Self-Compacting Concrete

The design, production and use of SCC by Rob Gaimster and Noel Dixon

Awareness of self-compacting concrete (SCC) within the construction industry has grown year on year since it was developed in Japan in the late 1980s¹, and the quest for further understanding as to its capabilities and limitations has generated considerable interest in research worldwide. This paper attempts to summarize key aspects and to outline the current situation.

Khayat² defines SCC as ‘a highly flowable, yet stable concrete that can spread readily into place and fill the formwork without any consolidation and without undergoing any significant separation’.

Feature/benefit analysis on the use of SCC would suggest that the following benefits should result:

• increased productivity levels leading to shortened concrete construction time
• lower concrete construction costs
• improved working environment
• improvement in environmental loadings
• improved in-situ concrete quality in difficult casting conditions
• improved surface quality.

Non-vibrated concrete is already commonplace in the construction industry and is used with acceptable results, eg in piling and shotcrete applications. Development of SCC has mainly focused on congested civil engineering structures, and its acceptance within the marketplace has primarily grown through its ability to solve technically difficult casting conditions. It is a niche product — a problem-solver.

Okamura and Ouchi³ have commented on the reduction in the number of skilled workers affecting the quality of construction work in Japan. However, as SCC reduces the dependency of concrete quality on the workforce, further market penetration is expected.

Materials and mix design

Before looking at designing a mix for SCC an understanding is needed of the properties required for self-compaction and how this can be optimized utilizing materials currently available. The two main requirements are for a highly fluid material which has significant resistance to separation.

To achieve a highly mobile concrete a low yield stress is required, and for high resistance to separation a highly viscous material is required. Water can be added to decrease the yield stress but, unfortunately, this also lowers the viscosity. The addition of a superplasticiser will also lower the yield stress and will only lower the viscosity slightly. The viscosity of a mix can be increased by changes in the mix constituents or the addition of a viscosity modifier but this will increase the yield stress of the paste. Thus, it is necessary to find a happy medium between the two parameters. Figure 1 shows the relationship between shear rate and shear stress.

Advances in admixture technology have played a vital part in the development of SCC. Modern superplasticisers (based on polycarboxylic ethers) promote good workability retention and can be added at any stage of the batching cycle. They achieve this through a mechanism of electrostatic repulsion in combination with steric hindrance.

Viscosity modifiers can be added to increase the resistance to segregation while still maintaining a high fluidity, thereby allowing the concrete to flow through narrow spaces.

Many authors have different mix-design theories but all attempt to achieve the above. Most separate the mix into a two-phase design: ‘continuous’, which covers the water admixture, cement and fillers with a particle size less than 0.1mm; and ‘particle’, which considers the coarse and fine aggregate. A few of them are summarized below.

Ozawa’s⁴ ‘General Method’, originating in the late 1980s in Japan, is a very simplified method looking at basic values such as the coarse aggregate content being restricted to 50% of the concrete volume. This method is very conservative giving cement contents in excess of 600kg/m³.

Petersson’s⁵ ‘CBI Method’ examines the overall grading of the combined aggregate, allows for any size of aggregate and considers actual construction criteria. It determines aggregate volumes from which a paste content can be established.

Sedran’s⁶ ‘Compressive Packing Model’ considers the material properties such as bulk density, apparent particle density, absorption and particle size distribution, and uses this information in software models to predict the flow behaviour from blocking and segregation risks. This produces a theoretical optimum mix from the above and this is trialled and modified through laboratory tests.

Saak’s⁷ ‘Segregation Control Theory’ looks at how to optimize material additions to control yield stress, viscosity and the density of a cement-paste matrix. Thus, the rheology of the matrix can be engineered to produce SCC.

Fine particles play an integral part in the design of SCC and similar-sized particles to cement grains, such as pulverized fuel ash, ground granulated blast-furnace slag and silica fume, can be added to the mix to aid the plastic and hardened properties of the concrete. Limestone filler is used extensively on the Continent.

Plastic concrete

There are three main areas to be considered in the concrete’s plastic state: filling ability, resistance to segregation, and passing ability. These properties are looked at in turn, along with methods of assessment8.

Filling ability

This property of fresh SCC concrete is related entirely to the mobility of the concrete. The concrete is required to change shape under its own weight and to mould itself to the restricted formwork in place.

To allow this to occur, the inter-particle friction of the materials must be reduced. This can be achieved in two ways: first, surface tension can be reduced by the inclusion of superplasticisers; and secondly, optimizing the packing of fine particles can be achieved by the introduction of fillers or segregation-controlling admixtures.

Measurement of the plastic properties can be achieved by the following tests — slumpflow or BTRHEOM rheometer.

