Your browser is no longer supported

For the best possible experience using our website we recommend you upgrade to a newer version or another browser.

Your browser appears to have cookies disabled. For the best experience of this website, please enable cookies in your browser

We'll assume we have your consent to use cookies, for example so you won't need to log in each time you visit our site.
Learn more

Deep soil mixing


Report on the joint BGS/ICE meeting 'Environmental and engineering applications of deep soil mixing', held at the Institution of Civil Engineers on 13 January 1999, by Tamsin Silvester, Arup Geotechnics.


The main speakers at the meeting were Dr Nick O'Riordan of Arup Geotechnics, environmental consultant Mike Smith and Dr Martyn Lambson of British Petroleum. Following the discussion, three brief presentations on current research were given.

Nick O'Riordan outlined the char- acteristics of the deep soil mixing process (DSM) and described its history. DSM was developed by the Japanese, primarily to control liquefaction under seismic loading, although it was also used for temporary retaining walls. Wet grout is injected into an augered hole both as the auger is inserted and withdrawn. No soil is removed as the process remoulds the soil insitu. Some typical characteristics are shown in Table 1.

There is a potential with DSM to produce a 'designer soil' with specific strength, stiffness, permeability and chemical characteristics. However there are still problems and unknowns with the process; it is difficult to know how to control the rate of grout injection and the end product can be very sensitive to ground conditions, more so than CFA piling or jet-grouting.

DSM is not yet a technically mature process and thus previous experience and case histories are very important. A number of applications were described.

Hibino (1996) outlines cellular disposition of DSM columns to 10m depth, used to support a five storey apartment building (Figure 1). Typical exhumed columns are shown in Figure 8. In ground comprising loose sands and sandy silts, settlement was kept to between 30mm and 100mm.

O'Rourke and O'Donnell (1997) describe a DSM retaining wall with plunge steel sections used as temporary works for a cut and cover tunnel on the Boston Central Artery. DSM buttresses and jet-grouting were used to improve the ground on the passive side of the wall and limit movement (Figure 2).

DSM retaining structures are generally only used for temporary works and there can be problems with interlock shear, movement control and toppling failure. Obtaining consistent mix characteristics in the field is a major area of difficulty with DSM.

In general, UCS of DSM insitu is proportional to the Young's Modulus and durability and inversely proportional to permeability. Thus variations in strength of the mixed soil will result in similar variations in these parameters. Where reported in the literature, unconfined compressive strength (UCS) of cores was found to be highly variable with coefficient of variability of around 40% commonly found.

Much research has been carried out on DSM in recent years. Dong et al (1996), conducted model tests using 400mm diameter counter-rotating blades of varying blade thickness. Examining the effect of degree of mixing in a clay-sand with UCS of 29kPa they show that:

The degree of mixing improves at higher rotary speed.

Thin blades generally achieve better mixes and therefore higher UCS.

UCS is proportional to blade speed.

UCS is proportional to total number of blade revolutions (Figure 3).

This indicates that UCS is inversely proportional to the rate of penetration, ie a slow installation rate at high blade speed is needed to approach the strength of the 'perfect' mixture.

Masuda et al (1996) report on various case studies of ground deformation due to DSM. One example measured horizontal deformation due to DSM in a firm plastic clay adjacent to railway lines. 50mm lateral deflection was recorded with DSM 5m from the rails. This indicates that the method can weaken soil in a temporary condition (Figure 4). The installation sequence is therefore important and monitoring essential.

It is also vital to measure grout strength both in the field and in the laboratory. Asano et al (1996) report on the differences in strength of field and lab mixes. For example 0.3 to 0.5 field/lab strength with E50=50 to 300 UCS (field) for FGC (flyash, gypsum, cement) grout and E50=140 to 500 UCS (field) for CDM (cement deep mixing) grout (Figure 5).

Calabresi et al (1994) describes some 'dry' DSM work in a 20m deep quarry backfilled with mine waste of cu=55kPa. The mine waste was pre-augured to reduce heterogeneity and cement grout was introduced under compressed air. With a cement/soil ratio of 0.2, undrained strength of the mixed waste varied widely between less than 0.25 and 1.25MPa, on the basis of hundreds of tests carried out at varying depths.

Yoshida (1996) reports on the direct shear testing of interface and intra-column cores from 1m diameter DSM columns installed in loose silt. The interlock shear strength was only 70% of the column strength. Importantly, strength was also shown to be time-dependent and zero strength was noted if the lapping columns were installed greater than six days apart (Figure 6).

