Grain Size and Plasticity

Coarse soils are sorted by grain size, fine soils by how they behave with water.

Before this Lesson

Fill a jar with marbles and about 40% of it is still empty. Fill the same jar with peas and it's still about 40%, because equal balls leave the same share of gaps whatever their size. But pour peas in among the marbles, then sand in among the peas, and the gaps start to disappear. For packing, the mix of sizes matters more than the size. A stack of sieves measures that mix.

Many sizes pack tighter than one

Coarse soil works like the jar. Nearly one size leaves the gaps open. With a wide range of sizes, smaller grains sit in the gaps between bigger ones, and the void ratio from Soil as Three Phases falls. A soil like that is called well graded. The name runs backward from everyday English: a well-graded soil is thoroughly unsorted, with some of every size.

Nearly one size A wide range of sizes
The same big grains, alone and with smaller grains in their gaps

A soil that falls short is poorly graded, and it can fall short in two ways. It can be uniform, with nearly one size, or gap graded, with a wide range but a band of sizes missing.

Two sieves mark the size classes. The No. 4 sieve has openings of 4.75 mm, about the width of a BB, and what it catches is gravel. The No. 200 sieve has openings of 0.075 mm, about the smallest grain your eye can pick out. What gets through the No. 4 but not the No. 200 is sand, and what gets through the No. 200 is fines: silt and clay. The Handbook's soils pages print the 0.075 mm but never the 4.75 mm, so that's the one to remember.

What the sieves catch

Stack the sieves with the coarsest on top and a pan at the bottom, pour in an oven-dried sample, and shake. Each grain falls until it meets an opening too small for it, so the mass on a sieve is only the grains that made it that far and no farther. That's the mass retained. The number you usually want is how much got through, and anything that got through a sieve is sitting somewhere below it. The Handbook prints no procedure for this, but it's only bookkeeping.

No. 4, 4.75 mm No. 10, 2.00 mm No. 20, 0.85 mm No. 40, 0.425 mm No. 60, 0.25 mm No. 100, 0.15 mm No. 200, 0.075 mm Pan 40 g 90 g 90 g 110 g 60 g 40 g 50 g 20 g 330 g stayed 170 g passed the No. 40
A 500 g sample after shaking, with the No. 40 sieve picked out

Take the No. 40 sieve. Its own catch is 110 g, but the grains on the sieves above never got through it either, so what stayed up is the cumulative mass retained: 40+90+90+110=33040 + 90 + 90 + 110 = 330 g. The other 170 g got through, so 34% of the sample passes the No. 40.

percent passing=total mass−cumulative retainedtotal mass×100\text{percent passing} = \frac{\text{total mass} - \text{cumulative retained}}{\text{total mass}} \times 100
Worked Example Reducing the 500 g sample
1 of 3

Find the percent passing each sieve for the sample in the figure, and its percentages of gravel, sand, and fines.

Mass retained on each sieve
SieveOpening (mm)Retained (g)
No. 44.7540
No. 102.0090
No. 200.8590
No. 400.425110
No. 600.2560
No. 1000.1540
No. 2000.07550
Pan—20

The gradation curve

Plot percent passing against size and the whole analysis becomes one line, the gradation curve. The size axis is logarithmic, because grain sizes span several powers of ten. By long custom it also runs backward, with the largest size on the left. That part is only convention.

10 1 0.1 0.01 0 20 40 60 80 100 Grain size (mm), log scale Percent passing by weight No. 4 No. 200 Gravel Sand Fines D60 D30 D10
The gradation curve of the 500 g sample

Three points on that curve have names. D60D_{60} is the grain size that 60% of the soil by weight passes, and D30D_{30} and D10D_{10} are the same idea at 30% and 10%. The Handbook's Phase Relationships section defines them just that way.

