In obtaining meaningful sieve analysis data, six
major steps are recommended. 1) Obtain a
representative sample of the material to be evaluated.
2) Prepare the sample for evaluation; this may involve
washing and/or drying the sample. 3) Reduce the
sample to a size suitable for the sieve analysis
procedure. 4) Perform the actual sieve analysis
procedure. 5) Compute the data and convert the data
into a usable format. 6) Organize the data and
assemble the information for presentation.
Granular and powder materials are prone to
segregation during movement and storage of the
products. This segregation can be due to the disparity
of the particle sizes and the varied densities for
blended products. When forming a stockpile of
material, the larger, coarser particles are heavier and
tend to roll to the lowest portion and outer perimeter of
the cone. The finer particles are lighter and more
angular and remain concentrated at the top and
through the vertical center of the cone. Obtaining
samples from only the outer perimeter or from the top
of the cone would not provide a sample which would
be representative of the entire batch.
Sample extraction and preparation is the most
commonly overlooked variable in sieve
standardization programs. Testing bias can be added at
many places along the progression from the raw
materials received from a supplier, samples taken at
each stage of production, sample reduction procedures
and samples when the product is ready for shipment to
the customer. The way the samples are extracted from
the original bulk volume varies with the way the
materials are received, produced or stored. The ideal
sampling method is one, which provides the most
representative sample with the least amount of
material required.
The following paragraphs were first published in the
ASTM technical publication STP 447 A. The
collaborative efforts of the authors have produced a
section on sampling technique which will aid in
obtaining representative test samples from larger test
sources.
Sampling from a chute or belt
Accuracy in sampling is obtained where material is
flowing from a chute or belt conveyor. The ideal place
to collect the sample is where the material drops from
the chute or belt. If the material stream is small enough,
use a pail or other suitable receptacle which can be
swung completely across the flowing stream in a brief
interval of time and with uniform movement. The
sampling receptacle should not be allowed to
overflow, because the overflow would tend to reject a
higher proportion of the larger particles that exist in a
representative sample. Mechanical sampling devices
are available for selecting samples automatically from
a stream at uniform time intervals.
Sampling from carload shipments of coarse bulk
material
For coarse materials, such as crushed stone and
gravel, shipped in railroad cars, a recommended
method is to dig three or more trenches at least 30.48
cm (1 foot) deep and approximately 30.48 cm (1
foot) wide at the bottom. Equal portions are taken at
seven equally spaced points along the bottom of the
trench by pushing a shovel downward into the material
and not by scraping horizontally. Samples from trucks,
barges, or boats should be taken in the same manner as
from railroad cars, except that the number of trenches
should be adjusted to the size of the transportation unit
and tonnage involved.
Sampling from carload shipments of fine bulk
materials
One established method for sampling a carload of
bulk granular material is to take eight equal samples,
(approximately 700 to 1000 grams each) from the
bottom of a 30.48 cm (1 foot) conical excavation.
Samples should be suitably spaced to represent the
length and width of the car and then combined into a
single gross sample.
Sampling bulk shipments of fine material with a
sampling tube
An alternate and simpler method of sampling a
carload, or other bulk quantity of fine or granular
material is by use of a sampling tube which, for this
purpose, should be 38.1 mm (1 1/2 inches) by
approximately 1.829 m (6 feet). Five or six insertions
of the tube will produce approximately, a 2 pound
(907 g) sample.
Sampling from a carload of bagged material
One method of sampling a carload of material
shipped in bags is to select, at random, a number of
bags equal to the cube root of the total number of bags
in the car and to take suitable portions (800 to 1000
grams for minus 6 mm material) from each of the
selected bags for a combined gross sample.
Sampling from a pile
In sampling from a pile, particularly material like
crushed stone or coal containing large particles, it is
extremely difficult to secure samples that are truly
representative. At the apex of a conical pile, the
proportion of fines will be greater, while at the base;
the percentage of coarse particles will be greater.
Therefore, neither location will be representative of
the whole. In a shoveling process, every fifth or tenth
shovel, etc., should be taken depending on the amount
of the sample desired. The sample should consist of
small quantities taken at random from as many parts of
the pile as are accessible and taken in a manner that
the composite will have the same grading as the larger
amount.
Reduction of gross sample to test size for sieve
analysis
After the gross sample has been properly obtained,
the next step is to reduce it to a suitable size for sieve
analysis without impairing in any way the particle size
distribution characteristics of the original sample. This
phase of the operation should follow the applicable
procedures described in the succeeding sections and
should be performed with as much care as was used in
the collection of the gross sample and in performing
the sieve test.
