Cut growth rate evaluation of antidegradants in model tire sidewall compounds.

Cut growth rate evaluation of antidegradants in model tire sidewall compounds

Service failure of tire rubber components, such as sidewalls, generally occurs as a result of fatigue. Fatigue in a material leads to fractures creating new surfaces resulting in the disintegration of the material. Fracture also takes place in a material under static loading because of chemical degradation due mainly to the presence of oxygen and ozone in the atmosphere. The basics of both, fatigue and fracture, were recently reviewed in depth by Doyle (ref. 1) and Hamed (ref. 2).

To measure or predict the service life of a rubber compound, designed to be fatigue resistant, several traditional methods are available e.g., DeMattia (crack initiation), Du Pont Flex and fatigue-to-failure. A rubber vulcanizate in such a testing is subjected to a repetitive mechanical stress cycle until complete failure occurs. The plot of lifetime (number of cycles to failure) versus the cyclic stress amplitude gives the familiar S-N plot. This traditional approach, however, in spite of its widespread practical usefulness, does not define fatigue resistance as an inherent compound property in terms of a parameter related to the compound composition and being independent of test specimen geometry/shape. In traditional methods the crack initiation phase represents a large fraction of lifetime. Furthermore, the traditional method is unable to distinguish between the three well recognized (ref. 3) and successive stages of a failure process:

* crack initiation;

* crack propagation;

* catastrophic fracture.

The crack propagation stage is the rate determining step governing the service life of a rubber compound. A well defined cut initiated in a rubber vulcanizate will grow incrementally during this stage due to cyclic mechanical straining.

The crack growth rate in a rubber compound is dependent on the strain/stress conditions, the chemical composition of the base polymer, the type of crosslinking and the compounding ingredients, in particular the crack inhibitors or the antidegradants. Lindley (ref. 3) has reported the effects on cut growth of crosslinking systems, carbon black and the antioxidant phenyl-[BETA]-naphthylamine (PBN) in unfilled NR vulcanizates. Gent (ref. 4) has reported the same effects of PBN and other PPD derivatives in unfilled NR and SBR.

This article presents a first systematic evaluation of the effect of widely used chemical classes of antidegradants on pure shear cut growth rate (CGR) in NR, SBR, model NR/BR tire sidewall, and model NR earthmover tread compounds. Table 1 gives a list of the antidegradants investigated in the present study.

A study of this type is important for the following reasons:

* the service failure of a tire component due to fatigue is an inherent compound property;

* the rate of failure of a compound can be minimized by judiciously selecting the antidegradants;

* the predominant mode of failure of a compound during its service is that of a cut growth normally independent of the specimen geometry;

* the stress conditions such as in traditional methods responsible for a catastrophic failure are not voluntarily imposed on a compound during its service. [Tabular Data Omitted]

Theory

The present work measures the dynamic cut growth rate under relaxing conditions and tear energy using a pure shear test specimen. The method and the theory behind pure shear cut growth and tear energy has already been reported and well documented by Thomas (ref. 5), Rivlin (ref. 6) and recently by Young (ref. 7). Therefore, only essential definitions are discussed here.

Pure shear specimen

The pure shear geometry test specimen used to measure cut growth rate and tearing energy is shown in figure 1. It consists of a thin (1 mm), long (220 mm) rubber vulcanizate with a narrow width (19.5 mm) and ribs for clamping purposes. A constant rate of cyclic deformation is imposed to cause the initiated cut to propagate. The portion of the specimen away from the tip of the initial cut of length, [C.sub.o], and the opposite uncut edge is in pure shear deformation.

Tear energy

Tear energy is defined as the change in stored elastic energy when a crack of unit length is created. It is again known (ref. 3) that there are two extremes of this energy, namely, threshold tear energy, [T.sub.o], and critical tear energy, [T.sub.c].

Threshold tear energy, [T.sub.o], is the minimum tearing energy below which no crack propagates due to mechanical energy input in the vulcanizate. Thus, it is an intrinsic property of a rubber compound and below this energy value the mechano-oxidative cut growth can occur (ref. 3).

Critical or catastrophic tear energy, [T.sub.c], corresponds to the energy at which the crack propagates catastrophically.

