Dimensionally stable PET fibers for tire reinforcement.

Dimensionally stable PET fibers for tire reinforcement

High-modulus, high-tenacity polyethylene terephthalate (PET) fibers have gained wide acceptance in reinforcement for rubber products such as tires and hoses; in geotextiles, and in ropes and cordage. Accordingly, a great deal of fundamental research has been conducted on how the processing and resulting morphology of these thermoplastic fibers affect their physical properties[1-9]. The translation of these starting fiber properties to those in the final end-use product has received much less attention.

This article compares the structure-property relationships of recently-developed PET yarns possessing high dimensional stability (i.e., high modulus and low shrinkage) with conventional PET tire yarn. The yarns under investigation are new fibers developed and manufactured by Allied Fibers called DSP. If tire cord performance were predicted solely on the basis of yarn mechanical properties, then DSP fibers would not have gained their current commercial acceptance because their yarn strength is at least 10% lower than conventional tire yarn. However, the dimensionally stable products, particularly their advanced versions - two of which have recently been introduced commercially - exhibit a more efficient strength conversion into treated cord. Thus, yarn properties could be used for quality control, but not for predicting treated cord properties. Therefore, it is prudent to go one step further and base tire design on the properties of the cord in the tire (in-tire properties). This article will identify some yarn features that give an indication of superior treated cord properties.

Dimensionally stable yarns initially gained commercial acceptance in tires because of their ability to reduce sidewall indentations (SWIs) in monoply radial tires. This is a cosmetic problem in which a depression occurs at the body ply fabric splice. These SWIs form because, under inflation pressure, the higher cord density at the splice restricts tire cross-section expansion more than in the adjacent region, which has a lower cord density. All other things being equal, SWI depth decreases linearly as the number of overlapped cords in the splice decreases, becoming non-existent for butt splices (i.e., where there is no overlap). For tire safety, most manufacturers prefer to overlap the body ply.

There are presently two schools of thought as to why DSP cords reduce SWIs. Once faction believes that lower shrinkage per se reduces the differential body cord contour lengths between the splice and the adjacent non-splice region. This implicitly assumes that the higher cord density in the splice exerts a higher localized shrinkage force on the tire and hence more shrinkage occurs. The other faction believes that it is the higher cord modulus which restricts cord extension during inflation and hence reduces SWI depths. This study will show that in-tire modulus seems to be the governing factor.

Interest in these materials is further stimulated by their ability to improve tire uniformity and, for some tire manufacturers, eliminate the need for post-cure inflation during tire manufacturing. Where possible, cause and effect relationships will be developed.

Experimental procedure and testing

Background

The study was conducted in two phases:

* testing and analysis of the base yarns, and

* testing and analysis of subsequent treated tire-cord.

The first phase detailed the properties of three 1,000 denier polyester yarns manufactured by Allied Fibers: 1W70, a conventional PET fiber, as well as two DSP yarns, 1W90 and A330. The second phase studied the properties of tire cords made from these three yarn types plus A340, which is the most dimensionally stable of these products. Please note that the 1X90 product used in treated cords is equivalent to 1W90 with the exception that 1X90 has improved polymerization conditions and 0.03 higher intrinsic viscosity (IV).

Phase 1 - testing of base yarns

Tensile measurements were performed using a Model 1125 Instron tester at a strain rate of 120% min. Initial modulus, ultimate elongation (UE) and tenacity measurements were taken according to standard testing procedures.

Shrinkage was determined under loads of 5 to 100g at 177 [decrees] C.

Measurements were taken at "equilibrium," which occurred after approximately one minute. Retractive force values were taken as the maximum of the retractive-force versus time curve at 177 [degrees] C.

Density ([Rho]) was measured in an n-heptane/carbon tetrachloride density gradient column at 23 [degrees] C. Crystallinity using density analysis ([X.sub.d]) was subsequently calculated according to ref. 10 using 1.335 and 1.495 g/[cm.sup.3] for the amorphous and crystalline densities, respectively[11,12]. Thermal analysis was performed with a Perkin-Elmer DSC-7 equipped with a data station. Melting points ([T.sub.m]) were taken as the peak of the melting endotherm. Small samples(2mg) were used for [T.sub.m] analysis for enhanced resolution and accurate temperature measurement. Large samples were tested for heat-of-fusion characterization. Crystallinity using thermal analysis ([X.sub.c]) was calculated assuming a crystalline heat of fusion of 140 J/g[13].

