Effect analysis of mixing process variables on viscoelastic properties of NR tread compounds.

Statistical methods have considerable utility in rubber research and development. Such methods are a cost-effective way to help a user to formulate questions in a way as to make a research program more productive. The use of statistical methods is particularly fruitful in cases where simultaneous balancing of different parameters like cost, property level and processing behavior seems to be difficult without making undue compromises. Statistical methods are also useful in making sound decisions in light of extreme variability in the data with conflicting requirements. Research in rubber compounding for tire application is one such example.

The theory of viscoelasticity, at least in its linear regime, is well understood, modeled and described. For rubber compounds, it can be considered that the elastomers used in these compounds exhibit quasi-linear viscoelasticity behavior in the tire deformation range. For these elastomers, the mechanical response to strain energy input may be described by the well known complex modulus E *, expressed as a function of an elastic E' and loss modulus E".

The relationship of rubber compounding ingredients with viscoelastic behavior of the vulcanizates has been well explored over the last 50 years or so. But very little information has been published on the impact of compound processing variables with viscoelastic behavior. This is also the case with statistical analysis on rubber processing behavior.

In this article, an attempt is made to statistically evaluate the impact of changes in processing variables on the viscoelastic behavior of a bias truck tire tread compound.

Experimental

Materials

The formulation used for this study is shown in table 1. The compounding ingredients for the experiments included natural rubber (RSS-3), carbon black, aromatic oil, zinc oxide, stearic acid, n-phenylene (1,3 dimethyl,butyl) p-phenylene diamine (6PPD), suffur and N-oxy-bis benzthiazole sulfenamide (NOBS).

Statistical design of experiment

A two level factorial design was used throughout the experiment. Factorial designs allow for the simultaneous study of the effects that several factors may have on a process. When performing an experiment, varying the levels of the factors simultaneously rather than one at a time is efficient in terms of time and cost, and it allows for the study of interactions between the factors. Interactions are the driving force in many processes, and without the use of factorial design, many important interactions may remain undetected. It is better to use a "full factorial design" to evaluate the effect of each of the variables and its interactions with other variables. However, while dealing with a large number of factors, it is often convenient to use a "fractional factorial design" to minimize experimental cost and time. A defined fraction of the full factorial has been selected to assess the influence of each of the factors and the important interacting factors on the overall average effect. The number of experiments is drastically reduced with a fractional factorial design. For example, with seven factors at two levels, full factorial design calls for [2.sup.7] or 128 experimental runs; whereas, a 1/4 factorial (fractional factorial) design calls for only 32 experimental runs. Although three factor and higher interactions are aliased for a 1/4 fraction and cannot be determined, all the two factor interactions and single factor effects can be safely assessed.

All the seven process variables with two level design were codified and put into Minitab Statistical Software (Release: 13) to get randomized statistical experimental runs (ref. 4). Accordingly, the test compounds were mixed.

Mixing procedure

The mixing has been divided into two groups according to mastication, i.e., pre-masticated rubber (PMR) and mast in situ (MIS).

In the preparation of PMR, mastication was done for about 2-1/2 minutes. The controlling factor was 0.15 kwh power input (PI) maximum and the dump temperature was 120 [+ or -] 20[degrees]C. After eight hours, the PMR was used for masterbatch mixing. The batch weight of 1,300 gms contains 100 phr RSS-3 and 0.2 phr PCTP peptizer.

In the preparation of MIS, mastication was done up to 0.04 kwh (PI) with rubber and peptizer, and this PI and corresponding time were excluded from total masterbatch kwh (PI) and time.

Process variables

Seven process variables have been selected for experimentation with two levels of variation. The selection of variables is arbitrary in nature based on past experiences on processing. This is shown in table 2.

Mixing

The compounds were mixed in the lab size internal mixer (model: 00) with a mixing chamber volume of 1.5 liters. The mixing was done in three stages, i.e., masterbatch 1, masterbatch 2/remill and final batch. The same materials were used for all 32 batches, and the time interval between the mixing of different batches was kept constant. In doing so, background noise factor is reduced. Mixing parameters were varied for the first two stages only (master-1, master-2/remill).

Preparation of rubber vulcanizates

The cure properties of the compound were determined by an MDR 2000E from Alpha Technologies at a temperature of 141[degrees]C for 60 minutes. This was done to check the cure characteristics of the experimental compounds. The compounds were cured at 141[degrees]C for 60 min. using an electrically heated press of platen dimension 12" x 12" for dynamic mechanical analysis specimen preparation.

Dynamic mechanical characterization

Dynamic mechanical analysis (DMA) was done with a VA 4000 viscoanalyzer at 30[degrees]C with 5% strain and at a frequency of 11 Hz in tension-compression mode.

Results and discussion

The test results were analyzed with ANOVA (Analysis of Variance) in the Minitab Statistical Software (ref. 4). The main effects and interaction effects of the factors on the responses were studied in the above mentioned software. The compound design, along with its properties, are shown in table 3.

