An elasticity/viscosity tester for raw rubbers and compounded stock.

Melvin Mooney in his Charles Goodyear Medal address in 1962 castigated the industry for still using his disc rheometer to assess processing (ref. 1). Little has changed. We still have a plethora of vibratory instruments being used against his "viscometer" of 1934.

This is despite the general knowledge that the Mooney of a rubber has little correlation with quality or processing (ref. 2).

Present need

The biggest tire companies in the late 1980s bemoaned the lack of technology to meet the needs for consistent products for just-in-time supply of rubber products (ref. 3).

Accordingly, the International Rubber Study Group approached me to propose a solution. This tester is the outcome.

Theoretical analysis

Since Mooney's day, we know that rubber is a linear polymer of a wide molecular distribution. It is convoluted approximately in accord with the mathematical problem of random flight (ref. 4). On mastication the chains are extended, some to rupture point in their middle monomer units (ref. 5).

The extended chains have a tendency to return to their most probable end-to-end distance (ref. 6). This results in a force normal to the direction of extension if the elastic recovery is constrained, the so-called normal elastic force. This elasticity (E) can be defined as the force exerted normal to the direction of extension by originally 1 sq. cm cube of rubber on extension by 1 cm and not permitting the rubber to retract.

Viscosity (V) is caused by the attractive secondary forces between units of neighboring rubber chains resisting shearing of the rubber. The viscosity can be defined as the force per sq. cm of surface to move it 1 cm in relation to an opposite similar surface 1 cm apart.

The above defining of units is not being pedantic. The purpose is to provide E and V values which would give the same values, whatever the type of tester is used for the measurement, providing of course that the tester is capable of such a measurement. In the present case, any other suitably-designed tester than the one to be described will give the same results. (This is more than can be said for any current instruments.)

Now, processing operations essentially cause flow of the rubber. This flow depends on the properties E and V of the rubber. In a mixer, the flow pattern is too complicated to be modeled, and this is by far the most important operation for which one wants to describe the "processability" of the rubber. So, one has to choose a model system of greatest mathematical simplicity so that E and V can be measured, but with the stipulations that the typical rates of shear, strains of the rubber and cycle times of a mixer are employed.

We now take into account that a main purpose of the mixer is to soften the rubber, i.e., to reduce E and V. Hence, any processability numbers must be at least three to describe the breakdown curve of E vs. time t and V vs. t, as in figure 1.

[ILLUSTRATION OMITTED]

We can now measure processability quantitatively. It is by the change of E and V during mixing. It can apply to the gum raw rubber or to the compounded stock. It strictly applies to the particular mixer in which these changes are effected. One could laboriously extract samples during mixing and measure their E and V on a tester, but this is obviously impracticable for routine use.

A practical alternative is to measure the E vs. t and V vs. t curves of figure 1 of the rubber or lightly compounded stock with the above tester. Then, from experience with the particular mixer in the factory, to draw the dotted limit curves of figure 2, which give rubbers promising satisfactory operating conditions and consistency of product when reproducibly vulcanized. This is a suitable processability.

[ILLUSTRATION OMITTED]

The curves of figure 1, if reproducible for a bale, give an unambiguous measure of its quality. The initial values [E.sub.0] and [V.sub.0] give the initial elasticity and viscosity of the bale. The higher these numbers, the higher is a product gum-vulcanized with a reproducible curing recipe going to have properties depending on the smaller number of chain-ends in the network, such as tensile strength and elongation at break.

However, it is not simply a matter of obtaining bales with the highest [E.sub.0] and [V.sub.0] values. This is also reflected in the stiffness of the bale. The factory wants, from experience, bales of stiffness suitable for its processing equipment and operations. So again, the factory chooses the quality in terms of [E.sub.0] and [V.sub.0] values which it finds suits its operations and product properties.

What is more important is the choice of bales within a limit of [E.sub.0] and [V.sub.0], or better, within limits as in figure 1, so that processing and products are consistent.

In summary, processability and quality cannot be measured by single numerical values. Bearing in mind that it must be related to the individual mixer and its product line by limit curves, initial and at least two further values of E and V of figure 1 give quantitative values to these two elusive properties of raw or compounded rubbers.

Purchasing raw rubbers

Apart from dirt content, there is no quantitive quality standard for raw rubber. This is technically not good for purchasers. It is a lot worse for suppliers; they have little information on the quality of their raw materials and its variability when being fabricated and used. Seasonal changes and blending for natural rubber and batch syntheses for synthetic rubbers may give variations which are not found to be unacceptably variable until the product is made in large numbers. New suppliers, even new supplying countries, have difficulty breaking in to markets for quality products.