Wide-angle x-ray scattering was used to measure crystallite sizes (L) employing a Phillips vertical goniometer with [CuK.sub.[Alpha]] radiation generated at 45 kV and 35mA. Resolution of overlapping peaks was accomplished with a profile-curve-fitting program provided by Phillips. Small-angle x-ray scattering was used to measure long periods (l) employing a Franks' camera and a 1-D position-sensitive detector. Crystalline orientation values needed for computation of amorphous orientation were obtained for an azimuthal scan of the 105 peak.

Birefringence ([Delta]n) measurements were made with a Leitz Model NR.5533 polarizing microscope using a 23-order berek compensator. Amorphous orientation ([f.sub.a]) was calculated using crystal and amorphous intrinsic birefringences of 0.220 and 0.275, respectively[14].

Sonic modulus measurements were made on a Morgan dynamic modulus tester Model PPM5 at a frequency of 5kc/sec.

Phase 2 - tire cord preparation and testing

Unless otherwise indicated, tire cords were prepared by twisting the yarns into an 8 tpi (turns-per-inch) by 8 tpi, 1,000d/3 or 1,500d/2 construction and then applying a typical RFL adhesive system during tensilization. A two-zone stretch/relax oven tensilization was performed using 450 [degrees] F in both zones and 40/60 second residence times in the stretch/relax zones, respectively. Each cord was tensioned by first stretching to +6% and then relaxing 0, 2 or 4%. Since the total cord denier and hence twist multiplier are the same for 1,000d/3 and 1,500d/2, there are no significant property differences between these constructions.

Tensile and shrinkage tests were performed in the same manner as used for base yarns.

The fatigue resistance of these materials was measured using a Goodrich disk fatigue tester according to ASTM D885-64T. In this test, the cords were imbedded in rubber and exposed to a compression/extension cycle, after which cord breaking-strength-retention was measured. For this study, the cords were strained +3%/-9% at 2,500 RPM for 48 hours.

Results and discussion: Base yarns

Mechanical properties

Representative stress-strain curves for the three yarns studied are depicted in figure 1 and data derived from these traces are summarized in table 1, along with representative shrinkage and retractive force results.

The conventional 1W70 yarn exhibits the highest initial modulus but the lowest LASE-5 (load at specified elongation of 5%). This indicates that it has the most non-linear stress-strain curve. To quantify this behavior, a stress-amplification ratio (SAR) was defined as the secant modulus at 5% elongation (i.e., LASE-5/0.05) divided by initial modulus. As indicated in table 1, SAR follows the order A330>1W90>1W70. Alternatively, one can describe the stress-strain response as instantaneous modulus-strain curves. This measurement serves to further emphasize the rank ordering. All the yarns show a minimum instantaneous modulus at about 2.2% strain. However, A330 shows the desired feature of a more nearly constant modulus at low strains. In fact, A330 has a 50% higher minimum instantaneous modulus than 1W70. If one attributes the reduction in modulus at 1-3% strain to yielding, then it seems reasonable to attribute the more linear behavior to a suppression of this yielding. Consequently, the advanced DSP yarns would exhibit lower hysteresis and have more recoverable strain. Such was found to be the case[15].

Tenacity exhibits a trend opposite to that of SAR, with 1W70 and A330 exhibiting the largest and smallest tenacities, respectively (table 1). Shrinkage at 177 [degrees] C as a function of constraining force shows a trend similar to tenacity. Therefore, A330 exhibits the paradoxical, yet desirable, characteristic of high LASE-5 and low shrinkage.

Although one would expect retractive force to mimic the shrinkage values, this parameter is relatively invariant among the samples (table 1).