For pure shear specimens the tear energy is independent of the cut length (ref. 6) and is simply a product of the strain energy density, [W.sub.o], and the unstrained specimen width, [1.sub.o], T = [W.sub.o] * [1.sub.o] (1)

The strain energy for a pure shear specimen is given by W = [C.sub.10] * (I-3) + [C.sub.20] * [(I-3).sup.2] (2) (I-3) being the strain invariant which expressed in function of extension ratio (E.R.) is equal to [(E.R.-1/E.R.).sup.2]. The coefficients [C.sub.10] and [C.sub.20] in [N/cm.sup.2] may be compared to, [2.C.sub.1] and [2.C.sub.2], the classical Mooney-Rivlin coefficients (ref. 8). The coefficient, [C.sub.10], is sometimes taken as an indicator of the degree of crosslinking.

In the linear crack propagation region, the cut growth per cycle according to Gerspacher and Borowczak (ref. 9) can be approximated to a power law of tear energy as given: dC/dN = 1/G [( T - [T.sub.o] ).sup.n]/( [T.sub.1]) (3) where, dC/dN is the cut growth rate (CGR) in cm/Mc, (Mc = megacycle), G is the cut growth constant in Mc/cm, n is the tear energy exponent (no units), T is the tear energy at applied deformation in N/cm, and [T.sub.1] is a statistically determined constant in N/cm. The cut growth rate and tear energy are experimentally determined. Thus, the cut growth equation 3 is a dimensionally homogeneous equation. In the present work, the parameters G and n in equation 3 are considered to be [greater than] 0. This is because the CGR is always a positive quantity and if G is [less than] 0 then CGR has a negative algebraic sign implying a healing process of the cut. In order to reduce the CGR it is desired to maximize the cut growth constant, G, and minimize the tear energy dependency, n.

The equation 3 is not different than the empirical equation 4 widely used in the literature. dC/dN = constant * ([Delta]T)sup.m] (4)

The threshold tear energy, [T.sub.o], for the present work is assumed to be equal to zero since a measurable cut growth rate was observed at the lowest strain of 10% studied. Further it is also assumed that, [T.sub.1], the empirical energy parameter is equal to 2 N/cm, at a 95% confidence interval (ref. 10). The equation 3 becomes, dC/dN = 1/G [( T )sup.n]/( 2 ) (5)

To obtain n and G the equation 5 is transformed into natural logarithmic form, and [In (dC/dN)] versus [In (T/2)] is plotted. Fitting the experimental data using the linear regression model, the slope of the regression line gives the tear energy exponent, n, and the intercept gives the cut growth constant, G. For two antidegradant systems, A and B, under conditions of identical deformations and tear energy dependence it can be shown that, [(CGR).sub.A] = ([G.sub.B]/[G.sub.A]) * [(CGR).sub.B] (6)

Experimental-Rubber compounds

Mixing

All rubber compounds studied were obtained from a masterbatch prepared in a conventional internal mixer. Incorporation of sulfur, accelerators and antidegradants was done according to standard compound mixing practices on a laboratory mill (300 x 200 mm). The mixing conditions used were: prewarm temperature (40 [degrees] C), typical batch size (1 kg), total mixing time (7 min) and stock temperature (75-80 [degrees] C).

Respective sulfur type and level, accelerator level and antidegradant type and level used are given in the data tables 3-7.

Properties measurement

The cure characteristics (rheometer) of each compound were determined according to ASTM D 2084. Tensile properties of vulcanizates were determined as per ASTM D 412. Heat aging of test specimens was performed according to ASTM D 865 (test tube method, using a Scott Model LG aluminum block).

Molding of pure shear test specimens

Each rubber compound was first sheeted to ca. 2.5 mm thickness over a cold roll mill as quickly as possible. Pure shear test specimens were then molded with the grain direction along the length of the mold and cured to their respective [t.sub.90] at temperatures indicated. The specimens were examined carefully for flaws and thickness (1 [+ or -] 0.05 mm). The flash on the specimens was carefully removed to minimize errors in the effective extension ratio due to distortion when the samples were clamped.

