PRÜFEN UND MESSEN TESTING AND MEASURING

41KGK · 04 2017www.kgk-rubberpoint.de

Processing LAOS branched polymers “green tire recipe” polymer/filler in­teraction silane functionality

In the last years, characterisation of poly-mers by Fourier analysis of Large Ampli-tude Oscillating Shear (LAOS) signals be-came a powerful, widely spread tech-nique. Oscillatory rheometer measure-ments were carried out in the non-linear viscoelastic domain for summer tire tread compounds based on the “green tire reci-pe”. Compound variations were realized by changing the polymer base well the as filler system including silane. A main fo-cus lay on the question if LAOS is able to detect differences of polymer architec-ture even in the filled compound. A further question was which potential can be attributed to the method regarding polymer-filler interaction rised by diffe-rent surface area and/or structure or vari-ations of silane functionality.

Charakterisierung von Sommerreifenlaufflächen­mischungen mittels der „Large Amplitude Oscillating Shear (LAOS)“­Methode Verarbeitungsverhalten LAOS ver­zweigte Polymere „Grüner Reifen Re­zeptur“ Polymer/Füllstoff­Wechsel­wirkung Silanfunktionalität

In den letzten Jahren hat sich die Fourier-Analyse von „Large Amplitude Oscillating Shear (LAOS)“-Signalen zur Charakterisie-rung von Kunststoffen als aussagekräftige Technik etabliert. Es wurden Untersu-chungen des nicht linearen Regimes mit-tels oszillierenden Rheometermessungen an Sommerreifenlaufflächenmischungen auf Basis der „Grüner Reifen Rezeptur“ durchgeführt. Die Compounds wurden hinsichtlich der Polymerbasis wie auch des Füllstoffsystems inkl. Silan variiert. Das Hauptaugenmerk liegt auf der Be-antwortung der Frage, ob es mit LAOS auch noch in einem gefüllten Compound möglich ist, die Polymerarchitektur abzu-bilden. Außerdem soll das Potential der Methode ermittelt werden, Polymer-Füll-stoffwechselwirkungen in Abhängigkeit der Füllstoffoberfläche und/oder -struktur bzw. Variationen in der Silanfunktionalität abzubilden.

Figures and Tables: By a kind approval of the authors.

IntroductionIn the context of the production of rub-ber compounds, especially of highly filled Silica/ Silane compounds, it is often mentioned that processing becomes more difficult. An unambiguous defini-tion of “processing” seems difficult since it encloses such different aspects like behaviour in an internal mixer, on an open mill, a roller die, in an extruder or calender. Despite the difficulties of a pre-cise meaning, it was always desired to get an idea of processing behaviour by rheological characterization of com-pounds (with the Mooney measurement as the best-known method). Thus it is of obvious interest to examine possible contributions of Large Amplitude Strain Analysis (LAOS) in this respect.

LAOS became a widely spread tech-nique to analyse the dynamical behav-iour of elastomer melts using a Fourier Transformation (FT) [1, 2, 3]. Also for the analysis of filled compounds this tech-nique was recently applied [4, 5 ,6].

LaosGenerally speaking, Large Amplitude Os-cillating Shear means to carry out oscilla-tory measurements at a certain frequen-cy and at high values of shear strain (mostly at 1000 %). Now different ways of testing are associated with the term “LAOS” regarding the used strain values as well as the FT analysis technique. Thus investigations in the present study were done for the entire accessible strain re-gime, i.e. from the linear to non-linear regime. The Fourier analysis in this study was provided by a discrete Fourier trans-formation (not FFT) based on 5 cycles each containing 90 data points.

All tests were carried out by a Rubber Process Analyser from TA at a fixed tem-perature of 80 °C and a frequency of 0,2 Hz. The relatively low frequency was chosen to enable large strains. It is known that some authors suggest – es-pecially if the molecular polymer archi-tecture is in the focus of the analysis - to adjust the measuring temperature such that a common viscosity is established (H. Burhin shows that by an example of

PIB samples with different molecular weight [7]). Because the approach re-quires the existence of a kind of “tem-perature-frequency equivalence” for non-linear viscoelasticity, it is not rele-vant for our compound studies. Instead the boundary conditions were fixed to establish the opportunity for a direct comparability of LAOS results.