The effect of each factor on DMA response can be clearly defined by the main effect plots. Thus, for example, the variable ram pressure gives higher E * (complex modulus) than the fixed ram pressure (figure 1).

[FIGURE 1 OMITTED]

On the other hand, an interaction plot shows the combined effect of each level of a factor, with the level of a second factor held constant. Interaction plots are useful for judging the presence of interaction effect of the factors on the response. For example, the interaction effect of 0.52 kwh (PI) and variable ram pressure gave higher E * than the 0.52 kwh (PI) and fixed ram pressure (figure 2). It was also observed that the interaction effect of 0.63 kwh (PI) and 18 passes through the dump mill reduces E *. Similarly, the interaction effect of dry mixing and remill produces higher E * (figure 2). This may be due to the fact that ram float reduces the shearing force; so the chain breakage is reduced leading to higher E *. At the same time, higher PI and increasing the number of passes through the mill causes extensive chain breakdown, leading to lower E *.

[FIGURE 2 OMITTED]

E' is affected in similar ways as the E * (figures 3 and 4).

[FIGURES 3 & 4 OMITTED]

The main effects plot clearly shows that the process variables PI, ram pressure and type of 2nd stage have a major effect on E" (viscous modulus) (figure 5). The interaction effect of variable ram pressure and 0.52 kwh (PI) gives higher E" than the fixed ram pressure and 0.63 kwh (PI) (figure 6). This is possibly due to relatively poorer dispersion of carbon black in the former case compared to the latter. A higher number of passes in the dump mill and ram pressure had pronounced effects in reducing E". The com

[FIGURES 5 & 6 OMITTED]

bination of 0.63 kwh (PI) and 18 mill passes produced the lowest E" (figure 6).

The process variables of PI, ram pressure and type of 2nd stage also had a major influence on the tan [sigma] value (figure 7). The interaction effect of 0.63 kwh (PI) and 18 passes reduced tan [sigma]. Similarly, the interaction effects of: nine passes and masterbatch-2, dry mixing and masterbatch-2, and variable rotor rpm and normal mixing reduced the tan [sigma] value (figure 8). This may be due to better filler dispersion achieved in the above combinations.

The interaction of MIS (mast in situ) and 0.63 kwh (PI) slightly reduces the D" (loss compliance) value. The other interactions do not have much effect on D" (figure 10). (Note: D" = E"/[(E *).sup.2]--applicable for constant stress environments).

A summary of the main effects of the process variables on viscoelastic behavior is shown in table 4. The segregation is based on the quantitative effects the variables can induce on parameters, within the range of experimentation. Thus, >4% effect is termed as "major," <1% effect is termed as "no effect" and "minor effect" is in between the two extremes.

Summary

A two-level fractional factorial design can be used as an effective tool in screening rubber compound mixing processing variables having different degrees of effects on viscoelastic behavior. This is helpful in redefining future studies in this line. However, the scenario may change, depending on the type of formulation, machinery and processing variables selected. Further studies in this line are in progress.

Table 1--formulation

Ingredients    Amount (phr)

RSS-3            100
Peptizer         0.2
N-330             50
Aromatic oil       7
Zinc oxide         5
Stearic acid       3
6PPD               2
Sulfur           2.5
NOBS             0.5

Table 2--process variables

Process variables              Levels

                        1                 2

Mastication type        Pre-masticated    Mast in situ
                        rubber
Mixing PI * (kwh)       Low (0.52)        High (0.63)
Ram pressure            Fixed at 70 psi   Variable
                                          (70 psi to float)
Rotor RPM               Fixed at 60       Variable
                                          (60 to 45 to 30)
Mixing type **          Normal sequence   Dry mixing
  for masterbatch 1                       sequence
Passes through          9                 18
  dump mill (total for
  3 mix stages)
Type of 2nd stage       Masterbatch 2     Mixer remill
                        (MB 1 + 10 CB
                        and 1 oil)

* PI is power integrator value associated with a mixing stage. Higher
value indicates extensive mixing. Note:
0.52 kwh (PI) = 0.30 + 0.12 + 0.10 for 1 st, 2nd and 3rd stage
respectively.
0.63 kwh (PI) = 0.35 + 0.18 + 0.10 for 1 st, 2nd and 3rd stage
respectively.

** Mixing type:

Normal sequence--polymer, then CB, oil, chemicals added together.
Dry mixing sequence--polymer + 30 phr CB; remaining CB, oil,
chemicals added at 0.18 kwh PI.

Number of passes through dump mill.

9 = 3 + 3 + 3 for 1st, 2nd and 3rd stage respectively.

18 = 8 + 7 + 3 for 1st, 2nd and 3rd stage respectively.

