for the other measured results (Gsb, Gssd, and absorption), the
differences were statistically significant between two groups of
blends—with and without adding the clay.
Figures 3-34 and 3-35 show the correlations between the
clay content and the Gsb and absorption results for the natural
sand and RC limestone, respectively. The correlations suggest
that the presence of a clay material (even in a small amount)
can significantly affect the measured results, especially Gsb,
Gssd, and absorption. It should be noted that the clay material
(sodium bentonite) used in this study can absorb a large
amount of water and swell 15 to 18 times its dry size. As shown
in Table 3-34, the Atterberg limits of the clay were very different
from those of the limestone P200.
The amount of the relatively clean P200 material also
affected the test results, especially Gsb, Gssd, and absorption.
Although this effect was less significant for the natural sand,
it was more profound for the RC limestone. As shown in Figures
3-32 and 3-33, when the amount of the relatively clean
P200 material was 20 percent or more, it affected the measured
Gsb, Gssd, and absorption values; however, the differences
were not statistically significant based on a 95-percent
confidence interval (significance level of 0.05).
Due to the significant effect of clays and P200 on AASHTO
T 84 test results, it is desirable to have a test method, such as
the sand equivalent test (AASHTO T 176), that can be used
to identify fine aggregate materials containing P200 that may
have adverse effects on AASHTO T 84 test results. An analysis
was conducted to assess the correlation between P200 contents
and sand equivalent test results of the fine aggregate
blends tested in this study. Figures 3-36 and 3-37 show the
correlations for the natural sand and RC limestone blends,
respectively. The effect of 1.25 percent of clays was similar to
that of 10 percent of relatively clean P200 on the sand equivalent
test results, and the effect of 2.5 percent of clays was similar
to that of 20 percent of relatively clean P200.
Due to the significant effect of clay on test results, a sand
equivalent threshold should be selected so that P200 materials
(containing clays) having adverse effects on test results should
be tested separately. Based on Figure 3-36, this cut-off sand
equivalent value is approximately 75 percent, and it is approximately
65 percent based on Figure 3-37. Thus, it is proposed
that a sand equivalent threshold of 75 percent is selected, and
it can be used as an option to determine if the “+200” and
P200 materials for a fine aggregate should be tested separately.