To illustrate the impact of large strains on the dynamical behaviour of filled compounds figure 1 shows the depen-dence of shear modulus on strain (the so called “Payne-test”). The drop of stiffness is usually associated with “non-linearity” and becomes more explicit when the stress-strain curves of the measure-ments are considered.

Figure 2 shows the results at 1 % shear (the material was the same as that used for figure 1): the graph displays the typi-cal loss ellipse. The results from the mea-surements at high strains of 1000 % are presented in figure 3: – the stress strain curve deviates significantly from an el-liptical shape, which indicates non-linear behaviour.

Providing a harmonic oscillating strain input, the viscoelastic stress response at steady state can be written as a time-do-main Fourier series of odd harmonics [8]:

𝜎𝜎 (𝑡𝑡,𝜔𝜔. 𝛾𝛾0) = 𝛾𝛾0 � {𝐺𝐺′𝑛𝑛 (𝜔𝜔. 𝛾𝛾0) sin𝑛𝑛𝜔𝜔𝑡𝑡 + 𝑛𝑛 :𝑜𝑜𝑜𝑜𝑜𝑜

𝐺𝐺"𝑛𝑛 (𝜔𝜔. 𝛾𝛾0) cos𝑛𝑛𝜔𝜔𝑡𝑡}.

𝜎𝜎 (𝑡𝑡,𝜔𝜔. 𝛾𝛾0) = 𝛾𝛾0 � {𝐺𝐺′𝑛𝑛 (𝜔𝜔. 𝛾𝛾0) sin𝑛𝑛𝜔𝜔𝑡𝑡 + 𝑛𝑛 :𝑜𝑜𝑜𝑜𝑜𝑜

𝐺𝐺"𝑛𝑛 (𝜔𝜔. 𝛾𝛾0) cos𝑛𝑛𝜔𝜔𝑡𝑡}. (1)

For sufficiently small strain amplitudes γ0 a linear material response is observed indicated by the fact that the 1st har-monic provides the only significant con-

Characterization of Summer Tire Tread Compounds by Large Am-plitude Oscillating Shear (Laos)

AuthorsM. Heinz, Wesseling, J. Kroll, Leverkusen, T. Rauschmann, Wetzlar

Corresponding author:Dr. Michael HeinzEvonik Resource Efficiency GmbH, Bruehler Str. 2, D-50389 Wesseling Phone: +49 2236 763472E-Mail: michael.heinz@evonik.com

PRÜFEN UND MESSEN TESTING AND MEASURING

42 KGK · 04 2017 www.kgk-rubberpoint.de

tribution (see fig. 4). For larger deforma-tion amplitudes, higher odd harmonics appear proving that the material now shows a non-linear response (see fig. 5 for an analysis of the real part of the 1000 % signal in the typical alternating sequence).

To focus explicitly on non-linear ef-fects, the higher Fourier coefficients will be normalized with respect to the linear contributions, i.e.: ri = |Ri|/R1, i = 3, 5. Fig-ure 6 shows the relative values r3, r5 and r7 as a function of strain. For the present analysis it is sufficient to consider the real (i.e. in phase) part of the signal. This is due to the dominating elasticity achieved for the materials and test con-ditions under consideration.

The logarithmic display underlines the fact that only r3 and r5 are of significance at strains below 1000 %. At 1000 % strain some non-trivial contributions of r7 are recognized. But overall, it can be stated that in usual large strain domains (i.e. up to 1000 %) non-linear strain contributions of order 7 and higher can be neglected.

CompoundsOn basis of the “green tire recipe” com-pounds with variations of BR-type, filler and silane are produced. Table 1 shows the general ingredients, where tables 2a to 2d display the individual variations. Additionally figure 7 provides the infor-mation about surface area and structure of the carbon blacks used.