Those percentages rarely land on a sieve, so you interpolate between the two sieves on either side. The curve runs straight between sieves on its log axis, so going a fraction ff of the way up in percent means going ff of the way across in log size. On a log axis, going part of the way across means multiplying. Halfway from 0.1 mm to 1 mm is 0.1×100.5=0.3160.1 \times 10^{0.5} = 0.316 mm, and the exponent is how far across you've gone. Between two sieves it's the same move:

f=P−PlowPhigh−PlowDP=Dlow(DhighDlow)ff = \frac{P - P_{\text{low}}}{P_{\text{high}} - P_{\text{low}}} \qquad D_P = D_{\text{low}} \left( \frac{D_{\text{high}}}{D_{\text{low}}} \right)^{f}

Here PP is the percentage you want, and "low" and "high" mark the sieves on either side of it. Straight-line interpolation on the sizes themselves is a common shortcut. On this soil it runs 3% to 8% high, which is usually close enough, but the form above is the one that matches the curve.

Worked Example The three sizes of the 500 g sample
1 of 3

From the reduced analysis, find D60D_{60}, D30D_{30}, and D10D_{10}.

Percent passing
SieveOpening (mm)Passing (%)
No. 44.7592
No. 102.0074
No. 200.8556
No. 400.42534
No. 600.2522
No. 1000.1514
No. 2000.0754

Two numbers for the shape

A mix of sizes can go wrong in two ways: the range can be too narrow, or it can be wide with a band missing. So the curve gets two numbers, both in the Handbook's Phase Relationships section, which calls the second the coefficient of "concavity (or curvature)".

CU=D60D10CC=(D30)2D10D60C_U = \frac{D_{60}}{D_{10}} \qquad C_C = \frac{(D_{30})^2}{D_{10} D_{60}}

The coefficient of uniformity CUC_U says how far apart the sizes are. If every grain were the same size it would be exactly 1, and it can't go lower. The coefficient of curvature CCC_C checks the middle. It equals 1 when D30D_{30} sits exactly halfway between D10D_{10} and D60D_{60} on the log axis, and a missing band shoves D30D_{30} away from there.

The Handbook's Unified Soil Classification table sets the pass marks, writing the two coefficients there as Cu and Cc. They're conventions drawn from experience, not laws. A soil has to clear both tests. Miss either one and it's poorly graded.

What well graded takes
SoilCUC_UCCC_C
Gravel≥4\geq 41 to 31 \text{ to } 3
Sand≥6\geq 61 to 31 \text{ to } 3

Adapted from the FE Reference Handbook, ASTM D2487-11 Unified Soil Classification System table

Sand needs six. Gravel gets by on four.

Worked Example Grading the 500 g sample
1 of 4

The sample is mostly sand, with D60=1.03D_{60} = 1.03 mm, D30=0.356D_{30} = 0.356 mm, and D10=0.114D_{10} = 0.114 mm. Is it well graded?

One symbol, two jobs

CCC_C comes back in Compressibility and Preconsolidation as the compression index, a different quantity entirely. The Handbook prints both with the same symbol and doesn't comment. If the question is about sieves, it's curvature. If it's about settlement, it's compression.

What the shape says

The coefficients are the curve's shape turned into numbers, so you can read one from the other.

10 1 0.1 0.01 0 20 40 60 80 100 Grain size (mm), log scale Percent passing Uniform Well graded Gap graded missing sizes
Three soils with the same axes

The uniform sand is nearly vertical: going from 10% to 60% passing barely moves the size, so D60D_{60} and D10D_{10} sit close together and CUC_U is only 1.8. Its CCC_C of 1.14 is in range, and that doesn't rescue it. Uniform sounds like a virtue, but it means the gaps have nothing to fill them.

The gap-graded soil runs flat across a shelf, because no grains fall in that band of sizes. Its CUC_U is a healthy 17, and only its CCC_C of 0.16 gives it away. The well-graded soil is the long, smooth sweep between them.

D10D_{10} also has a name of its own, the effective size, though the Handbook's soils pages don't use the phrase. It's simply the size 10% of the soil is finer than. The name comes from water: the finest grains choke the narrow throats between the bigger ones, so that tenth has an outsized say in how easily water gets through.

Checkpoint
1 of 3

A clean sand has CU=1.3C_U = 1.3. Which of these could be its gradation curve? All three use the same axes as before.

