Printer Friendly

Effects of carbon black on elastomer ultimate properties - IR compounds.

Effects of carbon black on elastomer ultimate properties - IR compounds

Applications in industrial rubber products, such as for belts, hosing or vibration mounts, require a balance of mechanical and chemical properties. While the basic properties of compounds are determined by the polymer used, carbon black selection has a critical effect on stiffness, dynamic properties, abrasion resistance and fracture properties of the compound. Among those properties, relatively little effort has been focused on how the carbon black affects the fracture properties[1].

The mechanical processes of fracture can be described in three stages: crack initiation, fatigue crack propagation and catastrophic fracture[2]. The crack initiation in elastomers could be caused by either structure heterogeneities or damage, such as ozone cracking, mechanical cut, etc. In this article, the effects of carbon black on the crack initiation of rubber compounds will be discussed.

Two methods most often used for characterizing the fracture properties are tearing energy (T)[3-6] and J-integral (J)[7-13] methods.

Tearing energy

Rivlin and Thomas[3] developed methods to characterize the tear strength of elastomeric material by extending Griffith's tearing energy concept. The Griffith's concept of tearing energy can be described as the amount of energy dissipated per unit of new surface area created under the stress field. In a mathematical form, it can be defined as (1) [Mathematical Expression Omitted] where T is the characteristic energy (tearing energy) needed to create the new surface, [Mathematical Expression Omitted], and [Mathematical Expression Omitted] is the strain energy density change. The subscript, [Lambda], indicates that the differentiation is carried out under conditions of constant displacement of the parts of the boundary which are not force-free.

J-integral

An alternative approach to characterizing the fracture properties of elastomers is the J-integral concept. The J-integral concept was proposed by Rice[7] and successfully applied by Bagley and Landes[8] as a failure criterion to metal.

Donovan et al[9-12] published a series of papers to report the application of J-integral to characterize carbon black filled NR system. In two dimensional problems, the J-integral can be expressed as a line integral with two terms (2) [Mathematical Expression Omitted] where P is any smooth contour surrounding the crack tip and W is the strain energy density. T is the traction vector defined according to the outward normal along P, u is the displacement vector and ds is an element of arc length along P. This line integral is path independent for all paths starting from one side of the crack, enclosing the crack tip, and ending on the other side of the crack. The two terms represent the decrease in strain energy and the work done on the material inside the volume as the crack advances.

Pure shear test

A sketch of pure shear test sample is shown in figure 1. In general, the sample is a thin rectangular strip and the width is much larger than the height. For an incompressible material such as rubber, a state of pure shear may be achieved by stretching the sample in the height direction, and maintaining the width dimension unchanged. The extension ratios in the three principal directions are [Lambda], 1, and 1/[Lambda], the [Lambda] ([Lambda] > 1) being that in the direction of stretching force. The major advantage of using the pure shear sample is the tearing energy or J-integral does not depend on the crack length[3].

By definition, the tearing energy of a sample under the pure shear strain field can be expressed as[3]. (3) [Mathematical Expression Omitted] where [t.sub.o] is the thickness of the sample in the underformed state, c is the crack length, W is the energy density of the material in a state of pure shear of amount defined by an extension ratio [Lambda] in the height direction and the [h.sub.o] is the initial height of the test sample.

On the other hand, for a pure shear sample, the J-integral evaluation can be much simplified and the same formula as T was obtained[7,13]. In other words, for a test sample under the pure shear strain field

(4) J = T = W [h.sub.o] Our main interest at this stage is to study the effect of carbon black on the crack initiation of rubber compound. T or J calculated at crack initiation is called critical T or J (Tc and Jc).

Experimental

Material

Three quite different rubber systems were used in this study. SBR is a non-crystallizable rubber with reasonably good interaction with carbon black surface. NR is a strain-crystallizable polymer at room temperature. HNBR (hydrogenated nitrile rubber) is a relatively new polymer. It has good oil and fuel resistance similar to NBR, and has heat resistance similar to EPDM. Due to the small amount of carbon-carbon double bonds, in general, HNBR has weak interaction with carbon black.

