← Course 2: Reading Soil-Water Behaviour

Lesson 1: Measuring suction

Unsaturated Soil Mechanics · Course 2 · about 75 minutes

You will learn

Every curve starts with a measurement

In Course 1 you calculated suction from simple models. In practice, suction is measured or controlled, and every soil-water curve you will read in this course is built from those measurements. Before you trust a curve, you need to know how its numbers were obtained and what each method can and cannot see.

By the end of this lesson, you will be able to:

Name the main methods for measuring or controlling suction, and state whether each one gives matric or total suction and over roughly what range.
Explain how the axis-translation technique works, calculate the suction it applies, and recognise its main limitations.

You will need a calculator. Suctions are in kPa; 1000 kPa = 1 MPa.

A quick reminder

Matric, osmotic and total suction

Recall from Course 1 that matric suction, s = u_a − u_w, comes from capillary effects and water held on grain surfaces, and that dissolved salts add an osmotic suction. Together they make the total suction.

total suction = matric suction + osmotic suction

This matters for measurement because instruments respond to different parts. Some sense the pressure of the pore water directly (matric). Others sense the humidity of the air in equilibrium with the soil, which depends on both capillary and salt effects (total). A curve labelled simply “suction” may be either, so always check.

The main methods

What each method measures

Methods fall into two groups. Measuring methods read the suction that is already in a sample or in the ground. Controlling methods impose a chosen suction and wait for the soil to come into equilibrium with it.

Method	Measures or controls	Typical range	Typical time per reading
Standard tensiometer	Measures matric suction	0 to about 80 kPa	Minutes to hours
High-capacity tensiometer	Measures matric suction	Up to about 1.5 MPa	Minutes to hours
Axis translation (pressure plate, suction-controlled cells)	Controls matric suction	Up to about 1.5 MPa, limited by the ceramic disc	Days to weeks per step
Filter paper, in contact with the soil	Measures matric suction	Wide, best above a few tens of kPa	About 1 to 2 weeks
Filter paper, not in contact	Measures total suction	Wide, best at high suctions	About 1 to 2 weeks
Chilled-mirror (dew-point) device or psychrometer	Measures total suction	Roughly 1 MPa upwards	Minutes

Ranges are indicative. They depend on the device, its calibration and the care taken. Always check the manufacturer’s information and the standard being followed.

Why a standard tensiometer stops at about 80 kPa

A tensiometer holds water in tension. As the tension approaches one atmosphere, dissolved air comes out of solution and the water cavitates, so the reading is lost. High-capacity tensiometers (Ridley and Burland, 1993) delay this with very careful saturation and a fine porous stone, but they still cavitate eventually.

The laboratory workhorse

The axis-translation technique

Measuring a large negative water pressure is hard because the water cavitates. Hilf (1956) avoided the problem by raising the pore-air pressure instead. If both pressures go up by the same amount, the suction (the difference) is unchanged, but the water pressure is now positive and easy to measure or control.

A simplified axis-translation cell. The ceramic disc stays saturated, so water can flow through it, but air cannot pass until its air-entry value is exceeded.

The key part is the high-air-entry (HAE) ceramic disc. Its pores are so fine that, once saturated, it lets water through but blocks air, up to a pressure difference called its air-entry value (for example 100, 500 or 1500 kPa). The suction you apply must stay below that value.

s = u_a − u_wThe same definition as in Course 1. Only the reference level of both pressures has moved.

Putting numbers in

The pressure plate and the suction-controlled triaxial

In a pressure plate, many specimens sit on one large ceramic plate in a sealed chamber. The water below the plate drains to the atmosphere, so u_w = 0. The applied air pressure is then the suction. In a suction-controlled triaxial or oedometer cell, both pressures are set separately, together with the cell (confining) pressure.

Worked examples

(a) Pressure plate

Air pressure 200 kPa, water draining to atmosphere (u_w = 0). s = 200 − 0 = 200 kPa.

(b) Suction-controlled triaxial

Cell pressure σ₃ = 300 kPa, u_a = 200 kPa, u_w = 50 kPa.

s = 200 − 50 = 150 kPa

net confining stress = σ₃ − u_a = 300 − 200 = 100 kPa

The specimen experiences a suction of 150 kPa and a net confining stress of 100 kPa: the two stress variables from Course 1.

Equilibrium. After each change of suction, water flows out (drying) or in (wetting) through the disc until the soil reaches the new suction. You know equilibrium is close when the volume gauge stops changing, within a criterion you set before the test. In fine-grained soils this can take days or weeks for every step.

From practice: waiting is the experiment

In my PhD I controlled suction with the axis-translation technique, both in a pressure plate and in a suction-controlled triaxial set-up. The pressures are set in minutes. The real work is waiting for the water to stop moving at every step and checking that it really has stopped. Changing the suction gradually, step by step, is also closer to how water enters and leaves the ground, which is why one test on a compacted soil could run for around three months.

Know the limits

When axis translation can mislead

The disc sets the ceiling. Above its air-entry value, air breaks through and control is lost.
Air diffuses through the water. Over long tests, dissolved air passes through the disc and appears as bubbles below it. These look like water volume change unless they are flushed out and accounted for.
Near saturation it is questionable. The method assumes the pore air is continuous. When the air is in isolated bubbles, raising the air pressure compresses the bubbles instead of translating the pressures.
Slow equilibrium can be cut short. Ending a step too early records a water content that belongs to a lower (or higher) suction than the one applied.