Cut growth rate measurement

A detailed description of the procedure is given in the literature (ref. 5). A summary of the important steps is given here.

The theoretical extension ratios of the cams used in the present study ranged from 1.10 to 1.55 as given in table 2. An illustration of a cam is given in figure 1. The desired extension ratio was achieved by an appropriate geometry of the cams made out of fiber glass reinforced laminate. The cams were calibrated for each blank vulcanizate by mounting a conditioned sample free from any slack and manually raising it to each high point ([H.sub.i]) and bringing back to each low point ([L.sub.i]) of the cam. The ratio of the mean [H.sub.i] to mean [L.sub.i] gave the calibrated value for each elastomer as given in table 2. Test specimens were mounted along the length as shown in figure 1, and to break in the compound each specimen was run prior to cut-growth measurements for 100 cycles (frequency 4.7 [s.sup.-1]) using an E.R. of 1.55 (Cam G or Universal Cam given in table 2). After the cam of a given extension ratio is mounted a cut of fixed length (25 [+ or -] 0.1 mm) is made into the specimen with a sharp pair of scissors. The specimens were then clamped and any slack present was removed by adjusting the movable clamp. The specimens were next cycled and the cut growth (2 runs in each case) was recorded every 30 minutes over a period of 5 hours. The cut length for E.R. 1.10 to 1.30 was measured using a traveling microscope. While the cut length at higher E.R. was measured using a thin paper ruler approaching as close as possible to the actual crack. An auto shutoff timer was used to turn off the cycling at every 30 minutes.

Table : Table 2 - calibrated extension ratios for NR (series 100), SBR (series 200) and NR/BR (series 300) in pure shear cut growth study

Cam           Extension ratio (E.R)
      Theory          Experimental
                NR        SBR    NR/BR
A(*)   1.10    1.11      1.17     1.17
B(*)   1.20    1.25      1.31     1.33
C(*)   1.30    1.32      1.34     1.39
D      1.35    1.38      1.42     1.49
E(*)   1.40    1.39      1.44     1.45
F      1.45    1.45      1.53     1.52
G      1.55    1.68      1.72     1.82

Pure shear specimens were conditioned with Cam G for 100 cycles except where noted (*) - indicates Cams with identical low points

The cams for upper and lower clamps were always mounted with their low points coinciding so that when a cycle is stopped for measurement at low point, both the upper and the lower clamp specimens are under identical strain. The clamps were machined precisely to clamp pure shear specimens along the rib leaving the wide test portion of the pure shear test specimen free from undesired stresses (see cross-sectional view in figure 1).

The tear energy measurements were made according to the method reported by Rivlin and Thomas (ref. 6). Pure shear specimens were used to measure the strain energy and the Mooney-Rivlin coefficients [C.sub.10] and [C.sub.20]. The tear energy was then calculated using equations 1 and 2.

Results and discussion

The present work studied pure shear cut growth behavior for antidegradants (table 1) in NR, SBR and model NR/BR (50:50) sidewall compounds. Also studied was the DeMattia crack growth in a model NR earthmover tread compound in an attempt to correlate the results with a pure shear cut growth approach. Conventional tensile properties were measured in each case. [Tabular Data Omitted]

Cursory study of antidegradants in NR

This study was undertaken to select candidate antidegradants from those given in table 1. The E.R. employed for CGR measurements are ranging from 1.35 to 1.55 (cams D, E, F and G in table 2).

Effect of sulfur level

Table 1 gives the CGR parameters for 1, 2 and 4 phr (compound no. 18, 1 and 19) S-level. The catastrophic TE ([T.sub.c]) decreases, as expected, with an increase in S level and in crosslinking density ([C.sub.10]). The compound with 4 phr S cracked completely and instantaneously at the 1.35 E.R. A level of 2 phr S was used as the reference.

Effect of phenolic (BHT, BPH) antidegradants

The cut growth constant (G) and [T.sub.c] for BPH is greater than that of BHT. This suggests that, to minimize cut growth, hindered bisphenols are effective. However, the exponent, n, increases for BPH suggesting a stronger TE dependence of CGR. BPH also shows CGR improvements compared to BHT at lower TE and at higher TE it shows a crossover. Both antidegradants minimized cut growth at low TE (or E.R.) based on extrapolation using a linear regression model.

