Deep soil mixing is a ground improvement technique in which a binder — usually cement, sometimes with lime or other additives — is mechanically blended into the soil in place, forming columns of treated material that are stiffer and stronger than the ground they replace. The columns reduce compressibility and increase shear strength without excavating and replacing the soil.
Two variants exist, distinguished by how the binder is delivered.
Dry mixing injects binder in powder form and relies on water already present in the soil to hydrate it. It is used in saturated soils, where sufficient water exists.
Wet mixing injects binder as a cement–water slurry. It can be used at any degree of saturation, including in soils that do not contain enough water for dry mixing to work.
The design problem is not the same as designing a concrete mix. The soil is a component of the mixture, its properties vary along the column, and the quantity of binder is specified indirectly through parameters the contractor can actually control at the mixing head.
A contractor cannot control the strength of a mixed column directly. They control the binder factor and the volume ratio, and strength follows from them.
The Phase Relationships
Mixture design is built on a phase diagram of the mixed material, and the notation follows FHWA practice.
Before mixing, the soil consists of solids, water, and air. The binder is then added — as dry powder, or as a slurry containing its own water. After mixing, the resulting material contains soil solids, binder solids, and the combined water.
| Quantity | Meaning |
|---|---|
| Vs, Ws | Volume and weight of soil solids |
| Vw,soil, Ww,soil | Volume and weight of water in the soil before mixing |
| Va | Volume of air before mixing |
| Vb, Wb | Volume and weight of binder solids |
| Vw,slurry, Ww,slurry | Volume and weight of water in the slurry, for wet mixing |
| Vsoil, Wsoil | Volume and weight of soil before mixing |
| Vmix, Wmix | Volume and weight of the mixed material |
From these follow the parameters that define the mixture.
The Five Mixture Design Parameters
| Parameter | Symbol | Definition | Controlled by |
|---|---|---|---|
| Binder factor | α | Weight of binder per unit volume of soil treated | Dry mixing |
| In-place binder factor | α in-place | Binder weight per unit volume of mixed material | Derived |
| Binder content | aw | Binder weight as a percentage of dry soil weight | Derived |
| Total water-to-binder ratio | wT:b | All water in the mix divided by binder weight | Target from strength |
| Volume ratio | VR | Volume of slurry injected per unit volume of soil | Wet mixing |
The distinction between the binder factor and the in-place binder factor is a frequent source of confusion. The first is referenced to the volume of soil before treatment; the second to the volume of the mixed material after treatment, which is larger because binder and slurry have been added. Specifications and field records do not always state which is intended, and the difference is not small.
Two further quantities are needed: the specific gravity of the binder Gb, and the dry unit weight and natural water content of the soil, γd,soil and w, which determine how much water is already available.
The Central Design Relationship
The link between mixture proportions and strength is the total water-to-binder ratio.
wT:b = ( Ww,soil + Ww,slurry ) / Wb
Unconfined compressive strength decreases as this ratio increases, following the same underlying logic as the water–cement ratio in concrete: more water per unit of binder produces a weaker cemented matrix. The relationship is presented in FHWA guidance as a design curve of qu against wT:b.
The design sequence therefore begins with the required strength and works backwards:
Required qu → wT:b from the design curve → binder factor and volume ratio → binder quantity per column
The critical point is that the soil’s own water counts. A soil at high natural water content already contributes a large numerator, so achieving a given wT:b requires more binder than the same target would need in a drier soil. This is why the same specified dosage produces different strengths in different parts of the same site, and why mixture design has to be referenced to the soil actually present rather than to a single site-wide value.
Dry Mixing: The Design Steps
In dry mixing the contractor meters the delivery rate of dry binder, which means the binder factor α is under direct control.
- Establish the required unconfined compressive strength qu for the design.
- Read the corresponding total water-to-binder ratio wT:b from the design curve.
- Using Gb, wT:b and the dry unit weight of the soil, calculate the binder factor α, the in-place binder factor, and the binder content aw.
- Calculate the binder weight for the column: Wb = Vsoil · α
- Calculate the unit weight of the resulting mixture γmix, which is required for the analysis.
