Aggregates for concrete - Concrete Mix Design, Quality Control and Specification - page 58

Rock from an untried source must be tested for susceptibility to alkali-
aggregate reaction. Although comparatively rare, this reaction produces
such catastrophic results that its occurrence should not be risked without
at least a petrographic report. There is a rapid chemical test for reactivity
but it is not very reliable. The accelerated mortar test (ASTM C1260 or
C1567 for blended cement) is a better screening test. However, it will give
a large percentage of false negatives. Another approach is to add sufficient
pozzolans to the mix to control any possible ASR as advocated by Carse.
Another important feature of a coarse aggregate is its bond characteristics
(especially in high strength concrete and where flexural or tensile strength
is of special importance). This is a composite effect of its chemical nature,
its surface roughness, its particle shape, its absorption, and its cleanliness.
As an example of the importance of this feature is Day's experience with
two different basalts in Melbourne. One of these is superior to the other on
every tested feature, it is stronger, has a higher elastic modulus, is denser,
has less moisture movement, and a higher abrasion resistance. However,
the other aggregate was better able to produce concrete of average strength
over 60 MPa. We assume that this was due to the first aggregate being
so dense and impermeable that cement paste had difficulty in bonding to
it. It is interesting to note that the subsequent introduction of silica fume
reversed this situation, confirming the beneficial effect of silica fume on
bond and the interfacial transition zone.
The particle shape of the aggregate is influenced by the crushing process.
The stone type does have a distinct influence, some stones being more liable
to splinter into sharp fragments or to produce a larger amount of dust
than others. However, the crushing process also has a large influence. Cone
crushers are perhaps the most efficient and economical type of crusher, but
they do not produce as good a particle shape as a hammer mill. Other influ-
encing factors are the reduction ratio (a large reduction in a single stage
tending to produce a worse shape) and the continuity of feeding (choke
feeding giving a better shape).
The effect of a poor particle shape (flaky and elongated) is to require a
higher fine aggregate and water content (and therefore a higher cementi-
tious or admixture content) for a given workability and strength. The best
measure of this is the angularity number, being the percentage voids minus
33. Oddly enough Kaplan's work (1958) on the subject suggests that the
sharpness of the edges and corners tends to make more difference to this
parameter than flakiness and elongation.
The question of particle shape must include considering the relative mer-
its of crushed rock and rounded river gravel. Gravels are often reputed
to give inferior results, particularly for high-strength concrete. There is
no denying that this is true for a given water/cement ratio and that it is
true generally where tensile or flexural strength is concerned. However in
terms of compressive strength, with equal cement content and equal ease of