Nine grades of semi-reinforcing blacks were used in model SBR, NR and HNBR compounds. The blacks had surface area (CTAB) from 31 to 44 [m.sup.2]/g, structure (DBP) from 65 to 140 ml/100g (CDBP = 61-90 ml/100g), and TINT from 46-64% ITRB. The analytical properties of this set of blacks are listed in table 1.

Table 1 - carbon black analytical properties
Samples DBP CDBP CTAB Tint
 (cc/100g) (cc/100g) (m2/g) (%IRB3)
EX-1 66.2 61.3 38.3 64.3
EX-2 140.0 89.7 44.4 55.9
EX-3 66.8 61.1 32.3 53.9
EX-4 122.0 82.8 34.3 46.3
Sterling(*) SO 118.7 82.2 43.2 62.5
Sterling V 92.6 74.5 35.2 53.9
Sterling NS 71.8 64.5 31.0 55.0
Sterling 142 128.0 83.5 38.3 59.5
Sterling 105 134.0 85.3 39.8 61.6


(*)Sterling is a Cabot Corp. tradename

Mixing procedures and compound formulation of SBR, NR and HNBR model compounds were reported previously[14]. The compound processability, cure characteristics, tensile properties, hardness, rebound, dynamic properties and bound rubber were previously evaluated and reported[14].

Equipment set-up

The set-up of the fracture tester is illustrated in figure 2. A pure shear sample holder was designed and made for a modified servo-hydraulic Instron with computer control and automated data analysis. A macro photo lens on the video camera was focused on the inside surface of the crack tip. The high resolution camera and monitor with the appropriate lighting arrangement gave a very clear picture around the crack tip. The crack initiation was observed through a monitor without using coating material on the inside surface of the tip. The real time clock was also recorded on the screen.

A computer program was written to analyze the recorded pictures on an image analysis system, Kontron. The crack initiation time can be more accurately identified and the detailed crack initiation process and fracture surface can be examined using Kontron.

A cutting jig was made to ensure the clean and straight initial cut. It is important to have consistent cuts to produce a consistent crack tip.

Test procedures

The mixed compounds were first milled to thin strips and then fed to the extruder to perform the pure shear samples. The purpose of this step is to minimize the material flow in the curing press. The preformed samples were cured in the curing press at 160 [degrees] C for 30 minutes. The nominal sample dimensions are: L = 200 mm, H = 20 mm and the thickness is 0.6 mm.

In general, three pure shear samples were prepared for each compound. Each sample was conditioned by pre-stretching two to three times to eliminate the Mullins effect before testing. Each sample can be re-measured by re-cutting the crack tip as long as a reasonable ratio between width and height was maintained, i.e. W/H [is greater than] 5. The strain rate was set at 0.2/min.

Data analysis

The tearing energy or J-integral can be evaluated according to the following procedures. First, plot true stress vs. Cauchy strain ([Lambda.sup.2] - 1/[Lambda.sup.2]). A linear relationship should be obtained, and the true modulus, E, can be obtained from the slope. Then, Tc or Jc can be evaluated according to the following formula [Mathematical Expression Omitted] where [Lambda.sub.c] is the extension ratio at the first full crack formed.

Results and discussions

Crack tip morphology

Since most studies in the literature use coating material to detect crack initiation, the crack tip morphology change during the fracture process has seldom been reported[15]. From our observation, the crack initiation process can be described in two steps. First, the crack starts at one spot (or two) on the inside surface of the tip, then it propagates across the thickness direction of the sample to form a full initial crack before it propagates along the width direction of the sample. Different textures of torn surfaces were observed. For example, some of them were smooth, but some of them were rough and gritty. The first crack spot can be influenced by the initial cut and any defects in the sample. Any large lumps occurring in the propagation path can delay the formation of the initial full crack. These two factors can increase the variability of the test. The critical tearing energy (or J integral), Tc (or Jc), based on the initiation of the first full crack was calculated.