Asano et al (1996) also describe FGC and CDM grouts in a silty clay of cu=50kPa with 1:1 water/solids ratio and cement/soil ratios between 0.03 and 0.08. UCS generally increased gradually with curing time but was dependent on the mix components and slow pozzolanic reactions of flyash and gypsum components (Figure 7).

Kujala et al (1996) describe how the organic content in the grout mixes influenced strength, permeability and other characteristics of DSM columns. They report that the type of organics is important but that there is little improvement in strength if the humus content exceeds 2%.

O'Riordan concluded by emphasising that DSM is a potentially useful ground improvement tool but suffers from a lack of integrated guidance on its use. Applications are highly site-specific and extrapolation of results requires considerable care. There is currently no satisfactory specification for this type of work, although the CFA section of the ICE specification for piling works is the closest 'off-the-shelf' document, requiring little adjustment to suit DSM.

Trial mixes, both insitu and in the laboratory, are essential. Trial test results form the statistical basis for strength, durability and permeability characteristics, but detailed construction monitoring records are required, similar to those for CFA piling (injection rate, auger rpm, penetration rate etc). Laboratory tests on cored samples from complete DSM columns are essential. Experience shows that large numbers of samples are required to verify design assumptions. Trial loading should also be carried out to test strength and columns can be exposed to verify construction (Figure 8).

O'Riordan also advised that, when DSM is used in contaminated sites, regulatory monitoring and licensing may be required by the Environment Agency and this should be taken into account when programming site works.

Stabilisation/solidification processes

Mike Smith discussed DSM stabilisation/solidification processes for treating contaminated sites.

Although DSM can provide both stabilisation and solidification, the two are different processes. Stabilisation is the addition of reagents to produce more stable, less mobile constituents. Solidification is the addition of reagents to reduce fluidity and friability and to prevent access by mobilising agents. Depending on the circumstances, a treatment system may successfully deliver one process but fail to provide the other.

Selection of a treatment process depends on the objectives. For example, is the process required to meet an engineering objective, such as improving the soil strength? Or is it to modify the physical properties of the soil - and is the soil able to retain the modified properties? These objectives can often be achieved but reduction of the availability of contaminants to the environment is also a common objective and this can be more difficult.

Reduction in availability is governed by several factors, including whether the contaminant reacts to form low solubility compounds (chemical fixation), whether the contaminants are adsorbed by mix ingredients and the extent to which a low permeability mass is formed. Selection of stabilisation/solidification processes must take these three factors into account together with the ability of the system to retain the desired properties and performances in the long term (ie anticipated durability).

Although various treatment agents may be used, only those based on cementitious systems are really relevant to DSM. The most common systems employ Portland cement, pozzolans such as flyash, hydraulic slags and lime. Adsorbents are sometimes used to bind organic chemicals prior to solidification. Examples include activated carbon and organophilic clays.

Questions are raised, however, of how to determine the treatability and performance of the process. How should tests be conducted? How can it be proved that the process is performing as intended, particularly long term?

There are potential problems with DSM stabilisation/solidification processes, such as:

Incompatibility between contaminants.

Incompatibility of adsorbents with binder.

Effectiveness of fixation of a single contaminant.

Ensuring contact between the contaminants and the reagents.

Adverse effects of contaminants on setting, strength development and durability.

There are also practical problems of identifying and quantifying the contaminants in the ground. For example, aluminium can be toxic but is not often treated as a contaminant and rarely tested for. Contamination is never spread evenly across a site, with different contaminants concentrated at varying levels in different areas. Good site investigation information is essential to determine what treatment is required and where.

When carrying out DSM remediation, the type and quality of the reagents in the mix is an important issue. Some manufacturers include small quantities, often less than 1%, of a 'magic ingredient'. Smith was sceptical that such a small quantity of a reagent can work effectively in field conditions by finding the contaminant it has to react with.

Mix chemistry is important as tests have shown that the solubility/leachability of some contaminants, for example aluminium, can increase with time after treatment. It would be better if the required chemical reactions occurred before the cement mix set, ie within two hours, but it is not known whether this actually happens.

The US EPA Technical Resource Document (1993) on stabilisation/solidification and its application to waste materials reports that waste containing a large number of contaminants is generally more difficult to stabilise than waste containing only one or a few contaminants. It suggests that, as a general rule, a physical encapsulation process (solidification) may be the best compromise for multi-contaminated waste, whereas a chemical stabilisation process may be the best approach for single contamination or when the contaminants have similar chemical properties. However, if the solidification process breaks down in time, it is possible that the contamination will be re-released into the ground.