Because dry mixing depends on the soil supplying the water for hydration, it requires a saturated soil with sufficient natural water content. In drier ground the binder does not hydrate fully and the target strength is not achieved regardless of dosage.
Wet Mixing: The Design Steps
In wet mixing the contractor controls two variables: the water-to-binder ratio of the slurry itself, w:b, and the volume ratio VR — how much slurry is injected per unit volume of soil.
- Establish the required qu and read wT:b from the design curve as before.
- Select the slurry water-to-binder ratio w:b appropriate to the mixing equipment and pumping requirements.
- From Gb, wT:b, w:b, the soil dry unit weight and its degree of saturation, calculate the volume ratio VR required to achieve the target, together with α, the in-place binder factor, aw, and the dry unit weight of the slurry.
- Calculate the binder weight for the column: Wb = Vsoil · α
- Calculate the unit weight of the mixture, which for wet mixing includes the added slurry volume through the volume ratio.
The degree of saturation enters the wet mixing calculation explicitly, because air voids in an unsaturated soil are partly filled by the injected slurry. A single expression for a saturated soil does not transfer to a partially saturated one.
From Mixture to Engineering Properties
Mixture design produces a target strength. Analysis requires stiffness and shear strength.
Shear strength is taken from the unconfined compressive strength on the basis of a total stress model with the friction angle set to zero:
Sdm = qu / 2
The deformation modulus of the mixed material Edm is obtained from correlations with qu, with separate relationships for wet and dry mixing given in FHWA guidance. These correlations carry considerable scatter, which is one reason laboratory trial mixes and field verification coring are specified on projects of any size.
The Composite Modulus
A treated zone is not a uniform improved material. It is a grid of stiff columns within untreated soil, and the analysis must represent both.
The improved zone is treated as a composite medium whose modulus is obtained by area weighting:
Ecom = ( n · Ac · Ec + ( A − n · Ac ) · Es ) / A
where A is the plan area of the treated block, n the number of columns within it, Ac the cross-sectional area of one column, Ec the column modulus, and Es the modulus of the untreated soil.
The derivation assumes that columns and surrounding soil undergo equal vertical deformation. That assumption is reasonable beneath a rigid raft; it becomes less reliable beneath a flexible foundation, where the stiff columns can punch upward into the structure and the soil between them carries less than the area weighting implies.
Once the composite modulus is established, the soil profile is re-layered so that the treated zone becomes one or more layers with composite properties, and settlement analysis proceeds exactly as for untreated ground.
The improvement is not modelled as columns. It is modelled as a stiffer layer that happens to contain columns.
What Deep Soil Mixing Is Used For
| Application | Mechanism |
|---|---|
| Settlement reduction | Composite modulus of the treated zone is higher than the untreated soil |
| Bearing capacity increase | Columns carry a share of the load through shaft and base resistance |
| Liquefaction mitigation | Stiff columns attract seismic shear stress away from the surrounding soil |
| Slope and excavation stability | Increased shear strength within the treated mass |
| Cut-off and containment | Overlapping columns form a low-permeability barrier |
The bearing capacity application borrows its mechanics from deep foundations: a column carries load through shaft friction and base resistance in the same way a pile does, and the same capacity methods apply when the columns are stiff enough to act as inclusions rather than as improved ground.
The liquefaction application deserves a note. The mechanism is shear stress redistribution: because the columns are much stiffer than the soil, they attract a disproportionate share of the earthquake-induced shear stress, reducing the cyclic stress ratio in the soil between them. The reduction is quantified through a stress reduction factor derived from the area ratio and the shear modulus ratio of columns and soil.
Verification
Because the mixed material is created in place from a variable input, verification is more central to deep soil mixing than to most construction materials.
Laboratory trial mixes established before production give the relationship between binder dosage and strength for the specific soils on the site. Wet-grab samples taken from the fresh column during production and cored samples taken from the hardened column confirm what was actually achieved. Continuous monitoring of the mixing machine records binder delivery rate, rotation speed, and penetration and withdrawal rates.
Strength in a mixed column is variable by nature — coefficients of variation well above those accepted for concrete are normal — which is why design values are selected with that scatter in mind rather than taken as the mean of trial results.