Table 3 shows the assignment of the variations to the compounds; different coloured and shades fields should under-

line the systematic of the variations. The same holds for the charts of results.

The compounds were mixed on a GK 90 E from Werner+Pfleiderer (with inter-meshing PS5 geometry) in three stages.

Figure 8 shows the results of the Mooney viscosity; it is increasing for the linear BR-type (providing a higher ML compared to the branched type), for bi-functional silane, for higher surface area of the filler (both silica and carbon black) and for higher carbon black structure.

Fig. 1: Impact of Large Amplitude Oscillating Shear on stiffness.

1

Fig. 2: σ over ε in the linear regime (filled compound, 1 % strain).

2

Fig. 3: σ over ε in the non­linear regime (filled compound, 1000 % strain).

3

Fig. 4: Fourier analysis for linear behaviour (filled compound, 1 %).

4

Fig. 5: Fourier analysis for non­linear behaviour (filled compound, 1000 %).

5

log strain

non-linear

linear

shea

r mod

ulus

G"

0,01 0,1 1 10 100 1000

PRÜFEN UND MESSEN TESTING AND MEASURING

43KGK · 04 2017www.kgk-rubberpoint.de

The cure behaviour is shown in figure 9 by virtue of the t95 values from a MDR test. Curing time is increasing for linear BR-type and, in addition, the influence of the different types of silanes is much more pronounced for this BR-type.

Figure 10 displays the stiffness (G*: absolute value of complex modulus) as a

function of strain (Payne test, RPA). Break down of filler network occurs at higher shear amplitudes, leading to the typical pronounced drop of stiffness.

Further physical tests of the cured samples were carried out. Significant observations are shortly summarized without further discussion:

■ Filler dispersion was detected by a tactual test method

A better filler dispersion was found for the highly branched BR-type. Dispersion of low surface area – low structure carbon black is extremely chal-lenging.

■ Hardness: ■ Slightly increasing for linear BR-type. ■ Significantly increasing for use of bi-

functional silane. ■ Generally increasing for high struc-

ture and/or high surface of filler. ■ Reinforcement:

■ Decreasing slightly for linear BR-type. Significantly increasing for use of bi-functional silane.

All the results are in agreement with a common understanding of filler-poly-mer interaction principles. Thus the chosen compound set provides a range of compound formulations, which cov-ers the main relevant aspects in that field.

Fig. 6: Relative Fourier coefficients r3, r5, r7 as a function of strain (filled compound).

6

Fig. 7: Surface area and structure of the used carbon black­types.

7

1 General ingredients “Green tire recipe”Ingredient Silica compounds/phr Carbon black compounds/phrBuna VSL 4526-2 96,25 96,25BR 30 30Silica 80 -Carbon black - 72OCTEO or Si 266® 4 or 5,8 -ZnO 2 2Stearic acid 1 1Vivatec 500 8,75 8,75Vulkanox HS/LG 1,5 1,5Vulkanox 4020 2 2Protector G3108 2 2Rhenogran DPG-80 2,5 -Perkacit TBzTD PDR-D 0,2 -Vulkacit CZ 1,6 1,6Sulfur 2 1,4

Fig. 8: Mooney viscosity.

8

Fig. 9: Curing behaviour.

9

PRÜFEN UND MESSEN TESTING AND MEASURING

44 KGK · 04 2017 www.kgk-rubberpoint.de

Results

ProcessingGarvey die tests are regarded as a signifi-cant method to estimate the extrusion behaviour of compounds being an im-portant aspect of processing. Thus these kind of tests were carried for a various screw speeds (up to 100 rpm). Figures 11a to 11c display pictures of several compounds at screw speed of 45 rpm.

It is remarkable that the BR.-type (lin-ear vs. branched) influences the extru-sion quality tremendously: if one assigns the Si 266®/ULTRASIL® 7000 GR filler system to be challenging from the pro-cessing perspective and N326 to achieve good extrusion behavior, the change of the BR polymer from linear to branched dominates these influences (e.g. CB1220 provides smooth stripes even for the Si 266®/ULTRASIL® 7000 GR system).