Effect of amines (ODPA, PAN)

The [T.sub.c] for PAN is greater than that of ODPA. This is not surprising, since it has been shown that PBN increases the critical tearing energy by about 50% in unfilled NR (ref. 3). The CGR behavior for both, however, was identical with the TE dependency exponent n being less than 1. Reduction in CGR at high TE for both is noteworthy. Therefore, PAN and ODPA were chosen as the candidate antidegradants for further evaluation. It was anticipated that PAN will show much better improvements in cut growth minimization when used in blends with MMBI.

Effect of p-phenylenediamines antidegradants

The [T.sub.c] and n increase in the order "mixed diaryl PPD" [greater than] 6PPD [less than] 77PD. The CGR for all three PPD's was lower at low TE and a noticeable shift in threshold TE can be observed. 6PPD was chosen as one of the candidate antidegradants for further evaluation because of high G and medium n values. Besides, 6PPD is a widely used antiozonant.

Effect of imidazole (MMBI) and quinoline (TMQ)

The [T.sub.c] for MMBI is greater than that for TMQ and almost equal to that of PAN and 6PPD, while [T.sub.c] for TMQ is the same as that of ODPA. Good improvement in CGR at low TE was observed with MMBI. Therefore, MMBI was chosen as the candidate for further evaluation. TMQ showed a cut growth behavior parallel to that of the blank.

TE exponent (n) vs. cut growth constant (G)

Figure 2 shows that the slope (n) vs intercept (In G) follows a linear relationship. An attempt to divide antidegradants in this graph into domains of chemical classes failed (e.g. overlap of phenolics and ppd's).

Candidate antidegradants

The candidate antidegradants chosen from the cursory

PHOTO : Figure 2 - slope vs. intercept for antidegradants investigated study are: MMBI, PAN, ODPA and 6PPD. The results discussed below are further evaluation of these candidates in NR (series 100), SBR (series 200), NR/BR sidewall (series 300) and NR earthmover tread (series 400).

Evaluation of MMBI, PAN, ODPA and 6PPD in NR (series

100)

The ER used for this study range from 1.10 to 1.55. The lower than 1.35 ER were chosen in order to obtain the threshold TE, [T.sub.O].

The NR masterbatch for this and the cursory study was not absolutely identical in spite of using the same ingredients. This is seen from the cure characteristics of the blank compounds and catastrophic TE measurements (table 8).

Table : Table 8 - comparison of NR blank compound

Study         t2      t90      [T.sub.c]    n       G
             (min)   (min)   (N/cm)              (Mc/cm)
Cursory        6.2    24.5      156        1.6         5
Evaluation     5.2    19.0      280        1.6         3

Series 100

The physical properties, however, for both are comparable. The differences in properties can probably be attributed to the types of impurities in the NR used. The CGR parameter, n, giving the TE dependence is identical for both masterbatches and the cut growth constant, G, is higher in the cursory study. Nevertheless, compounds in the cursory and evaluation studies can be considered comparable. Upon heat aging for 14 days at 80 [degrees] C the NR blank showed a less favorable CGR behavior compared to the unaged samples.

Effect of antidegradants on modulus

It is known that the antidegradants affect the cure rate of a compound giving vulcanizates of different tensile moduli (stresses) for an optimum or [t.sub.90] cure. In the range of elongation of interest for the present work, the moduli for an NR blank are usually lowered upon adding an antidegradant. However, for 6PPD and MMBI the moduli are considered to be the same (within experimental errors) in the elongation range of 10-70%.

Effect of single antidegradants 6PPD, MMBI, ODPA and

PAN

The ranking of unaged [T.sub.C] compared to blank is ODPA [is greater than] PAN [is greater than] MMBI [is greater than] 6PPD [is greater than] blank. Upon aging the highest [T.sub.C] was observed for MMBI. From values of the Mooney-Rivlin coefficient, [C.sub.10], MMBI gave a vulcanizate with lower crosslinking density, while others are of comparable crosslink density. This suggests that MMBI is interfering with the crosslinking reaction.