Table 1 - tensile properties of PET fibers

Property                  1W70   1W90   A330
Tenacity (g/d)             9.2    8.1    7.9
Initial modulus (g/d)      112     98    100
LASE-5 (g/d)              3.62   3.67   4.14
SAR                       0.65   0.75   0.83
Ultimate elongation (%)   13.7   11.1   10.5
Shrinkage(*) (%)          12.5    9.6    7.5
Retractive force (g/d)    0.32   0.32   0.35

(*) 9g tension

Morphological analysis

The results of thermal analysis, dilatometry, birefringence, sonic measurements and x-ray scattering are given in table 2. Good agreement exists between the crystallinity values calculated from the two approaches ([X.sub.c] vs. [X.sub.d]), and the extent of crystallinity does not vary much from sample to sample, as expected of highly drawn fibers.

Table 2 - morphological analysis

Parameter                  1W70     1W90     A330
[T.sub.m] ([degrees] C)    256.1    257.5    259.1
[X.sub.c] (%)                 35       37       38
[Rho] (g/cc)              1.3898   1.3857   1.3864
[X.sub.d] (%)                 37       34       35
[Delta] n                  0.232    0.203    0.197
[f.sub.a]                   0.88     0.71     0.68
1 ([Angstrom])               136      113      106
L010 ([Angstrom])             57       54       57
L110 ([Angstrom])             52       45       47
L100 ([Angstrom])             43       42       42
L105 ([Angstrom])             83       74       80

Figures 2-5 depict the effect of [f.sub.a] on shrinkage, tenacity, sonic modulus and LASE-5. Included in these figures are data for a series of 1W90 yarns prepared at different draw ratios. These additional results are included to enhance the accuracy of the trends.

Both tenacity and sonic modulus increase with [f.sub.a] as expected, and the results for the three yarn types fall along the same regression line. The similar dependence of sonic modulus on [f.sub.a] for all the samples is attributed to the "conventional" dependence of modulus on amorphous orientation at very low strains.

This contrasts with the LASE data of figure 5, which indicates that both A330 and 1W90 exhibit an elevated LASE-5 at a given [f.sub.a] when compared to "conventional" 1W70. This observation is analogous to the previously mentioned stress-amplification.

Shrinkage follows the historical trend of increasing with increasing [f.sub.a]. Unfortunately, no data were available to establish a relaxation line for a given base yarn. However, the rate of change of shrinkage with [f.sub.a] is similar to that found for conventional yarns in other studies.

Up to this point the tenacity, initial modulus and shrinkage characteristics of the fibers have been explainable in terms of classical morphological analysis. However, an important question remains: Why does A330, and to a lesser extent 1W90, possess significant stress-amplification? Clearly, some morphological characteristic, the exact nature of which is unknown at present, is giving rise to this stress-amplification. Perhaps the interfibrillar regions play a role in polymer yielding. Research is continuing to clarify the origin of this interesting phenomenon.

Treated cords

Background

All the reported data in this section are for 1,000/3 or 1,500/2, 8x8 tpi cords subjected to two-zone tensilization at 450/450 [degrees] F using a high stretch (+6%) followed by different relaxations (0, 2 and 4%) to vary the net stretch. Summary of the mechanical property data for cords at 4% free shrinkage is given in table 3. This shrinkage is taken to be representative of the shrinkage levels of the first generation DSP treated cords used in tires. During the course of this study, a variety of other treatments (higher temperature, one-zone tensilization, and both lower and higher stretch) were evaluated. The general conclusions of this study are applicable to a fairly wide range of tensilization conditions and to other cord constructions. Of course, the absolute magnitude of the properties will differ.

Dimensional stability

The tire industry is very much interested in PET tire cords of high modulus and low shrinkage because of the strong influence of these properties on tire performance. By plotting LASE-5 vs. shrinkage for cords tensilized to different net stretches, we can quantify dimensional stability by measuring LASE-5 at a given free shrinkage. Such a plot (figure 6) shows that dimensional stability, as measured by LASE-5, follows the order 1W70<1X90<A330<A340. For ease of comparison, the data of figure 6 can be interpolated to 4% free shrinkage as in table 3. Relative values for this dimensional stability representation derived as an average of multiple trials are shown in table 4, along with similar extrapolated values for initial modulus, minimum modulus, LASE-3, and elongation at 15 lbs. (E-15). All these property-shrinkage curves were quite linear and indicated similar trends. However, the relative differences were somewhat compressed at the lower strains. Since some tire companies use the sum of E-15 plus ASTM (constrained) shrinkage as a measurement of dimensional stability, this parameter was also included in table 4.