Coning and quartering
Pile the gross sample in a cone, place each shovel
full at the apex of the cone, and allow it to run down
equally in all directions. This will mix the sample.
Then spread the sample in a circle and walk around the
pile, gradually widening the circle with a shovel until
the material is spread to a uniform thickness.
Mark the flat pile into quarters, and reject two
opposite quarters. Mix again into a conical pile, taking
alternate shovelfulls from the two quarters saved.
Continue the process of piling, flattening, and rejecting
two quarters until the sample is reduced to the required
size.
Sample splitters and reducers
Gross samples, if not too large, may be reduced to
test sample size by one or more passes through a
sample splitter or Jones type riffle, which will divide a
sample in half while maintaining the particle size
distribution of the original sample. By repeated passes,
the sample can be split into quarters, eighths, and soon
until the size of the sample desired is obtained. For
larger gross samples, sample reducers are available
which will select a representative 1/16 part with a
single pass. After just two passes through such a unit,
a representative one pound sample can be obtained
from an original 256 pounds. Three passes will give a
one pound sample from two tons of material. Always
make sure that the passages in the splitter or reducer
are at least three times the size of the largest particle in
the sample. Do not attempt to arrive at exactly the
amount of material specified for the test. If a 50 gram
sample is desired, arrive as near to this amount as
practicable, because it will make no difference in the
test percentage results whether the sample is slightly
larger or smaller. In attempting to arrive at an exact
weight, the tendency is to discriminate by the removal
of sizes that are not representative of the whole, thus
destroying the representative quality of the sample.
Size of Sample in the Test
There is a natural tendency, although incorrect, to
use an excessively large sample in the test. In most
cases, a smaller sample will provide a more accurate
analysis. Beware, however, that the more you split, the
greater the chance of error. Testing sieves are a go or
no go gauge; if the sample is too large it will not
permit each of the particles an opportunity to present
themselves to the screen surface. Often the limiting
factor for reducing the sample size is the accuracy of
the weighing device used to determine the amount of
material retained on the sieve.
Generally a 25 to 100 gram sample is recommended.
However, if it is necessary to establish the correct
sample size, utilize the following procedure: Using a
sample splitter, reduce samples to weights (i.e. 25, 50,
100, 200 grams). Analyze these various sample sizes
on a selected nest of sieves for a period of five minutes
preferably using a mechanical sieve shaker. If the test
with the 100 gram sample shows approximately the
same percentage passing the finest sieve as the 50
gram sample, whereas the 200 gram sample shows a
lower percentage, this would indicate that the 200
gram sample is too large and the 100 gram sample
would be satisfactory. Then run the 100 gram sample
on the same set of sieves for the same time period to
see if repetitive results are obtainable.
A useful table of recommended sample sizes for
tests with 200 mm or 8" diameter sieves is presented in
Table 4. Note that the table gives sample sizes listed by volume. Recommended sample weights in grams
can be determined by multiplying the values in
Column 3 and 4 by the bulk density (grams per cubic
centimeter) of the material to be tested rounded out
within a reasonable tolerance. If the actual bulk
density of a certain material is not known, the typical
density factor for the most nearly similar material
listed in Table 5 may be used.
To perform the actual sieve analysis, sieves should
be chosen in a sequence as described earlier. Use
every sieve, every other sieve, or every third sieve, etc.
between the desired size parameters. The use of sieves
in this sequential order will allow for better data
presentation and a more meaningful analysis of the test
results. Care should also be taken in selecting the
proper sieves to avoid overloading any sieve with an
especially large material peak. For example, a
specification may require 96% of the sample be
retained above a #50 mesh sieve. The proper way to
perform an analysis of this nature is to use 'relief
screens', that is, sieves in the 30, 35, 40 and 45 mesh
ranges to remove some of the burden from the critical
cut point of 50 mesh. If the relief sieves are not used,
the particles of exactly 50 mesh size or slightly larger
may become wedged in or forced through the sieve
openings by the mass of material resting above them.
Large concentrations of material on one sieve reduce
the opportunity for near size material to pass through
the sieve resulting in a larger portion of the material
retained on the test sieve. The sieve cut point would
be inaccurate and the sample would not meet the
specifications for the test.
The selected sieves should be assembled with the
coarsest sieve at the top of the stack, and the balance
of the stack in increasing magnitude of fineness
(increasing sieve numbers with smaller openings). The
stack should include a cover on the top sieve and a pan
below the finest sieve. The sieve stack can either be
shaken then wrapped by hand, or mounted in a sieve
shaker with a motorized or electrostatic drive
mechanism.