Table 3--experimental design and test data

Runs    A     B        C          D         E

1      PMR   0.63   Variable   Variable   Normal
2      PMR   0.63   Variable      Fixed      Dry
3      MIS   0.52      Fixed   Variable   Normal
4      MIS   0.63   Variable   Variable   Normal
5      PMR   0.52   Variable   Variable   Normal
6      PMR   0.52      Fixed      Fixed      Dry
7      PMR   0.63      Fixed   Variable      Dry
8      PMR   0.52      Fixed      Fixed   Normal
9      PMR   0.63   Variable      Fixed   Normal
10     MIS   0.52      Fixed      Fixed   Normal
11     PMR   0.52   Variable      Fixed   Normal
12     PMR   0.63   Variable   Variable      Dry
13     MIS   0.63      Fixed   Variable      Dry
14     MIS   0.52      Fixed   Variable      Dry
15     MIS   0.63   Variable      Fixed      Dry
16     PMR   0.52   Variable      Fixed      Dry
17     MIS   0.63      Fixed   Variable   Normal
18     MIS   0.63      Fixed      Fixed   Normal
19     MIS   0.63   Variable      Fixed   Normal
20     MIS   0.52   Variable      Fixed      Dry
21     MIS   0.63      Fixed      Fixed      Dry
22     PMR   0.52   Variable   Variable      Dry
23     MIS   0.52   Variable   Variable      Dry
24     MIS   0.52   Variable   Variable   Normal
25     PMR   0.63      Fixed      Fixed   Normal
26     MIS   0.63   Variable   Variable      Dry
27     MIS   0.52   Variable      Fixed   Normal
28     PMR   0.52      Fixed   Variable   Normal
29     PMR   0.63      Fixed      Fixed      Dry
30     PMR   0.63      Fixed   Variable   Normal
31     MIS   0.52      Fixed      Fixed      Dry
32     PMR   0.52      Fixed   Variable      Dry

Runs   F    G    E *     E'      E"      Tan       D"
                 (Mpa)   (Mpa)   (Mpa)   [delta]   (Mpa-1)

1       9   M2   5.40    5.33    0.826   0.155     0.0284
2      18   M2   5.40    5.33    0.851   0.159     0.0292
3      18   M2   5.82    5.73    1.010   0.177     0.0299
4      18   RM   5.48    5.41    0.865   0.160     0.0288
5      18   RM   5.86    5.77    1.020   0.177     0.0297
6      18   RM   5.82    5.73    1.010   0.176     0.0298
7      18   RM   5.66    5.58    0.925   0.166     0.0289
8      18   M2   5.78    5.69    1.020   0.179     0.0305
9      18   RM   5.39    5.33    0.836   0.157     0.0287
10      9   RM   5.25    5.19    0.823   0.159     0.0298
11      9   M2   5.86    5.78    1.010   0.175     0.0293
12      9   RM   6.24    6.12    1.220   0.199     0.0313
13      9   M2   5.72    5.65    0.907   0.161     0.0277
14     18   RM   6.00    5.90    1.090   0.185     0.0303
15      9   RM   6.10    6.00    1.090   0.182     0.0294
16      9   RM   6.00    5.90    1.080   0.183     0.0299
17      9   RM   5.93    5.85    0.959   0.164     0.0272
18     18   M2   5.93    5.84    1.010   0.173     0.0288
19      9   M2   6.21    6.11    1.120   0.183     0.0289
20     18   M2   5.85    5.76    1.000   0.174     0.0294
21     18   RM   5.60    5.53    0.912   0.165     0.0290
22     18   M2   6.27    6.16    1.160   0.188     0.0295
23      9   RM   6.22    6.11    1.160   0.189     0.0299
24      9   M2   6.11    6.01    1.090   0.182     0.0292
25      9   RM   6.06    5.95    1.140   0.191     0.0310
26     18   M2   5.68    5.61    0.938   0.167     0.0290
27     18   RM   6.10    6.00    1.100   0.183     0.0296
28      9   RM   5.62    5.54    0.936   0.169     0.0296
29      9   M2   5.51    5.44    0.889   0.163     0.0293
30     18   M2   5.29    5.22    0.834   0.160     0.0298
31      9   M2   5.53    5.46    0.854   0.156     0.0280
32      9   M2   5.56    5.48    0.952   0.174     0.0308

A = mastication type; B = power integrator kwh (PI);
C = ram pressure; D = screw RPM; E = mixing type;
F = number passes in dump mill; G = type of 2nd stage

[GRAPHIC OMITTED]

References

(1.) I. Amiri Amrace and A.R. Teimorian, Tire Technology International, Annual Review '97.

(2.) F.E. Welsh, B.R. Richmond and R.J. Emerson, "Mixing concepts for enhancing carbon black performance, "presented at ITEC '96, Akron, OH.

(3.) Miller and Freund, "Probability and statistics for engineers," 6th Edition.

(4.) Data analysis, "Minitab software manual," Minitab Inc.