Table 4 - dimensional stability ranking (at 4% free shrinkage)

Initial Min. E-15+

mod. L-3(*) L-5(*) E-15(*) [Delta][Epsilon](*) mod. ASTM(*)

1W70   100    100    100    100        100              100   100
1X90   125    138    140     60         72              174    71
A330   135    154    160     52         65              204    66
A340   161    177    190     42         56              230    63

(*)Key L-3 = LASE-3 L-5 = LASE-5 E-15 = Elongation at 15 lb. [Delta][Epsilon] = difference in cord elongation ASTM = ASTM or constrained shrinkage

The preceding brings up the question of what is the best representation of tire cord dimensional stability? To some extent, the answer must depend on the application. For SWI, it relates to the difference in cord elongation ([Delta] [Epsilon]) for loadings representative of the ply and splice regions. This is schematically illustrated in figure 7 for a typical monoply tire. In a monoply tire, the splice region essentially has two plies and hence, for large splices, the loading per cord is roughly half that of the single ply region adjacent to the splice. As shown in figure 7, the [Delta] [Epsilon] for 1W70 is greater than that for A340. Relative ranking based on [Delta] [Epsilon] is also given in table 4 and is most similar to the LASE-5 case. Tire handling should be related to similar differential elongations at the stress levels prior to and during maneuvering.

A less used, but possibly more sensitive measurement of dimensional stability is instantaneous modulus. Qualitatively, the modulus-strain curves for treated cord are similar to the base yarns. Because one can vary net treating stretch, the minimum modulus can be markedly changed for a given tire cord type. The most meaningful comparison between materials would be the minimum instantaneous modulus at a given free shrinkage. Such a relative comparison is give in table 4. Also, a plot of minimum instantaneous modulus versus LASE-5 gave a single curve which covered all data points for all treated cords.

Figure 8 gives the treated cord SAR as a function of shrinkage. This indicates that more linear stress-strain curves are obtained at higher treating stretches. For a given shrinkage, treated cord SAR follows the same trend as noted for the base yarns and for the treated cord LASE-5.

The observed dimensional stability trend for these materials reflects and parallels the previously measured properties of the starting yarns. Most tire manufacturers compare the dimensional stability of materials under constrained rather than free shrinkage. The dimensional stability ranking of these tire cords is independent of the constraining force. However, the absolute value of the difference between the materials is reduced at a higher constraining force. We have focused on LASE-5 in the remainder of this article because it (a) can be readily and accurately measured in most routine testing programs and (b) also provides a reasonable measure of relative differences in minimum instantaneous modulus and SAR.

Tenacity and tenacity retention

Maintaining high cord strength is important for minimizing cord weight per tire, and hence cord cost. Tenacity is plotted versus free shrinkage in figure 9, and values for the four cords, interpolated to 4% free shrinkage, are provided in table 3. These results indicate the following rank order with respect to tenacity: [Mathematical Expression Omitted].

The fundamental reason for this ranking is uncertain, primarily because of the complex morphological origin of tenacity. It should be noted that most tire companies would rank 1W70 tenacity between 1X90 and A330, because the conventional 1W70 material is generally treated to 8% rather than 4% free shrinkage.

From a practical standpoint, it can be concluded that A330 possesses better dimensional stability than cords made with first generation dimensionally stable fibers, and that A330 has the excellent tenacity historically associated with the conventional products.

An important characteristic of dimensionally stable yarns is their high retention of tenacity following conversion to treated cord. This behavior is summarized in table 5 for cords treated under +6%/-2% stretch/relaxation conditions. Although the exact origin of this behavior is unknown, the data in table 5 show that tenacity retention clearly increases with dimensional stability.