The Tester will allow fabricators/purchasers of rubber to specify quality by limit curves as in figure 2. Suppliers will be able to blend and coagulate latices rationally to achieve desired values of E and V. Bales will be able to be sold in lots within specified limits. The issues of quality and consistency will be dealt with objectively, irrespective of subjective factors such as particular supplier or country of origin. The supplier will know from the outset what is the quality and consistency of the goods he supplies, not having to rely on the opinion expressed to him by the purchaser.

Processability

From the above theoretical considerations, the factory management will be able to organize the primary operation such that it obtains compounded stock within its selected limits of values of E and V. It can then be sure that the stock at this stage is sufficiently consistent. The subsequent flow processes of extrusion, calendering, molding, etc., may well be reproducible enough in how they further change E and V that the material arrives at the vulcanization stage consistent in its polymer structure and filler dispersion. Then, the curing adding reproducible numbers of crosslinks, the product should be consistent unit after unit.

The tester designs itself

Polymer science leads to the above practical solutions without an iota of experimental effort. It also leads to the optimum choice of design of the tester from the specifications:

* E and V need to be measured at temperatures from 100-180 [degrees] C for a standard instrument.

* Rates of shear should be able to be set at constant values from 50 to 100 [sec.sup.-1].

* The strains imposed on the rubber by the shear should be up to 500%.

To these specifications are added the practical ones of:

* The tester must be robust and operable in all climates.

* The test procedure should be largely automatic, including the print-out of the curves as in figure 2.

* The sample of rubber must be easy to obtain, insert and remove.

* Needless to state, the tester should justify its cost to the purchaser.

Three possible configurations of test-head occur to us:

* A tube for extruding the rubber, with measure of pressure required, dye swell and stress-relaxation. This is rejected as requiring a continuous supply of the rubber or lightly-milled stock.

* A biconical rotor-in-chamber a la the original theory of Piper and Scott (ref. 6). This is rejected mainly because of the incorrect assumption that has to be made that only laminar shear takes place in this design and that shear rate changes linearly throughout the thickness of the sample (implicitly assumed by Piper and Scott).

* A Couette viscometer with means of measuring normal force as in figure 2. This satisfies the condition that it is the simplest model for a mixer and it measures E and V with no assumptions except that the shear is laminar, as is expected from its configuration.

The design chosen is shown in figure 2. Chips of the rubber or stock are placed in the tray (a). The upper platen (b) forces these to be transfer-molded down the tubes (c). These then fill the measuring chamber (d), which is cut off from the tubes by rotating its upper plate (e). The central rotor (f) provides the shear on the annulus of sample (g). The torque applied to the rotor and the normal force on the bottom plate (h) of the chamber are recorded by electrical sensors (i) and (j) and the signals converted by the software to give the print-outs of E and V with time.

Results and discussion

The first thing to note from figure 3 is that the elasticity component of the theological properties is as significant as the viscosity component. With the universal practice of ignoring it, the modern rubber technologist is as disabled as the one-eyed man. He cannot appreciate the theological properties in the round and make full analysis and then use of them.

[GRAPH OMITTED]

The type of information provided by this tester will provide the technologist with 20:20 vision with regard to the meaning of the full data. For example, he will be able to plot the thixotropic properties of raw rubbers and stocks during mastication and, mixing against temperature and shear rate to optimize energy consumption. So will he not be restrained by the conditions set by Mr. Banbury in the 1920s on how he should carry out his primary operations, except by having a mixer which still limits him to these conditions.

Coming to the immediate purpose of selecting bales to form batches of consistent processing properties and giving reproducible products, the critical data of figure 3 and the complementary viscosity curves will enable him to purchase against the limiting curves for his mixers of figure 1. So will this in turn benefit the supplier.

References

(1.) Mooney, M., Rubber Chem and Tech 1962, 15, xxvii.

(2.) Izod, D.A.W. and Skam, G.D., Rubber World, 1968, 158, 48.

(3.) Int. Rubber Study Group Proceedings, 1988 and 1999.

(4.) Treloar, L.R.G., Physics of Rubber Elasticity, Oxford Univ., 1958.

(5.) Pike, M. and Watson, W.F., J. Polymer Sci., 1952, 9, 229.

(6.) Piper, G.H. and Scott, J.R., J. Sci. Instru., 1945, 22, 206.

This article by Dr. Bill Watson is his most recent in a continuous list of research papers from 1947 when he was with the British Rubber Producers Research Association. He founded the first rubber/plastics organization in 1962, now Rapra Technology Ltd. Watson-Brown Ltd. is currently being acclaimed for devulcanizing tire rubbers.