Wherever you are

If your laboratory has no axis-translation equipment, the filter paper method needs only a good balance, an oven, sealed containers and a stable temperature. It is slower and needs care, but it is used worldwide and is covered by a standard (ASTM D5298).

Matching method to question

Three questions to ask first

Which suction do you need? For most strength and volume-change problems, matric suction. Where salt content matters, you may need total suction too.
What range? A compacted fill near the surface in a wet climate may sit at tens of kPa. A dry clay in an arid climate may be at several MPa. No single method covers everything well.
Field or laboratory, and how quickly? Tensiometers can monitor the ground continuously. Filter paper and axis translation are laboratory methods that need days to weeks.

For a full soil-water curve, laboratories often combine methods: axis translation or tensiometers for the low and middle range, and a dew-point device or non-contact filter paper for the high range.

Separating the parts

Finding the osmotic suction

Because contact filter paper responds to matric suction and non-contact filter paper to total suction, measuring both on the same soil gives an estimate of the osmotic part.

Worked example

Filter paper tests on a clay give a total suction of 1200 kPa (non-contact) and a matric suction of 950 kPa (contact).

osmotic suction = total − matric = 1200 − 950 = 250 kPa

About 250 kPa of the total comes from dissolved salts. If the pore-water chemistry changes, for example by leaching, this part can change even when the water content does not.

Treat small differences with caution

Each measurement has its own scatter. If the difference between total and matric suction is smaller than the uncertainty of the methods, do not report it as a meaningful osmotic suction.

Reading other people’s data

What to look for in a test report

When you are given suction data, check that the report states:

the method, and whether the result is matric or total suction
the equilibrium criterion and how long each reading or step took
the sample: undisturbed, compacted or reconstituted, with its density and water content
whether the path was drying or wetting (you will see why this matters in Lesson 2)
the temperature, and any calibration used

If these are missing, the numbers may still be useful, but your conclusions must be more cautious.

Your turn

Work with suction measurements

Task (20 minutes)

In a pressure plate, the air pressure is 500 kPa and the water drains to atmosphere. What suction is applied? Would a 300 kPa ceramic plate be suitable?
In a suction-controlled triaxial test, σ₃ = 400 kPa, u_a = 250 kPa and u_w = 100 kPa. Find the matric suction and the net confining stress.
Non-contact filter paper gives 2.1 MPa and contact filter paper gives 1.8 MPa on the same specimen. Estimate the osmotic suction.
A colleague wants to monitor suction at 0.5 m depth beneath a road in a hot, dry season, where values above 500 kPa are expected. They suggest a standard tensiometer. What would you say?

Show the answers

1. s = 500 − 0 = 500 kPa. A 300 kPa plate is not suitable: the applied suction exceeds its air-entry value, so air would pass through the plate.

2. s = 250 − 100 = 150 kPa. Net confining stress = 400 − 250 = 150 kPa.

3. Osmotic ≈ 2.1 − 1.8 = 0.3 MPa (300 kPa), provided this difference is larger than the uncertainty of the two measurements.

4. A standard tensiometer cavitates at about 80 kPa, so it would lose its reading. Consider high-capacity tensiometers, which still need frequent checks at these suctions, or indirect sensors calibrated for the soil, backed up by samples tested in the laboratory.

Quick check

Five quick questions

Choose an answer to see the explanation.

Key points

What to take away

Some methods measure suction; others control it and wait for equilibrium.
Tensiometers and contact filter paper give matric suction. Dew-point devices, psychrometers and non-contact filter paper give total suction.
Standard tensiometers cavitate at about 80 kPa. High-capacity tensiometers reach about 1.5 MPa, but not indefinitely.
Axis translation raises u_a so that u_w stays positive. The suction is still u_a − u_w, and it must stay below the disc’s air-entry value.
Equilibrium is slow in fine soils. Ending a step early gives a wrong point on the curve.
Always check the method, the type of suction, the equilibrium criterion and the sample before using the data.

Up next: Lesson 2

The soil-water characteristic curve. Put the measurements together into a curve, convert between water-content measures, find the air-entry value and see why drying and wetting give different curves.

References

ASTM D5298 Standard Test Method for Measurement of Soil Potential (Suction) Using Filter Paper. West Conshohocken, PA: ASTM International.
Fredlund, D.G., Rahardjo, H. and Fredlund, M.D. (2012) Unsaturated Soil Mechanics in Engineering Practice. Hoboken, NJ: John Wiley & Sons.
Hilf, J.W. (1956) An Investigation of Pore-Water Pressure in Compacted Cohesive Soils. Technical Memorandum 654. Denver, CO: U.S. Bureau of Reclamation.
Lu, N. and Likos, W.J. (2004) Unsaturated Soil Mechanics. Hoboken, NJ: John Wiley & Sons.
Ng, C.W.W. and Menzies, B. (2007) Advanced Unsaturated Soil Mechanics and Engineering. Abingdon: Taylor & Francis.
Ridley, A.M. and Burland, J.B. (1993) ‘A new instrument for the measurement of soil moisture suction’, Géotechnique, 43(2), pp. 321–324.

Further learning

Read: the chapters on suction measurement in Lu and Likos (2004) or Fredlund, Rahardjo and Fredlund (2012) compare the methods in more detail.
Read: Ridley and Burland (1993) is a short paper describing the first high-capacity tensiometer.
Look: if your laboratory has a pressure plate or a suction-controlled cell, ask to see how the ceramic disc is saturated and how air is flushed from below it.

DAMILOLA AKINNIYI