It was observed that the change of the crack tip morphology for SBR, NR and HNBR compounds was different during crack initiation. For SBR and HNBR compounds, the crack propagation along the thickness direction can be clearly observed. In general, a nice straight full crack was formed before it propagated in the width direction for SBR compounds; while some of the cracks were rough and gritty for HNBR compounds.

On the other hand, for NR compounds, the crystallization process can be observed from the increasing amount of reflecting white spots in the tip region. The crack opened like a chevron shape above the initial cut and a V shape below the initial cut, and the crystal region was in between. The crystallization process hindered the crack propagation in the thickness direction, and no clear crack path in the thickness direction was observed. A full crack was identified when both edges in the thickness direction were torn.

The carbon black loading and morphology effects on the Tc are discussed separately in the following for HNBR, SBR and NR compounds.

Loading effect

* HNBR compounds - As shown in table 2, the tearing energy of HNBR compounds can be increased about 5 to 10 times when the carbon black loading changes from 20 to 60 phr. Although the true modulus of gum HNBR is lower than that of compounds with 20 phr loading, the tear energy of gum HNBR, 0.682 KN/M, was higher than that of compounds with 20 phr due to the larger lc. In other words, at 20 phr such low loading, carbon black showed adverse effect on Tc of HNBR compounds.

Table 2 - summary of fracture test results of HNBR compounds
Black type Loading True modulus Lamda Tc
 (phr) (KPa) C (KN/M)
Gum rubber 0 812 1.156 0.682
EX-1 20 1153 1.114 0.541
EX-2 20 1162 1.124 0.644
EX-3 20 1089 1.099 0.407
EX-4 20 1202 1.084 0.316
Sterling NS 20 982 1.081 0.246
Sterling 105 20 1126 1.096 0.397
Sterling 142 20 1187 1.108 0.522
Sterling SO 20 1187 1.093 0.381
EX-1 60 1547 1.219 2.519
EX-3 60 1617 1.253 3.376
EX-4 60 2127 1.232 3.806
Sterling NS 60 1580 1.250 3.235
Sterling 105 60 1706 1.252 3.615
Sterling 142 60 1785 1.241 3.401
Sterling SO 60 1831 1.216 2.809


* SBR compounds - As shown in table 3, carbon black loading, in general, had a pronounced effect on Tc (or Jc) for SBR compounds. For most compounds, Tc went through a maximum around 40 phr loading as the loading increased; however, for some compounds with low surface area and low structure carbon blacks, no maximum was observed before 60 phr loading. The Tc of gum SBR is 0.71 KN/M.

Table 3 - summary of fracture test results of SBR compounds
Black Loading True modulus Lamda T
 (phr) (KPa) C energy
 (KN/M)
Gum rubber 0 671 1.176 0.712
Sterling NS 20 823 1.206 1.212
EX-3 20 884 1.268 1.875
Sterling 142 20 931 1.176 1.253
Sterling 105 20 881 1.255 1.737
Sterling SO 20 900 1.183 1.160
EX-1 20 777 1.230 1.366
EX-4 20 847 1.208 1.232
EX-2 20 889 1.191 1.066
Sterling V 20 799 1.256 1.730
Sterling NS 40 1140 1.234 2.056
EX-3 40 1230 1.255 2.619
Sterling SO 40 1395 1.308 4.219
EX-1 40 1143 1.286 3.489
EX-4 40 1309 1.370 5.386
EX-2 40 1610 1.299 4.659
Sterling V 40 1272 1.311 3.825
EX-2 40 1614 1.345 5.480
Sterling V 40 1167 1.300 3.479
Sterling NS 60 1316 1.269 3.078
EX-3 60 1374 1.271 3.276
Sterling 105 60 1753 1.273 4.152
Sterling 142 60 1706 1.264 3.908
Sterling SO 60 1841 1.250 3.715
EX-1 60 1266 1.289 3.409
EX-4 60 1630 1.259 3.532
EX-3 60 1385 1.296 3.326
EX-1 60 1277 1.286 3.451
EX-2 60 2142 1.213 3.262
Sterling V 60 1532 1.256 3.359