Table 5 - yarn to treated cord conversion for cords treated at +6/-2%

               Tenacity                   Tenacity
                 g/d                        g/d
1W70                      A330
  Yarn           9.2         Yarn           7.9
  Cord           6.8         Cord           7.2
  % Retained     74          % Retained     91
1X90                      A340
  Yarn           8.1         Yarn           6.8
  Cord           6.8         Cord           6.4
  % Retained     84          % Retained     94

Improved toughness

Tire manufacturers also desire reinforcing materials which will yield tires with superior resistance to sidewall bruising and high speed impact. Improvement in these areas requires cords with increased toughness and ultimate elongation (UE)[17]. Since toughness can be increased for a given material by increasing relaxation and hence reducing LASE-5, comparison of inherent differences in materials is accomplished by plotting UE or toughness as a function of LASE-5 (figure 10). The figure indicates that A340 provides greater toughness for a given level of LASE-5, suggesting that A340 cords will provide the improved toughness required for bruise resistance. A case can be made that the enhanced toughness exhibited by A340 more than offsets the cord's low tenacity, thus actually improving tire durability.

Fatigue testing

Results of the Goodrich fatigue test for cords stretched +6%/-4% during tensilization are given in table 3. These values are typical for high tenacity tire cords tested according to the procedure described in the experimental section. In a running tire, hysteretic heat generation exerts a strong effect on tire temperature which in turn exerts a strong negative effect on fatigue life[18] because tire rubber thicknesses are sufficient to contain the generated heat. Consequently, dimensionally stable yarns, which exhibit low hysteresis and low heat generation[19] are expected to possess superior in-tire fatigue behavior. This behavior is not totally reflected in the Goodrich fatigue data since its small test samples tend to lose more heat to their surroundings than similar cords in actual tires.

Dimensionally stable cords do not always exhibit superior fatigue behavior in experiments using the Goodrich test, particularly when the more dimensionally stable products are treated to give higher modulus. Since the disk fatigue test strains all specimens to the same extent, more energy is input into the high modulus cords.

Differences in LASE-5 achieved by changes in tensilization stretch have a greater impact on fatigue strength than does cord type. Since cords in tire sidewalls are more likely to experience a constant stress or energy deformation, the fatigue data in table 3 may best indicate the inherent difference in fatigue resistance, since they were obtained using cords with similar LASE values. We are currently investigating test procedures designed to more accurately reflect actual tire conditions.

Cords in tires

Background

It is highly desirable to estimate the in-tire cord properties after tire manufacturing, since it is these properties which control overall tire performance. The tire manufacturing process is quite complex involving (a) cord extension as the carcass is expanded into the mold, (b) cord shrinkage as the temperature rises in the uncured carcass and the cord shrinkage force causes the cord to pull through the body ply toward the inner liner, and (c) further cord shrinkage as the hot, partially cured tire is released from the mold. When using a post-cured inflation (PCI) step, the cord is then extended when the tire is subsequently inflated. PCI extension is greatest for higher inflation pressure and for higher tire temperatures associated with shorter time intervals between release from the mold and PCI. This article is concerned only with the case where there is no PCI.

Simulated in-tire LASE

Since the extension during tire curing is small and the temperature is well below that seen during cord treating, it seems reasonable to simulate this complex process by allowing the cords to shrink to equilibrium under a given "effective" constraining force at a typical tire curing temperature of 177 [degrees] C. For the typical non-PCI radial passenger tire manufacturing process, a constraining force between that used for ASTM (=0.05g/d) and free (=9.3g) shrinkage tests gives the same residual LASE-5 as cords removed from virgin tires. To separate the contribution of internal stress-relaxation from property reduction due to actual shrinkage, cords were also constant-length-annealed at 177 [degrees] C. Afterward the tensile properties were measured at room temperature.

For these in-tire simulations, the subject treated cords were of 11x11 tpi, 1,000/2 construction. The first trial assessed the ability to increase in-tire LASE for a given material merely by using higher treating stretches. This was accomplished by treating 1X90 to three different net stretches. As shown in figure 11, the residual LASE after both the ASTM and free shrinkage tests were the same for all three net stretches even though the starting cords had significantly different starting LASE. Of course, all ASTM samples had higher LASE than their free shrinkage counterparts. Since the cords annealed at constant length exhibited LASE values more representative of the starting treated cord than the ASTM or free shrinkage samples, it is concluded that shrinkage plays a greater role than internal stress-relaxation.