LAOSIt is known that LAOS is an appropriate tool for a reliable evaluation of (long chain) branching in case of pure polymer melts. The branching variation in the present study is provided by the BR types, which show pronounced differences in the non-linear regime (see figure 12a and 12b).

Since the BR is underrepresented (30 PHR vs. 96 PHR of oil extended S-SBR) in the compounds the influence of blend-ing was first checked. Figures 13a and 13b show that the dynamic behaviour of the S-SBR dominates the LAOS analysis.

These observations suggest that non-linear analysis is not able to differentiate in between the BR polymers for the com-pounds considered. This is indeed the case as figure 14a and 14b might docu-ment by the example of the most differ-

2a Variation in BR­typesButadiene Rubber

Buna CB 22 Buna CB 1220Type Neodymium BR Cobalt BRMooney viscosity/ME 63 40cis 1,4 content/ % min. 96 min. 96degree of branching linear highly branched

2b Variation in Silica­typesSilica

ULTRASIL® VN2 GR ULTRASIL® 7000 GRBET (multi point) / m2/g 130 170CTAB / / m2/g 125 160pH 6,9 6,5

2c Variation in Silane­typesSilane

OCTEO Si 266®Type mono-functional bi-functional

2d Variation in Carbon Black­typesCarbon Black

N339 N326 N234STSA/m2/g 88 77 113OAN/ml/100 g 120 72 125COAN/ml/100 g 99 69 100

3 Assignment of variations to the compoundsBR Silica Silane Carbon black

Comp. 1 Buna CB 22 ULTRASIL®VN2 GR OCTEOComp. 2 Buna CB 22 ULTRASIL®VN2 GR Si 266®Comp. 3 Buna CB 22 ULTRASIL®7000 GR OCTEOComp. 4 Buna CB 22 ULTRASIL®7000 GR Si 266®Comp. 5 Buna CB 22 N339Comp. 6 Buna CB 22 N326Comp. 7 Buna CB 22 N234Comp. 8 Buna CB 1220 ULTRASIL®VN2 GR OCTEOComp. 9 Buna CB 1220 ULTRASIL®VN2 GR Si 266®Comp. 10 Buna CB 1220 ULTRASIL®7000 GR OCTEOComp. 11 Buna CB 1220 ULTRASIL®7000 GR Si 266®Comp. 12 Buna CB 1220 N339Comp. 13 Buna CB 1220 N326Comp. 14 Buna CB 1220 N234

Fig. 10: RPA Payne test.

10 11

Fig. 11a: Garvey die profile of compound 6 N 326, CB 22 => 24 bar

Fig. 11b: Garvey die profile of compound 4 ULTRASIL® 7000, Si 266®, CB 22 => 45 bar

Fig. 11c: Garvey die profile of compound 11 ULTRASIL® 7000, Si 266®, CB 1220 => 31 bar

PRÜFEN UND MESSEN TESTING AND MEASURING

45KGK · 04 2017www.kgk-rubberpoint.de

ent filler systems (Si 266®/ULTRASIL® 7000 GR and N234).

Therefore, the further discussion focuses on the systematic analysis of differentiations with regard to filler variations based on the CB1220 com-pounds.

Figure 15 shows the r3 values for the silica-filled compounds. For both silica types, one systematically finds a slight crossover at about 100 % shear for the mono- respectively bi-functional silane (which is recognized for the CB22 sys-tems in the same manner). Whether this

effect is of significance regarding the ac-curacy of the tests should be investigat-ed in the future.

Similar observations can be made for the influence of the surface area of the silica (and for r5 and r7 but at a reduced level of significance).

Fig. 12a: Lissajous figures of pure CB22 (linear: higher n.l. contribu­tions) and CB1220 (branched: smaller n.l. contributions) melts.

Fig. 12b: LAOS analysis of pure CB22 (linear: higher n.l. contribu­tions) and CB1220 (branched: smaller n.l. contributions) melts.

Fig. 13a: Lissajous figures of pure blend systems. Fig. 13b: LAOS analysis of pure blend systems.

