98 



THE INDIA RUBBER WORLD 



XOVEMBER 1, 1920 



The Effect of Compounding Ingredients on the Physical Properties 



of Rubber 



By C. Olin North 



RUBBER COMPOUNDING 



IT IS GENERALLY REALIZED that thc Compounding of rubber is more 

 or less of an art. It depends solely on a large number of un- 

 correlated and apparently unrelated facts. It is hoped that in 

 time this art will become a science with facts, theories and laws 

 so well substantiated that guess work and e.xperiments will be 

 reduced to a minimum. 



Before this ideal condition can be attained both rubber testing 

 and compounding must make considerable progress. Tensile 

 strength and ultimate elongation are important but tell only a 

 very small part of the whole story. Tests are needed which will 

 give us true measures of hardness, toughness, plasticity, resiliency, 

 internal friction, hysteresis, and many other properties. 



The purpose of this paper is to present some data as to the ef- 

 fect of certain common compounding ingredients on rubber and 

 to propose a method of visualizing the peculiar behavior of these 

 substances. 



It should be mentioned in the beginning that the tests on which 

 this work is based are very crude from the standpoint of scientific 

 accuracy but it is believed that the values obtained, the curves, 

 etc., are relative and as such will be more or less of interest to 

 other rubber technologists. 



COMPOUNDING EXPERIMENTS 



It was realized in the beginning that the usual weight method 

 of compounding was not only valueless but misleading. Conse- 

 quently a basis of 100 volumes of rubber was chosen to which 

 were added volumes of the different fillers varying from zero to 

 fifty. 



In the first experiments it was thought desirable to use a small 

 quantity of accelerator, for which purpose thiocarbanilide was 

 selected. Later this practice was discontinued and in all but two 

 of the experiments (barytes and zinc oxide) described below, no 

 curing agent other than sulphur was employed. A selected grade 

 of pale crepe was used with all fillers except barytes. Stocks 

 were prepared on small experimental mixing rolls and sheets were 

 vulcanized in the usual manner, in molds maintained at 140 de- 

 grees C. in a hydraulic press. Physical tests were performed on 

 a Cooey testing machine. 



Some years ago Dr. Warren K. Lewis called our attention to 

 thc fact that we were measuring tensile strength at the expense 

 of ultimate elongation. In the present methods of testing, tensile 

 strength is figured on the area of the test piece under no load. 

 This is very unfair to a stock high in rubber since the actual area 

 at break is considerably smaller than the original area and the 



r X d' 



relative decrease in cross-section of a soft stock is much greater 

 than for one heavily loaded. 



TENSILE STRENGTH AT BREAK 



Assuming that the volume cf a stuck remains constant through- 

 out elongation, it can be readily shown that thc tensile at break 

 can be arrived at by multiplying the tensile strength figured on 

 the area at rest, by the final length and dividing by the original 

 length. Thus, — 



Let L ^ the load necessary for rupture 



IV :^ width of a test piece before stretching 

 d = distance between the marks 

 t = thickness tf the test piece 

 I ' rn the volume =: wdt 



T = tensile figured on cross-ssction at rest 

 T' ^ tensile at break _ . 



Let u'd't and V represent the respective dimensions at break 

 Since V ■= V by assumption, then tvdt — w'd't' 



T = — T' — 



■wt ii't' 



.". Twt = T'w't' 

 Tut 



and T' = ■ But wdt = Wd't' 



w'f 

 wt d' 



w'f d 



Substituting, T' 



d 



Unfortunately the volume of a test piece does not remain con- 

 stant during elongation as has been shown by Schippel.^ Con- 

 sequently a correction factor should be used if absolute accuracy 

 is desired. 



.•\nother method of taking into account the decrease in area 

 and the corresponding increase in length is by Stevens' "ten- 

 sile product," which is obtained by multiplying the ultimate 

 elongation by the tensile strength as usually calculated. When 

 dealing with hard rubbers where the elongation is practically 

 zero, this method is absurd since the tensile product becomes 

 zero. However, on soft stocks it is a satisfactory unit of com- 

 parison. It is convenient when working in English units to 

 divide by 10,000. Tensile product is a less logical method of 

 attack than tensile at break but since the curves, when plotted 

 against volumes of filler, are parallel and the former is rather 

 generally used by rubber technologists this unit was chosen for 

 the comparisons given below. Correction factors are necessary 

 for absolute accuracy as in the case of tensile at break, but 

 in our experiments we did not use them, partly because of the 

 small error introduced by volume change of the test piece and 

 partly because of the lack of information about this phenomenon. 



1 Read before the Rubber Division of the American Chemical Society, at 

 the St. Louis meet ng, April 12-16, 1920. 

 " Industrial and Engineering Chemistry, Vol. 12 — 1, page 33. 