LASE-5 values for various cords originally treated to 4% free shrinkage and then subjected to the simulations are depicted in figure 12. These data indicate that the high LASE of the advanced DSP cords will translate to actual tires regardless of the effective constraining condition. Although the absolute values are lower, the relative values are similar to the starting treated cords compared at the same shrinkage.

Because tire manufacturers often tensilize different cords to a given LASE, simulated in-tire LASE is given for cords initially treated to a constant LASE-5. Under all constraining conditions (ASTM or free shrinkage), the advanced cords have superior in-tire LASE compared to the traditional cord and even to the first generation DSP product.

Sidewall indentation

SWIs are observed when the tire is inflated and the splice region does not extend as much as the non-splice region. Hence, we have derived a relationship (Appendix) which describes the SWI depth (D) in terms of tire geometry, inflation pressure and cord extension. For a given monoply tire, this relationship can be simplified to:

(Eq. 1) D = [r.sup.[Omicron]] ([Delta] [Epsilon]) where: [r.sup.[Omicron]] = uninflated tire cross-section radius; [Delta] [Epsilon] = difference in cord elongation at non-splice and splice tensions.

One can estimate relative differences in SWI due to materials by rationing the elongation differentials at the typical in-tire tensions;

(Eq. 2) D1/D2 = ([Delta] [Epsilon])1/([Delta] [Epsilon])2 where the subscripts 1, 2 refer to the two materials.

Using the appropriate elongation from the stress-strain curves for 1X90 and 1W70 removed from monoply radial passenger tires (where no PCI was used), one calculates a 21% SWI reduction for 1X90. Alternately, approximating D by assuming that the stress-strain curve had an average modulus equal to LASE-5/0.05 indicates a 20% SWI reduction for 1X90. Considering the limited available data, with its inherent variability, these computations compare well with the observed 32% SWI reduction, thereby supporting this SWI modeling approach.

The above results, based on cords from actual tires, is consistent with in-tire simulation data. For example, the "free shrinkage" simulation predicts a 17% SWI reduction for 1X90. Analogous predictions for A330 and A340 cords indicate 29% and 48% reductions, respectively, relative to 1W70. Since SWI improvements in actual tires seem to be even greater, our computations probably underestimate the SWI reduction.

The above analysis assumed that the cord lengths in the splice and non-splice regions were the same. This occurs when the cords develop high shrinkage force prior to rubber curing and hence the cords independently pull through the uncured rubber. For the more general case where this is not true, Eq. 1 would be written as

(Eq. 3) D = [r.sup.[Omicron].sub.ns] (1 + [[Epsilon].sub.ns]) - [r.sup.[Omicron].sub.s] (1 + [[Epsilon].sub.s]) where, [r.sup.[Omicron].sub.i] = uninflated tire sidewall radius for non-splice (ns) and splice (s) regions; [[Epsilon].sub.i] = elongation at non-splice (ns) and splice (s) tensions.

The most likely need to apply equation 3 would occur if the body skim either cured rapidly or merely had an ultra high modulus which prevented cord shrinkage. The resulting high residual cord retractive force in the splice region will cause the splice cord length to decrease relative to the non-splice cord length upon ejection from the mold. To determine the contribution of this mechanism, typical radial passenger tires (no PCI) were made using the 1X90 cord treated to 4% free shrinkage. Analysis of cords removed from these tires showed that the LASE-5 values in the splice were 9% lower than those in the adjacent non-splice regions. Based on the laboratory data for residual LASE after 0 (constant length), 2 and 4% shrinkage at the cure temperature, the cords in the splice shrunk about 0.5% more than the non-splice cords. Therefore,

[r.sup.[Omicron].sub.s] = 0.995 [r.sup.[Omicron].sub.ns] and Eq. 3 becomes

D = [r.sup.[Omicron].sub.ns] ([[Epsilon].sub.ns] - [0.995.sub.s] + 0.005)

Since ply cords in these tires elongate at least 2% at the inflation pressures used for SWI measurement, then the differential shrinkage contribution to SWI is small. To simplify the argument it is assumed that small length changes such as 0.5% do not permanently alter the cord morphology Thus, the lower modulus splice cords extend proportionately more, resulting in no significant contribution to SWI for cord extension well above 0.5%.

