The degree of carbon black dispersion is a very important
consideration in the manufacture of rubber both in
quality control and basic research. A study has been
made of various aspects of assessing dispersion and
the dispersion process of carbon black in rubber.
The main objectives of this work were:
(1) To investigate the relationship between the light
scattered at a fixed angle from the rubber surface
and carbon black dispersion and hence develop and
evaluate a new dispersion assessment system. (2) To
determine the normal variations of industrial rubber
mixing installations, and (3) To study the effect
of internal mixing variables on black dispersion and
other properties of rubber.
To achieve these goals a carbon black dispersion assessment
system based on an inverted Dark Field Reflected
Light (D.F.R.L.) microscope was developed in three
versions and successfully tested. The three versions
differed in their degree of sophistication and automation
and would be expected to find application ranging from
routine quality control to research. The basic principle
was that a rubber sample (cured or uncured) was cut
with a new razor blade and the surface observed in
a D.F.R.L. microscope. The light beam from the sample
surface is sensed by a photometer and its intensity
was shown to be related to black dispersion. In version
III an automatic stage driven by two stepper motors
was designed and fitted to the microscope to perform
object plane scanning. The photometer and the stage
were interfaced with an Apple II microcomputer providing
the following functions; stage control, photometer
data acquisition, statistical analysis, data storage and results output.
The system was evaluated
by taking measurements on several identically formulated
compounds differing only in black dispersion.
A general rubber goods and a tyre manufacturing installation
were studied. Several production batches were
sampled at various mixing stages and subjected to black
dispersion assessment, cure and vulcanisate properties
measurement. Analysis of variance of the results was
accomplished with a statistical computer package designated
GENSTAT Version 4.03.
Factorial experimental designs and multivariate regression
analysis techniques were used in studying the effect
of mixing variables on black dispersion and other properties.
The results are presented in the form of
response equations and contour graphs are used to enable
second order interactions to be readily identified.
A Doctoral Thesis. Submitted in partial fulfilment of the requirements for the award of Doctor of Philosophy of Loughborough University.