7 1691.0 1455.5 1191.3 985.9 N 20 -SS y 1 -SS y 2 2.6 12.9 507 24.4 56 50.0 1770.1 1480.3 1177.6 960.2 -0.04 0.62 0.42 -0.55 -0.59 -0.52 -0.58 -0.56 -0.50 -0.10 1.93 1.15 -1.61 -1.79 -1.47 -1.75 -1.65 -1.43 0.92 0.10 0.29 0.16 0.12 0.19 0.13 0.15 0.20 70 °C Korrelation für n = 8; R: t p Compound Table 2: Correlation analysis of the polymer study pansion, which lags behind the tension. The time delay leads to a phase shift. The loss angle δ can be determined from the phase shift. G* can be determined from the experimentally measured stress and strain, and G´ and G´´ can be calculated by using δ: [6] (Eq. 2) (Eq. 3) (Eq. 4) K RAUS developed the carbon black (CB) network model. This model 4 5 8 9: ; < 9: ; 46 4 5 =>? 466 4 5 >7@? Figure 10: left: Schematic representation of the dynamic shear modulus with an increase in the deformation amplitude; Middle: carbon black-filled vulcanizate; right: silica filled vulcanizate according to [6 - 8] TuS_3_4_2021.qxp_TuS_Muster_2021 03.09.21 13: 28 Seite 63 change within the filler phase. This model relates to the interactions that take place between the polymer and the filler. The assumption is, that the network arc density is dependent on the amplitude. The network arc density is a measure of the subchains per volume that were created by the chemical crosslinking of the rubber and defines the area between two crosslinking points. In the case of filled vulcanizates, the active fillers represent further network points and the network arc density is increased by this contribution. The resulting network points are stable as well as unstable network points. The unstable network points, however, depend on the amplitude and significantly influence the property profile of the vulcanizate. Unstable network points are crosslinks of the filler with one or very few points that can be broken up by mechanical or thermal stress (Figure 11). [6] The G´ and G´´ of N 10 -SS y1 -BR y , N 100 and N 10 -ES are shown in Figure 12. G´ of N 100 has an easily recognizable global maximum. G´´ decrease the most in this area. A very strong friction between the polymer and filler can explain this maximum, because the occurrence of possible sliding processes on the surface is here particularly large. B ÖHM suggests that the unstable chains cause the observable effect. A bound segment slides into a nearby segment. The sliding process is facilitated by free interaction positions that are nearby. [6] All eight compounds show a maximum in the measured G´´ with an increase of the elongation in the range from 0.3 to 1.0 % (Figure 13 & Table 3). The percentage change from the initial value ranges from 1.8 to 8.7 %. The mixture N 100 has its global maximum at an elongation of 1.0 % and is the only compound with an intersection between G´ and G´´ by an elongation of approx. 0.5 %, if both moduli are shown on one axis. The maximum in combination with the advanced stretching and the point of intersection could together be an indication of the formation of a relatively large variation of the demolding force. This assumption will be considered and verified in the following studies. Aus Wissenschaft und Forschung 64 Tribologie + Schmierungstechnik · 68. Jahrgang · 3-4/ 2021 DOI 10.24053/ TuS-2021-0021 assumes that the CB particles and agglomerates, which build up a 3-dimensional network through V AN -D ER - W ALLS interactions, result in a network collapse due to the increase of the amplitude. However, not all the observed effects can be explained with the CB network model. That’s why Maier developed the model of the variable network arc density. B ÖHM expanded M AIER ’ S model to include silica-filled vulcanizates. [6] In addition to the degree of dispersion, the filler network has a significant influence on the dynamic module. In vulcanizates filled with CB, Payne assumes that the filler-polymer interaction, the hydrodynamic effects and the modulus of an unfilled polymer has a constant contribution. [6] The active fillers are significantly stiffer than the polymer and produce a volume fraction that cannot be deformed. The reason for this is the colloidal spherical structure of the primary particles. The effect of the non-deformable volume fraction is called hydrodynamic reinforcement. [7] The contribution of the filler-polymer interaction is largely determined by the surface activity of the used filler. Depending on the surface activity, there are chemical as well as physical bonds between the polymer and the filler. On CB-filled vulcanizates, the surface of the filler aggregates is covered with an elastomer layer in the nanometer range. This means that there can be no direct contact between the filler units. The thin elastomer layer, which is absorbed by the