Since the preceding analysis employed room temperature stress-strain measurements on actual or simulated cords from virgin tires, it relates mostly to SWI measurement at the factory. SWI of tires that have been run is more difficult to predict, since the tire has experienced higher temperatures for extended periods. This is a subject for future work.

Tire uniformity

Reduction of the variation in radial deformation force during tire operation is a major goal of tire engineers. Since variation in cord spacing and cord modulus affect radial force variation, it follows that one desires cords which will undergo relatively small dimension and property changes throughout the tire manufacturing process.

Both the high modulus and the lower shrinkage associated with DSP cords are directionally correct for better tire uniformity. Still, the question arises as to whether it is better to tensilize a given cord under high or low stretch conditions. This choice in treating stretch is related to the ability to achieve equal (a) cord spacing through treating, calendering and tire building, (b) cord lengths, and (c) cord LASE. If the treated cord shrinkforce were too low, then there would be insufficient shrinkage to ensure that all the cords remained taut, resulting in wavy cords. Conversely, excessive treating tension results in a large differential between treated cord LASE and "equilibrium" in-tire LASE. Thus, mold segment differences in the "effective" constraining force during tire curing due to temperature and/or stock variation could result in different cord properties over a region sufficiently large to be detectable during tire uniformity testing.

The preceding implicitly assumed that the equivalent LASE-5 in figure 11 means equality in other properties such as stress at other strains and even creep. These measurements are currently being made.

Summary and conclusions

Tenacity, shrinkage and initial modulus of dimensionally stable yarns were shown to be primarily controlled by amorphous orientation, with increased orientation raising the parameters. However, these dimensionally stable yarns exhibit much higher stress (stress-amplification) at intermediate strains than expected on the basis of amorphous orientation. It is speculated that suppression of yielding mechanism may be responsible for the observed stress-amplification. Study of the interfibrillar regions is a reasonable starting point for future investigation.

Subsequently, the properties of cords made from such dimensionally stable fibers were studied to determine their translation into treated cord and implications regarding tire performance. Through testing and simulations of the tire manufacturing process, it was shown that the dimensional stability and stress-amplification trends of the yarns are maintained when they are treated and converted into cords. Based on these properties, we expect that the cords of advanced dimensional stability - A330 and A340 - can be used to manufacture tires with improved uniformity, bruise resistance and handling, as well as reduced sidewall indentations.

Table 3 - treated cord properties at 4% shrinkage

                       1W70   1X90   A330   A340
LASE-5 (g/d)            1.9    2.4    3.1    3.9
Modulus (g/d)            60     68     73     83
Tenacity (g/d)          6.4    6.8    7.1    6.4
SAR                    0.51   0.67   0.86   0.93

Retained(*)

Fatigue strength (%) 71 80 78 78 (*)+6%/-4% stretch/relax during tensilization. All materials had similar LASE.

[Figures 1 to 12 Omitted]

Appendix

Estimating sidewall indentations

Sidewall indentations depend on the cord properties and especially on the number of cords in the splice overlap. The analysis discussed below is for large splice overlaps where the splice essentially behaves as a tire built with the splice cord density. For smaller overlaps, this approach was found to give a good indication of relative SWI differences. 1. From the literature (ref. 20), cord tension in inflated tires is given by:

(1) T = [Pi]P([r.sub.c.sup.2] - [r.sub.w.sup.2])/N

T = tension per cord (lb. or N)

P = inflation pressure (psi or N/[m.sup.2])

[r.sub.c] = distance from wheel axle to crown (in. or m)

[r.sub.w] = distance from wheel axle to widest tire part

(in. or m)

