390
Proceedings of the 18
th
International Conference on Soil Mechanics and Geotechnical Engineering, Paris 2013
The pipette method itself was introduced by Middleton in
1930 as a means of determining the erosion potential of a soil.
The difference in the methods was that the samples collected in
the pipette consisted of particles of a maximum diameter of
0.05mm (Middleton, 1930). Volk’s method was, however,
preferred because studies in the southwest of the United States
indicated that the dispersion of the clay fraction (< 0.005 mm
diameter), was more significant in assessing the piping potential
of soils (Decker and Dunnigan, 1977).
The SCS involvement in earth dam construction in the
United States increased in the 1940’s and early 1950’s. The
dispersion test was employed as a routine procedure for all
samples submitted to SCS Laboratories. It was during this
period that the test procedure was adapted to use a hydrometer
instead of a pipette (Decker and Dunnigan, 1977). The
procedure is specified in ASTM standard D422-63 for Particle
Size Analysis of soils (ASTM International, 2007a & b).
With the increased use of the dispersion test and the growth
of the SCS testing facilities in the 1950’s and 1960’s, a few
inconsistencies were detected, such as, the test was not always
reproducible in the same laboratory, and there appeared to be
differences in the results between laboratories. Studies revealed
that water quality and moisture contents of the samples were
responsible for these errors and as such distilled water and in
situ moisture contents were specified (Decker and Dunnigan,
1977). Since that time, there appear to have been no discussions
in the literature regarding any anomalies encountered during
testing.
3 EVALUATION OF CURRENT TESTING METHODS
The test methods for hydrometer analysis currently in use are
the American standards (ASTM International, 2007), the British
standards (BSi, 1990) and the South African Technical Methods
for Highways - TMH1 (NITRR, 1986). Table 1 summarizes
procedures carried out for each test standard, focusing on the
main aspects of the testing procedure.
Table 1: Difference in testing procedures used for the determination of the clay fraction of a soil.
Property
ASTM- D422-63
BSi- BS 1377-2: 1990
TMH1- 1986 (A6)
Amount of dry
soil required for
the test:
100g of sand sized particles (i.e.:
particles less than 2mm in size) or
50g of soil fines if material has a
high percentage of silt/ clay.
Depends on type of soil, i.e.: 100g for sandy
soil, 50g for silty soil and 30g for clayey soil.
100g of soil fines (i.e.: particles
less than 0.425mm in size) or 50g
of soil fines if material has a high
percentage of silt/ clay.
Pre-treatment
None
With hydrogen peroxide if organic matter
present.
None
Dispersing agent
125mℓ sodium hexametaphosphate
solution at 40g/ℓ.
100ml sodium hexametaphosphate solution
comprising 33g Na-hexa. + 7g Na-carbonate in
distilled water to make 1 ℓ of solution.
NB: If soil does not disperse completely, allow
soil to settle, decant as much of the water as
possible and then add a further 100mℓ of
dispersant.
5 mℓ each of sodium silicate and
sodium oxalate solutions.
Soaking
Samples soaked for minimum of 16
hrs.
Soak in solution for min of 4hrs or overnight.
Minimum period of 2 hrs but
preferably overnight.
Hydrometer
readings
Taken at 2min; 5min; 15min; 30min;
1h; 4h and 24h.
Taken at 8min; 30min; 2h; 8h & 24h.
Taken at 18s, 40s & 1h.
Analysis
Equations used to calculate
maximum diameter of particles in
suspension
Equations used to calculate maximum diameter
of particles in suspension.
Maximum particle size calculated.
At 1h, max size is 0.005mm, at
40s max size is 0.05mm and at
18s, max size is 0.075mm.
Temperature
Constant temperature at or near 20°C
is required.
Constant temperature bath of 25°C, to an
accuracy of ± 0.5°C.
20°C when readings are taken or a
temperature correction has to be
applied.
Close observation of these test procedures illustrates little
variation in the method of determination of the dispersion ratio,
except with regard to the types of dispersing agents used. The
ASTM and BSi standards specify that sodium
hexametaphosphate be used as a dispersing agent. However, the
solution is prepared differently and at different proportions in
each standard. The volume required to disperse the sample in
the test is also very different.
TMH1 specifies that a combination of sodium silicate and
sodium oxalate be used as the dispersing agent. A study of the
past revisions of the South African standards shows that the
combination of sodium silicate and sodium oxalate was
specified in the 1958 and 1986 test methods, whereas, sodium
hexametaphosphate was specified as the standard dispersant in
the 1970 version. There is no reason found as to why the
combination of sodium silicate and oxalate is used instead of
sodium hexametaphosphate or why there was a change in the
standard in 1970.
It should also be noted that TMH1 states that the one hour
hydrometer reading indicates the percentage of the clay fraction
in the sample. Analyses using Stoke’s Law show that at one
hour, particles in suspension in all samples are in the range of
0.006 – 0.007 mm, which is silt-sized and not clay-sized.
4 TESTING EQUIPMENT AND METHODS USED
A study was conducted to evaluate the effects of the different
methods on the test results. For the study, testing procedures
that were carried out were the South African standard (TMH1)
and the American standard (ASTM), the primary difference in
the procedures, being the variation in dispersing agents. The
tests were carried out on three samples, one non-dispersive, one
highly dispersive and a third that was presumed to be moderate
to slightly dispersive (based on past experience and field
observations).
The dispersing agents used were combinations of sodium
hexametaphosphate (33g) plus sodium carbonate (7g) in a liter
of water and a sodium silicate (5mℓ) plus sodium oxalate (5mℓ)
solution. Samples were left to stand for approximately 16 hours
after dispersion before being mechanically agitated and then the
hydrometer readings were taken. Time intervals for the
hydrometer readings were at 1 hour, 40 seconds and 18 seconds
as specified in TMH1 Method A6. Duplicate samples were