* NR compounds - The measured Tc of gum NR is 4.97 KN/M, which was about seven times higher than Tc of gum SBR and HNBR. This was probably due to the strain crystallization effect on NR. As shown in table 4, compared with the Tc of gum NR, the carbon black loading had a moderate effect on Tc of NR compounds with different carbon black loading. For NR compounds, the tendency of forming crystals around the crack tip region under the strain appeared to have dominant influence on the Tc. Like most of the SBR compounds, Tc of NR compounds went through a maximum around 40 phr loading as the loading increased.

Table 4 - summary of fracture test results of NR compounds
Black Loading True Lambda T
 (phr) modulus C energy
 (KPa) (KN/M)
Gum rubber 0 600 1.554 4.973
Sterling NS 20 808 1.490 5.305
EX-3 20 748 1.528 5.777
Sterling 142 20 824 1.438 4.562
Sterling 105 20 909 1.420 4.740
Sterling SO 20 762 1.434 4.194
EX-1 20 732 1.506 5.263
EX-4 20 846 1.452 4.955
EX-2 20 871 1.443 4.904
Sterling V 20 739 1.504 5.240
Sterling NS 40 1109 1.433 6.000
EX-3 40 1137 1.456 6.727
Sterling SO 40 1032 1.437 5.675
EX-1 40 1027 1.411 5.144
EX-4 40 1062 1.449 6.103
EX-2 40 1281 1.430 6.877
Sterling V 40 1145 1.439 6.626
EX-2 40 1231 1.443 6.978
Sterling V 40 1022 1.457 6.090
Sterling NS 60 1320 1.302 3.816
EX-3 60 1423 1.326 4.684
Sterling 105 60 1818 1.269 4.259
Sterling 142 60 1790 1.265 4.100
Sterling SO 60 1513 1.330 5.110
EX-1 60 1385 1.317 4.330
EX-4 60 1590 1.275 3.828
EX-3 60 1392 1.331 4.700
EX-2 60 1681 1.299 4.740
Sterling V 60 1501 1.295 4.108


Carbon black morphology effect

Carbon black morphology effects on Tc of three different systems were also examined. The CDBP, CTAB and TINT were the factors used in a non-linear regression model, which includes the interaction terms. In general, a reasonably good correlation between Tc and carbon black morphology (particle size, aggregate size and structure) could be obtained at 40 and 60 phr loadings. However, the regression patterns were complicated, and they are compound and loading dependent.

* HNBR compounds - At 60 phr loading, the regression analysis showed that Tc can be correlated with CDBP and CTAB ([R.sup.2] = 0.78). It increased with increasing structure and decreasing CTAB. On the other hand, Tc of compounds with 20 phr loading increased with increasing CTAB and TINT, but with decreasing interaction between CTAB and TINT ([R.sup.2] = 0.79).

* SBR compounds - At 60 phr loading, the regression analysis showed that Tc can be correlated with CDBP, CTAB, and TINT ([R.sup.2] = 0.70). It increased with increasing structure and TINT, but with decreasing CTAB. As shown in figure 3, at a fixed level of TINT, the contourplot indicates that the regression pattern of SBR compounds at 60 phr carbon black loading was the same as the regression pattern observed for HNBR compounds at 60 phr carbon black loading.

On the other hand, Tc of compounds with 40 phr loading was correlated not only with CTAB, CDBP and TINT, but also with the interaction term CDBP* CTAB ([R.sup.2] = 0.94). The contourplots indicated that, at a fixed TINT value, the Tc increased with increasing CTAB, and the CDBP effect on Tc depended on CTAB level. In general, Tc increased with decreasing TINT. At 20 phr loading, no good correlation between carbon black morphology and Tc was obtained.

