Fatigue is defined as decay caused by cyclic deformations at an amplitude less than necessary for fracture in one cycle. Such failures are initiated by flaws which act as stress concentrators. These flaws occur in the material either through mechanical or chemical action during service or through agglomeration of certain ingredients during mixing and fabrication. This paper deals with the latter process, where the nature and size of the flaws as well as the properties of the matrix are contingent on carbon black variables. Using the tearing energy concept of fatigue developed by Lake and Lindley, it was shown that the size of the flaw is primarily determined by carbon black particle size. On the other hand, the cut growth constant depends on carbon black structure. When translated to actual fatigue life using the Monsanto Fatigue-to-Failure Tester, these relationships mean that under constant strain conditions, compounds containing coarse carbons will have a significantly higher fatigue life than those with fine carbons. Under conditions of constant strain, higher structure carbons will impart a slight positive effect. However, under conditions of constant stress, the beneficial effects of structure become magnified. Other factors known to affect fatigue life were also considered. These are : set, stress relaxation, hysteretic energy dissipation, and flaw size distribution.

Criteria for minimizing hysteresis in carbon-black-filled elastomer blends are: (1) large black unit size, wide distribution, and low structure; (2) higher black loading in the discrete polymer phase (large zones); and (3) polymer of lower hysteresis as the continuous phase (low black loading). Of the different strength properties that were evaluated, tear and fatigue resistance showed the greatest dependence on black location in NR-BR and NR-SBR blends. Both properties were markedly higher for NR-BR, with most of the black in the NR phase. In NR-SBR, tear strength was higher with high loadings of black in the SBR. Fatigue life showed a reverse pattern, but the variations were not as great in this system. Criteria for maximizing tear resistance are: (1) small black unit size, low structure; (2) higher loading of carbon black in the continuous polymer phase ; and (3) the polymer of higher strength as the continuous phase. There may also be optimum levels of polymer zone size and black size distribution which affect tear strength. Optimized performance in a 50:50 NR-BR radial truck tire tread stock was obtained with a wide-distribution N-375 (50:50 N-351-N-110 blend) black with 75% location in the NR phase. This black gave 5–6°C lower heat buildup, equivalent modulus, 35% higher tear strength, and almost double the fatigue life of a conventionally mixed N-220 at essentially equivalent tread wear resistance (−4%). Optimized performance in 50:50 NR-SBR was indicated for a wide-distribution N-231 type (50:50 blend of N-351-N-119) with 75% of the black in the SBR phase. In comparison to conventionally mixed N-234 and N-220, this black gave about 8–10°C lower heat buildup, about 15% lower modulus, and essentially equal tensile, tear, and fatigue properties. Alternatively, a standard N-231 type with 75% location in the NR phase showed about 5–6°C lower heat buildup, about 20% lower modulus, equivalent tensile strength, 60% higher tear resistance, and 75–120% higher fatigue life.

The influence of carbon black type on the dynamic properties of natural rubber is examined for nine carbon blacks, encompassing a wide range of structure and surface area. The dynamic properties are measured at 23° C and 100° C, over a 2 decade range of log frequency, and over a range of 1 to 25 per cent dynamic strain. The effect of carbon black structure and surface area on dynamic properties of a natural rubber compound are dependent on the combination of strain amplitude and frequency test conditions. The effect of carbon black surface area and structure at different temperatures varies on an absolute basis but not necessarily on a relative basis. Correlations of forced non resonant measurements of dynamic properties with dynamic measurements using other instruments confirms the strain amplitude and frequency interaction with carbon black type. Also, the Goodyear-Healy Rebound machine, the Goodrich Flexometer, and the Pirelli Hysteresimeter can be used to estimate basic dynamic properties over limited strain amplitude and frequency conditions. In addition, mixing procedure was found to have a significant effect on the dynamic properties of a natural rubber compound filled with N-327.

Conventional and diffracted beam electron microscopy can be used to resolve the ultimate crystallite units in standard and heat treated carbon blacks. They are very sensitive in the detection of small differences in crystallite size and orientation. Combined with x-ray diffraction methods and conventional vulcanizate testing, high resolution electron microscopy appears to be a useful and practical means to study carbon black surface activity. All commercial blacks appear to conform in part to the Heckman and Harling concentric crystallite model. Surface activity appears to be related to the extent of this alignment; the less the tangential orientation of the graphite layer planes to the surface, the greater the surface activity. Heating carbon blacks in an inert atmosphere for varying times and temperatures (850 to 1400° C) causes a gradual depression of surface activity with increasing treatment severity, defined by increases in the size and concentric orientation of surface crystallites. At no time was the effect of carbon black on vulcanizate properties changed significantly by heat treatment without disrupting the initial orientation of surface crystallites. Carbon blacks differ in the rate of change of crystallite orientation in response to heat treatment. The initial degree of crystallinity, porosity, and surface volatile content all appear to affect the rate of crystallite orientation and growth. Increases in average crystallite height, Lc, are a good measure of effects of heat treatment in excellent agreement with changes in surface crystallite size and orientation observed with the electron microscope. Commercial ISAF carbon blacks of varied structure have relatively minor differences in surface crystallite orientation, indicating small differences in surface activity. The higher modulus and treadwear associated with the use of high structure blacks appear to be predominantly a result of the chain structure itself. Further evidence was obtained showing that black structure and surface activity are independent properties each of which can have a significant influence on vulcanizate properties such as modulus, treadwear, and hysteresis.

The viscous modulus of oil-extended SBR loaded with carbon black relative to the gum stock increases with volume fraction of carbon black. This increase is much larger than predicted by the Guth-Gold equation and also greater than the increase in the elastic modulus. This behavior reflects the sum of at least three effects: geometrical (strain amplification), changes in the viscoelastic properties of the rubber, and formation of a three-dimensional network of carbon black and rubber. The geometrical effect increases with volume fraction of carbon black. Its per cent contribution to the viscous modulus decreases as the carbon black loading increases because of increased contribution of the viscoelastic effect at low loadings and of the network effect at high loading. The geometrical effect can be described reasonably well by the Guth equation with a shape factor and with the volume fraction increased by the additional immobilized rubber (bonded elastomer). The viscoelastic effect due to 60 phr black contributes at least 20 per cent to the viscous modulus and is related to structure, surface area, and volume fraction of the carbon black. It can be at least partially accounted for by rubber breakdown during mixing. At high volume fractions of carbon black (60 phr), network formation contributes about 25 to 45 per cent to the viscous modulus. It increases with decreasing interunit spacing which depends on the average volume of the carbon-black units and the amount of bonded elastomer. For a thermal black, N—990, the interpretation of the results is complicated by the distinct possibility that slippage contributes to hysteresis. We conclude, however, that at low loadings the increase of the viscous modulus is much greater than that of the elastic modulus and that network effects contribute to the viscous modulus at high loadings.