12

13

Fig. 14a: LAOS analysis of Si 266®/ULTRASIL® 7000 GR filled compounds. Fig. 14b: LAOS analysis of N234 filled compounds.

14

PRÜFEN UND MESSEN TESTING AND MEASURING

46 KGK · 04 2017 www.kgk-rubberpoint.de

Fig. 15: r3 values of the silica filled compounds in the VSL / CB 1220 – blend .

15

For carbon black filled compounds the influence of structure seems to be not recognisable by LAOS (see fig. 16), but regarding the surface area a behaviour analogous to that in case of silica is found: a higher surface area leads to higher r3 values for lower strains where the values are more or less similar at strain values about 100 %.

Comparing the high surface carbon black (N234) with the high surface silica (ULTRASIL® 7000 GR), it is remarkable that for the use of mono-functional si-lane the r3-values are similar to the val-ues of the carbon black filled compound, whereas a large difference is found for the bi-functionally bounded silica (see fig. 17).

ConclusionLarge oscillation strain analysis allows for a reliable differentiation of polymer melts where the effects noticed by the relative Fourier coefficients are usually dominated by the branching structure of the polymer. In S-SBR - BR blends the main part (here S-SBR) seems to deter-

Fig. 17: Comparison of r3 values of the carbon black vs. silica filled compounds.

17

Fig. 16: r3 values of the carbon black filled compounds in the VSL 4526­2 / CB 1220 – blend.

16

mine the non-linear behavior (as mir-rored by LAOS analysis) and masks the differences introduced by the polymer architecture of the BR-types (linear vs. branched) nearly completely.

The results for filled compounds show that the differences of the BR-types in the polymer phase remain nearly com-pletely undetected by analysis of the large strain regime (i.e. for strain above 100 %). This is particularly remarkable because the polymer variations induce a large impact on processing, here shown by the example of Garvey die extrusion. Thus processing behavior seems not to be predictable by “classical” (i.e. high strain) LAOS analysis.

In the case of filled compounds Fourier analysis of strains in the intermediate range (10 % to 100 % strain) shows poten-tial with regard to the estimation of filler polymer interaction (silica vs. CB, influ-ences of silane type or surface area). Since the capability of differentiations is main-tained up to the large strain regime in a reliable manner (a significant “crossover” was detectable), a further analysis of the

method with regard to deeper under-standing of filler polymer interaction is promising (as a kind of “Advanced Payne test”). Whether a further inspection of the intermediate strain regime also bears information directly utilizable for an esti-mation of processing behavior will also be subject of further investigations.

References[1] T. Rauschmann; Comparison of Rheological

Properties of Polymer melts Under Shear Flows in Closed Cavity and Open Boundary Rheometers; (2015), International Elastomer Conference, Cleveland, Ohio (USA).C. Laprais, “Industry Standards and Government Regu-lations”, (2009).

[2] K. Hyu, et al.; A review of Nonlinear Oscilla-tory Shear Tests: (2011), manuscript for pub-lication in Progress in Polymer Science.

[3] D. Dhupia et al.; Long Chain Branching; (2014) Tire Technology International.

[4] G. Nijman; Chancen für bessere Mischproz-esse durch Streuerung der Mischungs- rhe-ologie; (2014) 14th International Seminar on Elastomers, Bratislava (Slovakia).

[5] C. G. Robertson; Linear-Nonlinear Dichotomy of the Rheological Response of Particle-Filled Polymers; (2014) J. of. Applied Polymer Sience .

[6] J. L. Leblanc; Rubber-filler interactions and rheological properties in filled compounds; (2002) Progress in Polymer Science

[7] Stadler F. J., Leygue A., Burhin H., Bailly C.; The potential of large amplitude oscillatory shear to gain an insight into the long chain branching structure of polymers; in: The 235th ACS national meeting, polymer pre-prints ACS, 49, New Orleans, LA, USA (2008) 121.

[8] A. J. Giacomin, J. M. Dealy; Large-amplitude oscillatory shear; in Techniques in rheologi-cal measurement; (1993) Ch. 4, London: Elsevier.