* NR compounds - At 60 phr loading, the regression analysis showed that Tc can be correlated with CDBP, CTAB, TINT and the interaction terms, CDBP*TINT and CTAB*TINT ([R.sup.2] = 0.97). As indicated by the contourplots in figure 4, at a fixed TINT value, Tc increased with increasing CTAB, but with decreasing CDBP. Also, Tc increased with decreasing TINT in a moderate way. The similar regression pattern was observed for compounds with 40 phr loading. Tc correlated with CTAB, CDBP, CTAB*TINT and CDBP*TINT ([R.sup.2] = 0.97). At 20 phr loading, Tc was moderately correlated with CDBP and TINT ([R.sup.2] = 0.76). It increased with decreasing TINT and CDBP.

Compared with SBR compounds, the regression patterns of NR compounds at 40 and 60 phr carbon black loading were different from the patterns observed in SBR compounds.

True modulus

The true modulus obtained from the pure shear test was also analyzed. All the data were obtained after samples were preconditioned (without the Mullins effect).

* HNBR compounds - The true modulus of gum HNBR was 812 KPa. At 20 phr loading, it varied from 980 to 1200 KPa. No good correlation was found for compounds with 20 phr loading. At 60 phr loading, the true modulus ranged from 1540 to 2120 KPa with good correlation with CTAB and TINT ([R sup.2] -0.95). It increased with increasing CTAB but with decreasing TINT.

* SBR compounds - The true modulus in general increased with increasing loading. At 60 phr loading, the true modulus has good correlation with CDBP, CTAB, TINT, CDBP*CTAB and CDBP*TINT ([R.sup.2] = 0.99). The contourplots indicate that the true modulus increased with increasing CTAB and decreasing TINT at a fixed CDBP level, which is similar as previously found for HNBR compounds. However, unlike the HNBR compounds, it also increased with increasing CDBP. At 40 phr, the true modulus depended only on the TINT and CTAB ([R.sup.2] = 0.90). It increased with increasing CTAB and decreasing TINT. No good correlation was found for compounds with 20 phr loading.

* NR compounds - Like the SBR compounds, the true modulus of NR compounds increased with increasing loading. The true modulus of NR compounds only showed moderate correlation with carbon black morphology. No good correlation was found for compounds with 20 phr loading.

* Guth-Gold model and immobilized factor of occluded rubber - Efforts were made to fit the normalized true modulus (Ec/Eg) by modified Guth-Gold model[16,17]: [Mathematical Expressions Omitted] where [Phi.sub.e] is the effective volume fraction loading, which includes the filler and immobilized rubber in the compound. Medalia[18] has suggested that [Phi.sub.e] can be calculated as [Mathematical Expression Omitted] where [Phi] is the filler volumetric fraction loading alone. [Phi'] is the volumetric fraction loading including the occluded rubber and filler, and it can be calculated as [Mathematical Expressions Omitted]

The factor f represents the immobilized fraction of occluded rubber under deformation. Medalia used 0.5 as a constant in one of his[19] studies, and claimed that f can be affected by other factors.

In our study, when the correct effective filler volume fraction is used, [Phi.sub.e] = 0.48 [Phi.sup.1], the true modulus data of all compounds (of different rubbers and loadings) can be fitted by a single master curve, i.e., Guth-Gold equation. The results are shown in figure 5. According to this finding, the immobilized factor f can be calculated from Eq. 7 [Mathematical Expressions Omitted]

The factor f depends on DBPA. As shown in figure 6, no immobilized occluded rubber can be found until the DBPA is higher than about 95.

Conclusion

It was observed that the change of the crack tip morphology for SBR, NR and HNBR compounds was different during crack initiation. Carbon black loading showed the dominant effect on Tc of SBR compounds. But, it has moderate effect on Tc of NR compounds due to the high Tc of gum NR. The tendency of forming crystals around the crack tip region under the strain appears to have dominant influence on the Tc. For HNBR compounds, at 20 phr loading, carbon black showed adverse effect on Tc. For all NR compounds and most SBR compounds, Tc went through a maximum around 40 phr loading as the loading increased.