filler aggregate, is therefore restricted in its mobility. [7, 9] M AIER proved that the contribution is not constant, because the P AYNE model has some weaknesses (temperature and dispersion dependence). Further contradictions in the filler-filler network were discovered when dynamic measurements were carried out on CB-oil, statically stretched, swollen vulcanizates and filled cured and uncured samples. [6] The model of the variable network arc density does not seek the explanation of the P AYNE effect in the structural Figure 11: Schematic representation: Left: Crosslinking structure of unfilled polymers; Right: crosslinking of filler according to [6; 10] TuS_3_4_2021.qxp_TuS_Muster_2021 03.09.21 13: 28 Seite 64 Conclusion Eight different elastomer compounds with different polymer compositions were investigated regarding their demolding characteristics. Each of these trials series consists of ten demolding cycles and the mold was cleaned for each new trial series. After ten demolding cycles the mold was changed and photo-technically recorded. The tested recipes seem to have no significant influence on the demolding force level, but on the measurable variation of the global maxima. Oil-extended E-SBR (RAE oil) seems to stabilize the demolding force over the cycle number and forms a plateau. This recipe combination shows also a different cure characteristic and starts with the cross-linking reaction very early. This could be a reason for the visual mold fouling effects. The branching of the polymer could have an influence on the demolding force level. This assumption should be investigated in further trials with other recipes. A correlation between demolding force and other compound physicals is difficult to determine, but low elongations for the shear and loss modulus could be very promising to estimate whether the measurable demolding forces are subjected to a greater scatter. Aus Wissenschaft und Forschung 65 Tribologie + Schmierungstechnik · 68. Jahrgang · 3-4/ 2021 DOI 10.24053/ TuS-2021-0021 Figure 12: Overview of shore modulus (G´) and loss modulus (G´´) Figure 13: Overview of loss modulus of the polymer study G´´ max percentuale (>0,2 %) variance kPa kPa % % N 20 -SS y 1 -SS y 2 505.4 498.1 0.7 -1.4 Polymer with N 10 -SS y 1 -Br y 416.6 435.0 0.3 4.4 functionalization N 10 -SS y 2 430.6 468.1 0.3 8.7 N 20 -SS-BR 1 565.8 578.8 0.3 2.3 N 10 -SS-BR 2 619.8 612.4 0.7 -1.2 N 100 686.1 722.3 1.0 5.3 Compound with N 10 -ES 793.6 779.7 0.7 -1.8 100 phr NR N 10 -ES REA 719.7 735.2 0.7 2.2 negative percentaule variation = local maximum positive percentuale variation = global maximum Compound G´´ 0 Elogation G´´max Table 3: Maxima of loss modulus TuS_3_4_2021.qxp_TuS_Muster_2021 03.09.21 13: 28 Seite 65 [5] S. Nitzsche: Analyse und Modellierung des dynamischen Ablöseverhaltens von adhäsionsgeprägten Elastomerkontakten, Dissertation, TU Freiberg, 2021. [6] Johannes Böhm: Der Payneeffekt: Interpretation und Anwendung in einem neuen Materialgesetz für Elastomere, Dissertation, Universität Regensburg, 2001. [7] Hasan Kahraman: Experimentelle Untersuchungen und Beschreibung des de-formationsinduzierten anisotropen Werkstoffverhaltens von verstärkten Elastomeren, Dissertation, Rheinisch-Westfälischen Technischen Hochschule Aachen, 2015. [8] Donnet, J.-B.: Black and White Fillers and Tire Compound. Rubber Chemistry and Technology 71 (1998) 3, S. 323-341. [9] Lanzl, T.: Charakterisierung von Ruß-Kautschuk-Mischungen mittels dielektrischer Spektroskopie, Dissertation, Universität Regensburg 2002. [10] Vilgis, T. A., Heinrich, G. u. Klüppel, M.: Reinforcement of polymer nano-composites. Theory, experiments and applications. Cambridge: Cambridge University Press 2009. Aus Wissenschaft und Forschung 66 Tribologie + Schmierungstechnik · 68. Jahrgang · 3-4/ 2021 DOI 10.24053/ TuS-2021-0021 Literature [1] Chigusa Yamada: Rollwiderstand und Nasshaftung ausbalancieren. Beidseitig funktionalisierter S-SBR für Reifen, 2019. https: / / www.kgk-rubberpoint.de/ 24146/ rollwider stand-und-nasshaftung-ausbalancieren/ , abgerufen am: 22.03.2021. [2] A. Kuta, Z. Hrdlička, J. Voldánová, J. Pokorný, J. Plitz: Dynamic Mechanical Properties of Rubbers with Standard Oils and Oils with Low Content of Polycyclic Aromatic Hydrocarbons. KGK Kautschuk Gummi Kunststoffe April (2010), S. 120 -122. [3 Röthemeyer, F. u. Sommer, F.: Kautschuktechnologie. Werkstoffe - Verarbeitung - Produkte. München: Hanser 2001. [4] GDCh: Flüssige Polybutadiene: Spezialisten mit breitem Einsatzgebiet, 2020. https: / / faszinationchemie.de/ ma kromolekulare-chemie/ news/ fluessige-polybutadiene spezia listen-mit-breitem-einsatzgebiet-1/ , abgerufen am: 12.04.2021. TuS_3_4_2021.qxp_TuS_Muster_2021 03.09.21 13: 28 Seite 66