Smith stressed that the environmental impact of stabilisation/solidification methods must be considered carefully. It is possible for the contamination to spread off site, for example by dewatering or from liquids used to clean equipment. It is also possible to increase the solubility and hence the leachate potential of some contaminants, for example arsenic and lead, by changing the pH of the surroundings, which may occur with DSM stabilisation/solidification techniques. Figure 9 shows a solubility diagram for simple mineral phases containing lead. In particular the solubility curve for lead hydroxide shows a minimum at pH of around 10.

Treatability studies have been conducted on the effectiveness of stabilisation/solidification processes. Smith quoted the following extracts from an EPA report on the study of remediation technologies (EPA/540/A5-89/001).

Analyses of crushed samples show that there is often hardly any chemical binding which implies that most of the contaminants in the solidified waste are essentially mobile and will eventually leach out.

The treated waste contained large (up to 50mm) inclusions of untreated waste, this was overcome by screening to 1-2mm.

The solidified mass was porous, mixing was not highly efficient, and brownish aggregates passed through the process unchanged.

Organics contaminants were not immobilised for the most part.

Microstructural analysis indicated potential long-term durability problems.

However, treatability studies require representative samples and it is difficult to know how to assess both the short and long term chemical and physical properties of the materials. Should the material be ground up before testing or is it relying on bond for its performance? Although it is easy to focus on the contamination aspect, it is also important to know about the other properties of the components and how they react in a cement mix.

Smith said that he was deliberately sceptical about DSM stabilisation/solidification processes. Because his background was in cement and concrete chemistry he was very cautious about the idea of treating highly contaminated materials with cement-based systems under poorly controlled field conditions. He backed this up by reference to the EPA report, 1993, which states that well-documented stabilisation/solidification treatability data is scarce and procedures are generally badly described and contain insufficient details. Reliability on the data was not recommended as there was a bias towards successful applications and the failures did not tend to be reported.

Smith emphasised the need to test thoroughly before, during and after DSM stabilisation/solidification processes and to ensure that the contamination is not made worse by the treatment.

Applications of soil mixing - a client's view

Martyn Lambson of British Petroleum said clients need confidence and value for money. Confidence comes from knowledge and understanding of the risks and derives from consideration of three main elements:

The contaminated material to be treated.

The performance and capability of the treatment system.

Previous (similar) experience.

Client confidence must extend to long term performance. It is also important that the client works in partnership with the consultant and contractors so there is a common understanding.

Validation is a key to quality and confidence. Lambson explained the validation process should be incorporated throughout, ideally including:

Pre-installation characterisation including field trial.

Installation records.

Post construction durability and field monitoring.

Lambson advised that field trials should be encouraged as they reduce the number of unknowns and allow testing of a number of mixes before the main work starts. The client also wants to be confident in the control procedures used on site and Lambson backed up O'Riordan's earlier emphasis on keeping detailed site records.

After site work, the client needs proof of work quality, so again keeping detailed records is essential. They also want validation of the treatment's long term durability, how this is achieved and tested, and the long-term monitoring strategy.

Lambson then described the remediation of a former oil shale refinery and surfactant works at Pumpherston in West Lothian, Scotland (Ground Engineering February 1999). Stabilisation formed approximately 15% of the site works at Pumpherston and comprised two main methods, insitu and exsitu treatment, both using a DSM jet grouting rig.

Insitu stabilisation was used for discrete tar lumps buried in localised hotspots within a stable earthen matrix. Exsitu stabilisation involved importing tar from tar ponds and layering it with blaes in a prepared repository formed from clay bunds and the material stabilised.


David Greenwood of GCG opened the discussion by saying that as soil is the prime ingredient of DSM, and it is a highly variable material, variability in the end product should be expected. In his experience, organics can produce losses of strength of up to 10 times. He said that the volume of DSM is important and thick or large blocks need to be created to accommodate any areas of potential weakness or high permeability. Curing time is also important as large volumes create heat quickly and interlocking piles need to be drilled quickly to achieve a good seal. The choice of drilling tool must be chosen carefully to match the type of ground and penetration rate must be controlled to avoid creating enlarged holes requiring more grout.