Carbon black morphology effects on Tc of three different systems were also examined. In general, the reasonably good correlation between Tc and carbon black morphology (particle size, aggregate size and structure) can be obtained at 40 and 60 phr loadings. However, the regression patterns are complicated, and they are compound and loading dependent.

The true modulus obtained from pure shear test was also analyzed. All of the data were obtained after samples were pre-conditioned (eliminating the Mullins effect). In general, the true modulus increases with increasing loading. It showed good correlation with carbon black morphology at 40 and 60 phr loadings for SBR and HNBR compounds. However, it only showed moderate correlation with carbon black morphology for NR compounds.

When the appropriate effective volumetric fraction, [Phi.sub.e] = 0.48 [Phi.sup.1], was used, the true modulus data obtained from all compounds at different loadings can be fitted very well by a single master curve, i.e., the Guth-Gold equation. According to this finding, it was found that the immobilized factor f can be calculated from the carbon black structure (DBPA). [Figures 1 to 6 Omitted]

References

[1]A.I. Medalia, Rubber Chemistry and Technology, 60, p. 45 (1987). [2]M.J. Dole, Machine Design, p. 135-139, March 10 (1988). [3]R.S. Rivlin and A.G. Thomas, J. of Polymer Science, vol. X, No. 3, p. 291 (1953). [4]A. Kadir and A.G. Thomas, Rubber Chemistry and Technology, 54, p. 15 (1981). [5]D.G. Young, ibid., 58, p. 785 (1985). [6]D.G. Young, ibid., 59, p. 809 (1986). [7]J.R. Rice, J. of Applied Mechanics, p. 379, June (1968). [8]J.A. Begley and J.D. Landes, ASTM.STP 514, p. 1 (1972). [9]D.J. Lee and J.A. Donovan, Theoretical And Applied Fracture Mechanics, 4, p. 137 (1985). [10]R.F. Lee and J.A. Donovan, Rubber Chemistry and Technology, 59, p. 787 (1986). [11]R.F. Lee and J.A. Donovan, ibid, 60, p. 675 (1987). [12]H. Liu, R.F. Lee and J.A. Donovan, ibid., 60, p. 893 (1987). [13]K.A. Mazich, K.N. Morman, F.G. Oblinger, T.Y. Pan and P.C. Killgoar, Paper No. 58, ACS, Rubber Division 132nd meeting, Cleveland, Oct. (1987). [14]J.M. Funt and B. Chung, Paper No. 43, ACS Rubber Division meeting, Mexico City, May (1989). [15]A. Goldberg and D.R. Lesuer, Rubber Chemistry and Technology, 62, p. 272 (1989). [16]E. Guth and O. Gold, Physics Rev. 53, 322 (1938). [17]E. Guth, Proc. 2nd Rubber Technol. Conference, 353, London (1948). [18]A.I. Medalia, Rubber Chemistry and Technology, 47, p. 411 (1974). [19]A.I. Medalia, ibid., 51, p. 437 (1978).
COPYRIGHT 1991 Lippincott & Peto, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 1991, Gale Group. All rights reserved. Gale Group is a Thomson Corporation Company.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:industrial rubber
Author:Ouyang, G.B.
Publication:Rubber World
Date:May 1, 1991
Words:4117
Previous Article:Rubber and collagen-fiber blends for studless winter tire applications.
Next Article:Editorial.
Topics:


Related Articles
CB's role in compound curing behavior.
New blacks address compounding challenges.
Dynamic properties of rubber.
Use of reinforcing silica in model sidewall compounds: effects of carbon black type, polymer type and filler level.
The effect of carbon-black filling on the pyrolysis behavior of natural and synthetic rubber.
Carbon black's effects on vibration isolation.
Cure and mechanical behavior of elastomeric compounds containing devulcanized materials.
Viscous heating and reinforcement effects of fillers using the rubber process analyzer.
The challenge ahead - new polymer/filler systems. (Tech Service).
Cure system and carbon black effects on NR compound performance in truck tires.

Terms of use | Copyright © 2016 Farlex, Inc. | Feedback | For webmasters