Common Errors in Deep Soil Mixing Design
| Error | Consequence |
|---|---|
| Confusing binder factor with in-place binder factor | Binder quantity misstated, in the direction of under-dosing |
| Ignoring the soil’s natural water content in wT:b | Target strength not achieved in wetter zones of the site |
| Specifying dry mixing in unsaturated soil | Insufficient water for hydration; strength not developed |
| Applying one mixture design across a variable site | Consistent dosage, inconsistent strength |
| Using the composite modulus beneath a flexible foundation without checking the equal-strain assumption | Settlement underestimated |
| Taking the design strength as the mean of trial mixes | Scatter in mixed materials is larger than in concrete |
| Treating Edm correlations as precise | Stiffness carries more scatter than strength |
Where SETAF2018 Fits
SETAF2018 performs deep soil mixing mixture design following the FHWA methodology, within the same model that carries the foundation analysis.
The mixture is defined on the material itself. Entering the required unconfined compressive strength qu, the slurry water-to-binder ratio w:b, and the specific gravity of the binder Gb, together with the natural water content and dry and natural unit weights of the existing soil, the software calculates the binder factor α, the in-place binder factor, the binder content aw, the total water-to-binder ratio wT:b, the volume ratio VR, the dry unit weight of the slurry, and the resulting unit weight of the mixture γmix. Both dry and wet mixing are supported, with the degree of saturation entering the wet mixing calculation.
The shear strength of the mixture is obtained as qu/2 for a total stress model with the friction angle set to zero, and the deformation modulus Edm from the FHWA correlation. These become the column properties used in analysis.
Columns are placed beneath rectangular, circular, or polygonal foundations, individually by coordinate or as an automatically generated group at a chosen spacing. The improved zone is then re-layered with composite properties Ecom and νcom derived by area weighting, and settlement and bearing capacity are calculated through that profile. Where the ground is liquefiable, the effect of the columns is evaluated through the stress reduction factor for both compatible and incompatible deformation cases, connecting to the liquefaction assessment and the wider liquefaction analysis workflow.
The mixture calculation is available as a dedicated local report showing the water and cement quantities with the equations used, which is the document a contractor needs at the mixing plant rather than a summary of results. Column layouts, quantities, and the wider foundation analysis outputs are produced from the same model, and calculation reports carry the underlying expressions. Where the alternative technique is under consideration, the comparison with jet grouting is set out separately.
Frequently Asked Questions
What is deep soil mixing?
It is a ground improvement technique in which a binder, usually cement, is mechanically blended into the soil in place to form columns of treated material that are stiffer and stronger than the surrounding ground. It does not involve excavating and replacing the soil.
What is the difference between dry and wet mixing?
Dry mixing injects binder as a powder and relies on water already in the soil to hydrate it, so it requires saturated ground. Wet mixing injects binder as a cement–water slurry and can be used at any degree of saturation.
What is the binder factor?
It is the weight of binder added per unit volume of soil treated. The in-place binder factor is the same quantity referenced to the volume of the mixed material rather than the original soil, and the two are not interchangeable.
How is the required cement dosage determined?
From the required unconfined compressive strength, which gives a target total water-to-binder ratio from the FHWA design curve. That ratio, together with the specific gravity of the binder and the dry unit weight and water content of the soil, gives the binder factor and hence the binder weight per column.
Why does the same dosage give different strengths across a site?
Because the soil’s own water counts toward the total water-to-binder ratio. A wetter soil produces a higher ratio and a lower strength for the same binder dosage. Mixture design must be referenced to the soil actually present.
How is a treated zone represented in settlement analysis?
As a composite layer whose modulus is obtained by area weighting the column and soil moduli, assuming equal vertical deformation. The soil profile is re-layered with those properties and the settlement analysis proceeds normally.
How is deep soil mixing verified?
Through laboratory trial mixes before production, wet-grab and cored samples during and after production, and continuous monitoring of binder delivery, rotation speed, and penetration rate at the mixing head. Strength variability in mixed materials is substantially higher than in concrete.
Does deep soil mixing help against liquefaction?
Yes. Because the columns are much stiffer than the surrounding soil, they attract a disproportionate share of the earthquake-induced shear stress, reducing the cyclic stress in the soil between them. The reduction is quantified from the area ratio and the shear modulus ratio.