Unaged 6PPD showed improvements in CGR at lower TE and a crossover of the blank in the catastrophic region (figure 3). Aged 6PPD showed similar improvements except for an earlier crossover.

MMBI, both unaged and aged, showed overall CGR improvements by increasing the TE requirements for cut growth to occur compared to the blank without a crossover.

Unaged ODPA yielded CGR improvements at lower TE with a crossover in the catastrophic region compared to the blank. Upon aging an overall CGR, improvements compared to the blank are observed.

PAN produced CGR improvements both before and after aging. The TE requirement to propagate a crack in the presence of PAN compared to the blank is significantly increased.

Effect of binary antidegradants (MMBI+PAN) and

(MMBI+ODPA)

Unaged (MMBI+PAN) showed CGR improvements which upon aging were significant. (MMBI+ODPA) showed such significant improvements as well. The [T.sub.c] for (MMBI+PAN) is greater than (MMBI+ODPA) and higher than MMBI alone but comparable to PAN. Once again, due to the presence of MMBI the coefficient [C.sub.10] is decreased.

Effect of ternary antidegradants (MMBI+PAN+6PPD) and

(MMBI+ODPA+6PPD)

Both showed higher [T.sub.c] and similar unaged and aged CGR improvements compared to the blank. The coefficient [C.sub.10] is again decreased due to MMBI and is comparable to that of MMBI alone.

Effects of MMBI

With MMBI the cut growth resistance improvement after heat aging and in presence of amine antidegradants was significant. This is in agreement with the well known fact that MMBI shows synergism with amines (ref. 11) and other antidegradants.

Effect of conditioning cam on cut growth rate

Table 5 gives the CGR parameters for the blank, (MMBI+PAN) and (MMBI+ODPA) for the cut growth study performed by conditioning pure shear specimens at E.R. 1.55 referred to as "universal cam" condition (using universal cam G for all E.R.s studied) and at the respective E.R. referred to as "cam of test" condition for which measurements were being made. Each cam was calibrated for both conditions and the differences found for calibrated cams were within experimental errors (i.e. E.R. + .0045). Therefore, no differences in cut growth behavior are to be expected.

The reproducibility of the method is excellent if one compares the cut growth parameter TE exponent, n, obtained using the universal cam condition (table 5) with that in table 4 for the blank. Slight differences in values of cut growth constant, G, are related to changes in the NR masterbatch with storage (ca. 1 year). Good agreement of the cut growth parameters for both the universal cam and the cam of test conditioning is observed.

Binary blends of MMBI with amines again show an improvement in cut growth resistance by increasing the cut growth constant, G, and by increasing the catastrophic tear energy, [T.sub.c].

Evaluation of MMBI, PAN, ODPA and 6PPD in SBR

(series 200)

Effect of antidegradants 6PPD, MMBI, ODPA and PAN

(single, binary and ternary)

The ranking of [T.sub.c] for unaged samples is MMBI[is greater than]/=ODPA = PAN [is greater than] 6PPD = blank. The [C.sub.10] values are comparable and unlike in NR, no decrease in the coefficient due to MMBI in SBR was observed. Interestingly upon aging the [C.sub.10] values increase and the [T.sub.c] decreases, while the cut growth constant, G, is twice that of the unaged values with a very little change in TE exponent, n (figure 4). It is to be noted that upon aging MMBI improves at low TE and shows a crossover phenomena at higher TE. Also, the unaged and aged behavior of PAN containing samples is nearly the same as seen in the case of NR above.

The unaged mixtures of MMBI binary with ODPA and Pan and ternary with 66PD show the CGR improvements at low TE are similar to that observed in case of NR above. The aged behavior for all mixtures in the case of SBR was poor compared to the unaged. The [T.sub.c] values for the mixtures are comparable to one another.

Evaluation of MMBI, PAN, ODPA in NR/BR (series 300)

This evaluation was based on a proprietary realistic model sidewall compound. All antidegradant mixtures tested here are with a constant 3 phr level of 6PPD, to infer the necessary ozone protection required in practical compounds.

The cut growth behavior of MMBI at two different levels, 1 and 2 phr (figure 5), shows that compared to the blank, MMBI is effective in increasing the TE requirements for cut growth to occur. This effect seems to increase with the level of MMBI; however, greater levels than 2 phr were not investigated.