A.
10 0.01 100 0 Size (mm) % passing
B.
10 0.01 100 0 Size (mm) % passing
C.
10 0.01 100 0 Size (mm) % passing

Below the No. 200, water takes over

Sieves stop at 0.075 mm, and about there, size stops being the useful question. A sand grain is a chunk of rock, heavy for its skin, and it sits on its neighbors whatever the water is doing. A clay particle is a thin plate, nearly all skin. A gram of 1 mm sand has about as much surface as half a business card. A gram of the most active clay has more than a tennis court, and all that surface holds on to water. So fine soils are described by what water does to them, not by how big they are.

Work water into some dried clay a little at a time. At first it crumbles when you try to shape it. With more water you can mold it like putty. With more still, it slumps and flows like thick soup. Those changes happen at water contents you can measure, called the Atterberg limits:

  • The plastic limit, PLPL, is the water content where the soil goes from semisolid to plastic. In the lab, it's where a thread of soil rolled down to 1/8 in starts to crumble.
  • The liquid limit, LLLL, is the water content where it goes from plastic to liquid. In the lab, it's where a groove cut in a pat of soil closes after 25 drops of a standard cup.

These are water contents in exactly the sense of Soil as Three Phases, ω=Ww/Ws\omega = W_w / W_s. An LLLL of 60 means the water weighs 60% as much as the dry solids. No grain gets measured at all.

PI = LL − PL PL LL Semisolid Plastic Liquid crumbles molds like putty flows like thick soup Water content, ω, rising
One soil's states as its water content rises

The width of the plastic range is the plasticity index:

PI=LL−PLPI = LL - PL

The Handbook's Phase Relationships section prints this equation and names both limits, but it never says what they mean.

Checkpoint
1 of 3

What do the liquid and plastic limits of a soil describe?

A.
The largest and smallest grain sizes found among its fines
B.
How angular its grains are, from the roundest to the sharpest
C.
How much electrical charge its clay particles carry on their surfaces
D.
The water contents at the boundaries between its liquid, plastic, and semisolid states

What a high PI says

A plastic soil's particles slide past each other without losing their grip, and the water films on their surfaces make that possible. So the plastic range depends on how much active surface the soil has. Grains with little of it go from crumbling to soup with almost no range between, and grains with none have no range at all.

A wide range comes with a package of behavior. A soil that can hold that much water tends to swell when it wets, shrink when it dries, and squeeze under load, and it gives its water up slowly.

Checkpoint
1 of 2

Which of these soils would you expect to have the highest plasticity index?

A.
A fat clay
B.
A clean sand
C.
A silt
D.
A well-graded gravel

With some typical-looking numbers, the pattern looks like this. A clean sand has no plastic range, so it's called non-plastic, and the Handbook's AASHTO table says to use 0 for its PIPI.

Clean sand Silt Lean clay Fat clay no plastic range, PI = 0 PI = 5 PI = 20 PI = 50 PL LL 0 20 40 60 80 Water content, ω (%)
Plastic ranges of four soils, with illustrative limits

All of these numbers feed the next Lesson. USCS Classification takes the two sieves, the two coefficients, and the two limits, and turns them into a two-letter name.

Recap

  • The No. 4 (4.75 mm) and No. 200 (0.075 mm) sieves split gravel, sand, and fines. Percent passing comes from the running total retained, not one sieve's catch.
  • D60D_{60} is the size 60% of the soil passes, interpolated between sieves on the log size axis.
  • CU=D60/D10C_U = D_{60} / D_{10} and CC=(D30)2/(D10D60)C_C = (D_{30})^2 / (D_{10} D_{60}). Well graded takes CCC_C from 1 to 3, with CU≥6C_U \geq 6 for sand or CU≥4C_U \geq 4 for gravel.
  • A CUC_U near 1 is uniform, and a CCC_C outside 1 to 3 points to a gap-graded or otherwise poorly graded soil. D10D_{10} is the effective size.
  • LLLL and PLPL are water contents, not sizes, and PI=LL−PLPI = LL - PL is the plastic range. The more active clay, the higher the PIPI.