David Hartwell, a consultant, raised a point regarding the energy used in the mixing process and whether laboratory tests on mixing energy fully represented field conditions. Smith agreed that this was important and he believed mixing techniques are still quite crude and often not enough energy is input in the field to achieve sufficient contact between contaminants and reagents.

Hartwell also asked Lambson why the exsitu treated material at Pumpherston was not mixed above ground, ie. before being placed in the prepared repository. Lambson said volumes of material involved at Pumpherston were too great for mixing before placement. Large quantities of inert blaes material was available on site to use as a stabilisation matrix and placing it in layers and mixing with a jet-grouting auger avoided double-handling of the material on site.

A final question was asked by a representative of Binnie Black and Veatch regarding any experience of leaching of contaminants after DSM near a large body of water, such as the sea. Peter Barker of Bachy Soletanche said that he had worked on a site near an estuary where monitoring after DSM treatment had revealed no evidence of contaminants leaching from the site.

Current research activity

Following the discussion, Dr Abir Al-Tabbaa of Cambridge University, Tony Butcher of BRE, Bob Essler of Keller Ground Engineering and Dr Michael de Freitas of Imperial College gave brief presentations of current research.

Dr Al-Tabbaa outlined her research in time-related performance of soil grout materials and laboratory-scale soil mixing techniques (Ground Engineering November 1998).

Butcher then discussed the progress of the Brite EuRam research programme for EuroSoilStab, investigating development of design and construction methods to stabilise soft organic soils for the construction of road, rail and other infrastructure.

Essler shared his experiences of DSM work worldwide, illustrated with photographs of equipment and plant used and results achieved.

Finally, de Freitas described his team's research into the response of a surface to environmental change, in particular the grain boundaries of different types of silicate minerals.


Al-Tabbaa A, Evans CW and Wallace CJ (1998). Pilot insitu auger mixing treatment of a contaminated site. Part 2: Site trial. Proceedings of the Institution of Civil Engineers, Geotechnical Engineering 131, April, pp89- 95.

Asano J, Ban K, Azuma K and Takahashi K (1996). Deep mixing method of soil stabilization using coal ash.

Calabresi G, Pane V, Rampello S and Bianco O (1994). Geotechnical problems in construction over a thick layer of a mine waste, Proceedings of the first international conference of environmental geotechnics, Edmonton, pp449-454.

CIRIA Special Report. Harris MR, Herbert SM and Smith MA. Remedial treatment for contaminated land. Vol VII Ex-situ treatment of soils, sediments and sludges (CIRIA SP107) and Vol IX In-situ methods of remediation (CIRIA SP109).

Dong J, Hiroi K and Nakamura K (1996). Experimental study on behaviour of composite ground improved by deep mixing method under lateral earth pressure.

Hibino S (1996). Mo nitoring of subsidence of buildings on ground improved by deep soil mixing.

Kujala K, Makikyro M and Lehto O (1996). Effect of humus on the binding reaction in stabilized soils.

Lambson MD and Syratt WJ (1998). Contamination assessment technologies and remediation decision making at a former refinery site, Proceedings of the sixth international FZK/TNO conference, Contaminated soil '98, Thomas Telford, London, pp1289-1298.

Masuda T, Shimizu M and Aizawa, F (1996). Evaluation of ground deformation due to deep mixing in adjacent construction activities.

O'Rourke TD and O'Donnell CJ (1997). Field behaviour of excavation stabilized by deep soil mixing, Journal of Geotechnical and Geoenvironmental Engineering Vol 123 No6, pp516-524.

SGF Report 4:95, Swedish Geotechnical Society, Linkoping.

US EPA Morphology and microchemistry of solidified/stabilised hazardous waste systems (EPA/600/01-89/056)

US EPA report (1993) Study of remediation technologies. Hinsenveld M, final report NATO/CCMS.

US EPA report on the study of remediation technologies (EPA/540/A5-89/001)

US EPA Technical Resource Document (1993) Stabilisation/solidification and its application to waste materials (EPA/600/R-93/012)

Yonekura R, Terashi M and Shibazaki M (1996). Grouting and deep mixing, Proceedings of IS-Tokyo 1996, The second international conference on ground improvement geosystems, Tokyo 14-17 May.

Yoshida S (1996). Shear strength of improved soils at lap-joint-face.

Have your say

You must sign in to make a comment

Please remember that the submission of any material is governed by our Terms and Conditions and by submitting material you confirm your agreement to these Terms and Conditions. Please note comments made online may also be published in the print edition of New Civil Engineer. Links may be included in your comments but HTML is not permitted.