For PAN the cut growth behavior is improved compared to the blank and this improvement decreased with increasing level (figure 12).

For MMBI mixtures with PAN the cut growth improvements compared to the blank are significant (figure 13). The improvements increased with levels of both MMBI and PAN, viz. 1 phr each along with 6PPD at 3 phr. The cut growth constant, G, for [MMBI (1 phr)+PAN (1 phr)+6PPD (3 phr)] is higher than that for MMBI and PAN individually at levels of 1 and 2 phr (table 7). Thus, pure shear results are in agreement with the synergism phenomena known for mixtures of MMBI and amines. The optimum levels for the mixtures of MMBI and PAN are 1 phr each.

Two levels for ODPA, 2 phr and 1 phr in mixture with MMBI were studied (figure 14). The MMBI mixture and ODPA individually show similar improvements in CGR compared to the blank.

Tear energy exponent, n

The tear energy exponent, n (equation 5) obtained in the present study for the NR, SBR and NR/BR blanks is summarized and compared with the exponent, m (equation 4) reported by Clamroth and Eisele (ref. 12) for filled elastomers (table 9).

Table : Table 9 - comparison exponent n, m

Elastomer                 Exponent
               n (equation 5)    m (equation 4)
               (present work)    (ref. 11)
NR (blank)     1.6               2.0
SBR (blank)    2.1               2.4
NR/BR (blank)  2.2               -
BR             -                 3.0

The values obtained in the present study are in good agreement with the reported values. This indicates that the empirical equation 4 used by others and the equation 5 used here are comparable with the advantage that equation 5 is dimensionally homogeneous.

Evaluation of MMBI, PAN, ODPA

in NR (series 400)

This evaluation was based on a proprietary model earthmover tread compound. All antidegradant mixtures tested were used at a constant 3 phr level of 6PPD for the reason given earlier. The purpose of this study is to compare the traditional DeMattia crack growth results with the pure shear findings.

The crack growth data for 4 runs (except 6PPD aged-2 runs) for unaged and aged antidegradants considered are fitted using a second order polynomial regression. The equation used is: Y = [A.sub.O] + [A.sub.1] * X + [A.sub.2] * [X.sup.2] (7) where Y is the crack growth in %, X is the number of cycles in kc and [A.sub.i] are the polynominal coefficients. The crack growth improvements observed for MMBI mixtures with PAN are similar to those observed for NR, SBR and NR/BR in the pure shear study. For a clearer comparison the crack growth for these systems is calculated at 300 kc (a midpoint of the fatigue life span covering approx. 0-600 kc) using the coefficients given in table 10. The calculated unaged and aged crack growth is shown in table 10. The calculated unaged and aged crack growth is shown in figure 6. 6PPD at 5 phr, both unaged and aged, offers better crack growth resistance than all the other systems considered here. This is in agreement with the results found using the Du Pont flex test (ref. 13). PAN unaged is of no advantage but upon aging it reduces the crack growth compared to 6PPD at 3 phr. MMBI unaged is similar to 6PPD at 3 phr and improves upon aging, while the (MMBI+PAN) mixture, as discussed above in pure shear crack growth, shows significant improvements in both unaged and aged conditions. [Tabular Data Omitted]

Possible explanation of cut growth reduction by MMBI

MBI reduces cut growth propagation as seen above individually or in mixtures with amines. The explanation of such distinguishing property can be obtained by extrapolating the findings of Lukovinkov et al. (ref. 14). They have reported the effect of benzimidazole as inhibitors of the oxidation of poly(propylene). Figure 7 shows a possible scheme of MMBI attachment to an elastomer. If the predominant attachment is that of structure III then the abstraction of the thiol proton by a radical during oxidation process gives a sulfur radical. This is also favored considering the lower bond strength of S-H (368 kJ/mol) compared to that of the [is greater than]N-H (391 kJ/mol). When two such radicals come together disulfide linkages (-S-S-) with a bond strength of 269 kJ/mol are formed. This is similar to a healing mechanism. Such disulfide bonds may resist the cut growth explaining that in the presence of MMBI the Mooney-Rivlin coefficient, [C.sub.10], was decreased and this leads us to believe that MMBI is attached to the chain. The increased efficiency of MMBI upon aging supports the increased crosslinking and disulfide bond formation. This is not surprising as it has long been realized that the thiols are better chain transfer agents (or free radical scavengers) than amines exhibiting high chain transfer constants.

