426 



THE INDIA RUBBER WORLD 



March 1, 1921 



Fig. 1. 



MECHANICAL PICTURE OF RUBBER 



The diayram in I-"ig. 1, which was first suggested by a former 

 colleague, Doctor F. M. G. Johnson, of McGill University, helps 

 clarify one's mental picture of the thcrmodynamical phenomena 

 associated with rubber strains. Rubber may be viewed as a 



combination of a cylinder of gas, 

 a steel spring, and a friction mem- 

 ber. Following this picture, ex- 

 tension of the rubber is accom- 

 panied in the first instance by com- 

 pression of the gas, thus generat- 

 ing the reversible heat, Qr. In 

 the second place, the steel spring 

 is compressed, thus generating the 

 increase in potential energj' of 

 strain, E. Lastly, the friction ele- 

 ment operates through the exten- 

 sion, generating nonreversible heat, 

 Qf. When the rubber retracts, 

 the gas expands, the spring re- 

 tracts, and the friction element 

 contributes another increment to 

 the nonreversible heat. 

 Mechanical Picture Suppose the sample is extended 

 OF Rubber and we apply heat to the system. 



The gas in the chamber will ex- 

 pand so as to use up heat, raising the weight W, thus shorten- 

 ing the rubber and so constituting the Joule effect. 



FRICTIONAL HEAT OR HYSTERESIS 



Although the reversible heat has doubtless a decided technical 

 significance, by far the most important energy transformation is 

 that of useful work into heat through hysteresis, and a short 

 account will here be given of some experiments carried out under 

 the author's direction by H. F. Schippel. 



Briefly, the method consisted in generating hysteresis loops 

 by graphically recording stress-strain curves of extension and 

 retraction up to varying elongations. By means of the planim- 

 eter the area of the hysteresis loop was determined and the 

 readings calculated to foot-pounds of energy referred to one 

 cubic inch of rubber. In order to obviate the inertia of dead 

 weight tensile machines, and for other reasons of convenience, 

 a special machine was devised, the principal features of which 

 were the alinement of a helical steel spring with the sample and 

 the use of extremely light and nicely fitting parts. The rubber 

 sample was merely a standard te.st piece about 0.1-inch in thick- 

 ness, 0.25-inch w-ide, and 2 inches between shoulders. The ends 

 of the test piece were secured in special light weight clamps de- 

 signed practically entirely to obviate creeping. The spring ex- 

 tension measured the stress, and the separation of the clamps, 

 the strains. 



Through the use of this special machine it was possible to 

 generate stress-strain cycles under both rapid, or adiabatic, and 

 slow, or isothermal, conditions. 



IsoTHERMAi. Cycles Adopted. It is of course obvious that the 

 size and character of the hysteresis cycles will depend on whether 

 they are generated adiabatically or isothermally. Under the 

 former conditions, the reversible and frictional heat developed 

 on extension are only slightly dissipated, and so act to increase 

 the stiffness of the sample and alter the trend of the curves. 

 Owing to the difficulties of inertia, it, was not found possible to 

 generate adiabatic loops at speeds sufficient to allow of con- 

 cordant results. The method finally adopted was to generate the 

 cycles at low speeds, for example, 20 inches per minute, or under 

 practically isothermal conditions. 



Preliminary Extensions. It is well known that the area of 

 the first hysteresis loop is greater than that of the second, and 

 ?o on. in most cases, however, the third loop differs only very 

 slightly from the succeeding loops, and so in our work when it 



was the intention to generate the hysteresis loop up to an elonga- 

 tion of 300 per cent, the test jiiece which had not been otherwise 

 handled after cutting from the molded slab was put through two 

 preliminary cycles up to 300 per cent, and then clamped into the 

 machine, and its hysteresis loop graphically recorded. In tak- 

 ing a succession of loops at increasing elongations the same test 

 piece was used and two preliminary loops made at each elonga- 

 tion. The initial length upon which the cycles were based was 

 the length measured after the two preliminary extensions had 

 been made. 



Range of Compounds Used. The experimental results in- 

 cluded tests on a standard series of factory compounds used in 

 tire construction. They thus included practically pure gum fric- 

 tion compounds, lightly loaded breaker compounds, and more 

 heavily loaded tread stock. These various stocks were mixed in 

 the factory under standard conditions, and given laboratory cures 

 ranging from SO per cent of the optimum cure in each case up 

 to cures 275 per cent over the optimum in some cases. 



Hysteresis loops were generated at elongations ranging from 

 100 to 500 per cent. There is considerable difference in opinion 

 as to whether in measuring hysteresis one should work to a 

 fixed percentage of the breaking load, irrespective of the elonga- 

 tion, or work to a definite elongation, irrespective of the load 

 required. The latter method seems to the writer the only correct 

 one from the technical standpoint, in view of the fact that the 

 strains incurred, for example, by the skim coat, breaker, and 

 tread of a pneumatic tire are arbitrarily fixed by the inflation 

 pressure and the load. 



Relation Between Hysteresis Loss and Cyclic Elong.vtion. 

 Fig. 2 illustrates the results obtained with a typical pure gum, 

 high-grade tire friction with a breaking elongation of upwards 



of 900 per cent. This par- 

 ticular compound con- 

 tained 5 pounds of sulphur 

 to 100 pounds of rubber, 

 of which 60 were first 

 latex rubber and the 

 other 40 a soft-cured wild 

 rubber. The only other 

 ingredients were a small 

 percentage of thiocarban- 

 ilide and 5 pounds of zinc 

 as activator. The energy 

 units are expressed as 

 one-hundredths of a foot- 

 pound calculated to a 

 cubic inch of rubber. The 

 relationship is of the char- 

 acter of a rectangular 

 hyperbola, and the hysteresis increases very sharply for elonga- 

 tions exceeding 300 per cent. Viewing hysteresis as frictional 

 loss, it is natural to expect sharply increased friction to accom- 

 pany the rapidly increasing lateral compressions in the test piece. 

 Following our mechanical picture, it is analogous to contraction 

 of the friction element upon the moving arm. 



Adoption or Standard Loop. For comparison of different 

 compounds and for different cures it was decided to adopt a 

 standard cyclic elongation, and in order to reduce experimental 

 error it was of course desirable to select an elongation lower than 

 300 per cent, or lying on the flat portion of the curve. For higher 

 elongations the energy loss changes so rapidly with slight changes 

 in the elongation as to make concordant results difficult. More- 

 over, a brief calculation of the strains set up, for example in the 

 skim coat of a pneumatic casing run under service conditions 

 shows that under conditions of standard factory practice the rubber 

 is strained to an elongation of not much more than 200 per cent 

 each time the tire flattens against the road. For comparative 

 purposes we therefore adopted a standard cycle of 200 per cent 

 elongation. 



CYCLIC ELONGATION ■. 



Fig. 2 