N = total number of ply cords per tire 2. Cord elongation in the splice and non-splice regions can be estimated from the cord stress-strain curves using the stress equivalent to the tension on the cord in the tire. For the non-splice region this is calculated using equation 1 and the actual total cords per tire. Alternatively, one could establish an approximate relationship between cord tension and elongation using

(2) [Epsilon] = T/E where

[Epsilon] = Strain ([Delta]L/[L.sub.[Omicron]])

E = "effective" modulus (e.g., LASE-5/0.05) To estimate the tension per cord in the splice, a hypothetical tire built entirely with splice regions is assumed. For this case the cord count parameter is defined as [N.sub.s]. 3. To translate cord elongation ([Epsilon]) to sidewall indentation, it is assumed that the cords from a semi-circle as viewed from a meridional cross-section. Therefore, the less elongated splice region exhibits less of an "effective" circumference than the rest of the tire. Sidewall indentation is the difference in the radii (r) for the semi-circles representing the non-splice (ns) and splice (s) cords. Thus, the SWI depth (D) is given by:

(3) D = [r.sub.ns] - [r.sub.s] = [r.sup.[Omicron].sub.ns] (1 + [[Epsilon].sub.ns])-[r.sup.[Omicron].sub.s] (1 + [[Epsilon].sub.s]) where [r.sup.[Omicron].sub.i] is the radius of the uninflated tire in the splice or non-splice regions. When the splice and non-splice radii in the uninflated tire are equal, then

(4) D = [r.sup.[Omicron]] ([[Epsilon].sub.ns] - [[Epsilon].sub.s]) Alternatively, the "effective" modulus approach leads to: (5) D = [r.sup.[Omicron]] [Pi]P ([r.sup.2.sub.c] - [r.sup.2.sub.w]) ([N.sub.s] - N)/E [NN.sub.s]

References

[1]R.J. Samuels, "Structured polymer properties," Wiley-Interscience, New York, 1973, Chap. 4. [2]V.B. Gupta and S. Kumar, J. Appl. Polym. Sci., 26, 1877 (1981). [3]C.P. Buckley and D.R. Salem, Polymer, 28, 69 (1987). [4]T. Kunugi, A. Suzuki and M. Hashimoto, J. Appl. Polym. Sci., 26, 213 (1981). [5]P.N. Peszkin and J.M. Schultz, J. Polym. Sci., Polym. Phys. Ed., 24, 2591 (1986). [6]D.C. Prevorsek, G.A. Tirpak, P.J. Harget and A.C. Reimschuessel, J. Macromol. Sci., Physics, B9 (4), 733 (1974). [7]M.G. Northolt, A. Roos and J.H. Kampschreur, J. Polym. Sci., Polym. Phys. Ed., 27, 1107 (1989). [8]T. Thistlethwaite, R. Jakeways and I.M. Ward, Polymer, 29, 61 (1980). [9]J.O. Warwicker and S.G. Graham, J. Appl. Polym. Sci., 26, 3045 (1981). [10]R.J. Young, "Introduction to polymers," Chapman and Hall, New York, 1983, Chap. 4. [11]R. de P. Daubeny, C.W. Bunn and C.J. Brown, Proc. R. Soc. London Ser., A226, 531 (1954). [12]H.G. Kilian, H. Halboth and E. Jenckel, Kolloid Z., 172, 166 (1960). [13]B. Wunderlich, Macromolecular Physics, Volume 3, Academic Press, 1980, Chap. 8.4.5. [14]R.J. Samuels, J. Appl. Polym. Sci., Part A-2, 10, 781 (1972). [15]Allied Signal Internal Data. [16]Encyclopedia of Polymer Science and Engineering, vol. 5, pg. 80. [17]J.P.M. Dortmans, Textile Month, Dec., 43 (1988). [18]D.C. Prevorsek, C.W. Beringer and Y.D. Kwon, Macromolecular Secretariat, Organic Coatings and Plastics Division of ACS, (1983). [19]C.B. Freeman, Rubber and Plastics News, April (1990). [20]J.D. Walter, Journal of Engineering for Industry, Transactions of ASME, Paper #71-RP-A (1970-71).