Conclusions

Pure shear cut growth in the present study is not found to be affected by the specimen conditioning E.R. The findings obtained by this method are in agreement with DeMattia crack growth. The advantage of the pure shear CGR method over the conventional methods is that the pure shear method offers the cut growth parameters as opposed to the qualitative ranking. Also, testing times are less than half of that required for the conventional methods.

The E.R. studied in the present work ranges from 1.10 to 1.55. The strains ranging from 10% to 55% are important as rubber components undergo such strains during their service. The fatigue mechanism in this range of strains is primarily that of an oxidation based on the comparison of the CGR behavior of the blank with that of the compound with the antidegradants. Therefore, it is important that besides the selection of an antiozonant attention is paid towards selecting proper antioxidants. Unequivocal cut growth reduction in NR, SBR and NR/BR is obtained by MMBI, PAN and ODPA compared to other antidegradants. MMBI offers fatigue improvements by itself and the improvements are enhanced by using mixtures of it with PAN, ODPA and 6PPD. The dynamic properties and the ozone resistance (data not presented here) were not affected by these antidegradants. The cut growth reduction efficiency of the antidegradant systems parallels that of the catastrophic tear energy increase.

It is concluded here that MMBI, PAN and ODPA are potential antidegradants for improving fatigue-fracture resistance of tire sidewalls and tread groove regions.

References [1.] M.J. Doyle, "Fatigue crack growth in elastomers," presented at the Education Symposium "Fracture & failure of elastomeric materials" of the Rubber Division, ACS Meeting, Indianapolis, IN, May 8-10, 1984. [2.] G.R. Hamed, "Fracture of rubber vulcanizates," presented at the Education Symposium "Fracture & failure of elastomeric materials" of the Rubber Division, ACS Meeting, Indianapolis, IN, May, 1984. [3.] G.J. Lake and P.B. Lindley, J. Appl. Polymer Sci., 9, 1233 (1965). [4.] A.N. Gent, J. Appl. Polymer Sci., 6, 467 (1962). [5.] A.G. Thomas, J. Polymer Sci., XXXI, 467 (1958). [6.] R.S. Rivlin and A.G. Thomas, J. Polymer Sci., X, 291 (1953). [7.] D.G. Young, Rubber Chem. Technol., 58, 785 (1985). [8.] L.R.G. Treloar, "The physics of rubber elasticity," pp. 213, Clarendon Press, Oxford, UK, 3rd Edition, 1975. [9.] M. Gerspacher and M. Borowczak, "Fatigue limit determination," paper no. 63, presented at the Rubber Division, ACS Meeting, Montreal, Canada, May 26-29, 1987. [10.] G.E.P. Box, W. G. Hunter and J.S. Hunter, "Statistics for experimenters," pp. 473-8, J. Wiley, New York, NY, 1978. [11.] B.N. Leyland and R.L. Stafford, Trans. Inst. Rubber Ind., 35, 25 (1959). [12.] R. Clamroth and U. Eisele, Kautsch. Gummi, Kunstst., 28, 433 (1975). [13.] T. Kempermann and W. Redetzky, "Investigations concerning the long term effects of antidegradants," Technical Report, Bayer AG, W. Germany. [14.] A.F. Lukovinkov et al., Vyokomol. Soedin, 5, 1785 (1963); Chem. Abst., 60, 10876c (1964).

PHOTO : Figure 3 - series 100: NR + 6PPD

PHOTO : Figure 4 - series 200: SBR + 6PPD

PHOTO : Figure 5 - series 300: NR/BR + MMBI + 6PPD

PHOTO : Figure 6 - DeMatta crack growth at 300 kc calculated using 2nd order polynomial

PHOTO : Figure 7 - relative antidegradant efficiency of [is greater than] NH and -SH groups