While many applications still use the handshaken
method for sieving, motor driven shakers have proven
to be much more consistent, minimizing variations
related to operator procedures. In powder analysis
below the 100 mesh range, the sieve shaker should be
equipped with a device to impart a shock wave to the
sieve stack at regular intervals. This hammer or
rapping device is necessary to reorient the particles on
the sieve and impart some shear forces to nearsize
particles blocking the sieve openings.
Recommended Time Intervals
The duration of the sieving interval is usually
regulated by industry standards, or by inhouse control
specifications. Commonly, 10, 15 or 20 minute tests
are used as arbitrary sieving intervals. To determine
the best interval for a new material, or to double check
the accuracy of existing specifications, the following
procedure can be used. Select the desired sieves for the
analysis. 1) Weigh up a sample of the material to be
tested and introduce it to the completed sieve stack. 2)
Shake the sieve stack for a period of 5 minutes. 3)
Weigh the residue in the pan and calculate the
percentage in relation to the starting weight. 4)
Reassemble the stack and shake for one additional
minute. 5) Repeat the weighup procedure and
calculate the percentage. If the percentage of fines
increased more than 1% between 5 minutes and 6
minutes, reassemble the stack and shake for an
additional minute. The data can be plotted as
percentage throughput vs. time for each data point you
calculate. When the change in the percentage of fines
passing in the 1 minute period drops below 1 %, the
test can be considered complete. Record the total
testing time for subsequent analyses.
Another type of sieve analysis is the wet sieve test.
In this method, the sample is weighed and then washed
through the finest sieve in the stack with water, a
wetting agent (water based), or some other compatible
solvent. After thoroughly washing the fines from the
raw sample, the residue is dried either over a hot plate
or in an oven. The temperature of the sieve should be
maintained below 149°C (300°F) to avoid loosening
of the sieve cloth or failure of the solder joint. After
drying, the residue is then sieved normally on the
balance of the sieve stack. The loss in weight not
accounted for on the coarse screens is assumed to be
fines or soluble material.
Wet sieve analysis is especially helpful when
working with naturally agglomerated materials,
Ultrafine powders with severe static changes, and in
samples where fine particles tend to cling to the coarse
fractions in the blend. The disadvantages associated
with wet sieving are primarily the time period required
to perform the analysis due to the additional washing
and drying time and the possible damage to the sieve
mesh by overloading. A common practice with wet
sieving operations is brushing or forcing the sample
through the mesh while the liquid medium is directed
on the sieve. This pressure can distort the sieve
openings or tear the mesh at the solder joint through
stress. Therefore, this procedure is not recommended.
Once the sieving interval is complete, whether dry or
wet sieving is used, the residue on each sieve is
removed by pouring the residue into a suitable
weighing vessel. To remove material wedged in the
sieve's openings, the sieve is inverted over a sheet of
paper or suitable collector and the underside of the
wire cloth brushed gently with a nylon paint brush
with bristles cut to a 25.4 mm (1") length. The side of
the sieve frame may be tapped gently with the handle
of the brush to dislodge the particles between brush
strokes. At no time should a needle or other sharp
object be used to remove the particles lodged in the
wire cloth. Special care should be taken when brushing
sieves finer than 80 mesh. Brushing can cause
distortions and irregularities in the sieve openings. The
procedure is repeated for each sieve in the stack and
contents of the pan.
The individual weights retained on the sieves should
be added and compared to the starting sample weight.
Wide variations or sample losses should be determined
immediately. If the finished sample weight varies
more than 2% from the initial weight, the analysis and
sample should be discarded and the test performed
another sample. If the sample weights are acceptable,
complete the calculations and report the individual
weights retained on each sieve.
Presentation and analysis of the resulting data is
frequently made easier by plotting on one of a number
of graph formats. The most common graphic
presentation is the plotting of the cumulative
percentage of material retained on a sieve (plotted on a
logarithmic scale) versus percentage (plotted on a
linear scale). The resulting curve allows a quick
approximation of the sieve size at the fiftypercentile
point of accumulation. The curve also shows the
smoothness of the distribution by revealing the
presence of bimodal blends in the sample. Other
plotting techniques include loglog and direct plotting
of micron size versus percentage retained.
Care should be exercised in the analyzing the data in
relation to the length of time the test was run. If the
sample contains a large amount of elongated or nearsize
particles, the test results can be misleading. The
longer the sieving interval, the greater the opportunity
for these problem particles to pass through the sieve's
openings. Ideally each fraction should be inspected
microscopically after sieving to determine the integrity
of the sieve cut point.