The slumpflow test utilizes a British Standard slump cone, which is filled in one layer without compaction. The mean spread value in millimetres is recorded. Typical values lie between 650mm and 800mm. The test measures mobility/ deformity under a low rate of shear (self weight). Assessment of segregation can be made subjectively but the test does not completely measure the filling capacity of the SCC in question. A further evaluation can be carried out at the same time. This is the T50 value, which measures the time taken to reach a spread of 500mm. There is a question mark over the value of slumpflow results when viewed in isolation.

With the BTRHEOM rheometer the concrete is considered as a Bingham fluid and its behaviour is determined by the shear yield stress and the plastic viscosity. A low shear yield stress and a limited plastic viscosity value are required.

Resistance to segregation

SCC has to be stable under mobile conditions. Two areas therefore need to be addressed. First, the amount of moveable water needs to be minimized to avoid bleeding. This can be achieved by the use of superplasticisers to reduce the water demand and separation through a well graded cohesive concrete. Secondly, the liquid phase needs to be viscous in nature to be able to maintain the coarse particles in suspension, when mobile. This can be achieved by a high volume of fines in the mix and/or the introduction of a viscous modifier.

Measurement of the plastic properties can be achieved by visual inspection using the slumpflow method or by the GTM stability sieving test. This measures the degree of separation of the coarse and mortar fractions. Ten litres of fresh concrete are placed into a test container. Over a 15min period the coarse aggregate will settle to the bottom. The upper part of the concrete in the container is then wet sieved and the volume of mortar paste calculated. The higher the value the more segregation has occurred.

Passing ability

This is the ability of the concrete to be able to pass round immovable objects in the formwork, such as reinforcement. The need for this ability will depend on the reinforcement arrangement for the individual structures that are cast.

Factors to be considered will be the spaces between the reinforcement, which will influence the selection of the size and shape of the coarse aggregate and the volume of the mortar paste. The more congested the structure, the higher the volume of paste is required compared to the amount of coarse aggregate.

Measurement of the plastic properties can be achieved by adding a J-Ring to the slumpflow to assess the concrete’s passing ability, or by the L-Box test. This is useful in assessing different parameters such as mobility, flow speed, passing ability and blocking behaviour. The apparatus consists of a long, rectangular-section trough with a vertical column/ hopper at one end. A gate fitted to the base of the column allows discharge of the SCC into the trough. Adjacent to the gate is an arrangement of bars which permits an assessment of blocking potential to be made. The flow speed can be measured by the time taken to pass a distance of 200mm (T20) and 400mm (T40). Also, the heights at either end of the trough (H1 and H2) can be measured to determine the concrete’s levelling ability. The test appears to be useful although there is no standardization on the principal dimensions of the equipment.

These tests for assessing the plastic properties of fresh SCC are not a definitive list and at present are not recognized by any standards, but they are the most common tests in current use. The Advanced Concrete and Masonry Centre in Paisley is, however, co-ordinating a European working group which is investigating test methods for SCC.

Hardened concrete

In normal concrete, when vibrated, water will tend to migrate to the surface of the coarser particles causing porous and weak interfacial zones to develop.

If SCC has been well designed and produced it will be homogeneous, mobile, resistant to segregation and able to be placed within formwork without the need for compaction. This will encourage, between the coarse aggregate and mortar phase, minimal interfacial zones to develop. Thus the microstructure of SCC can be expected to be improved, promoting strength, permeability, durability and ultimately longer service life of the concrete. In-situ compressive strengths determined using cores have shown a closer correlation to standard cube strength than conventional concrete. Also, work has indicated that the reduction in compressive strength with an increase in column height is less pronounced, showing good homogeneity of SCC⁹.

Trials were carried out at the RMC Readymix technical centre to examine the hardened properties of SCC, using a total cementitious content of 480kg/m³ at a slumpflow of 700mm and a superplasticiser and a viscosity modifier. The concrete was poured into a U-shaped mould, as detailed in figure 2, with obstructions placed in the unit (shown shaded).

Ultrasonic pulse velocity tests were performed over the unit. Cores were taken to determine the in-situ strength and the density within the structure. The cores were also tested for chloride and oxygen diffusion.

Satisfactory self-compaction of the fresh concrete was confirmed by the consistently high UPV values and density measurements of the core samples taken throughout the unit. The mean estimated in-situ cube strength was 81% of the 28-day cube strength from concrete sampled during casting. The chloride and oxygen diffusion results, significantly less than those required by many specifications, were 0.304 x 10^–12 and
1.44 x 10^–8 respectively.

Production and transport

Owing to the need for the efficient dispersion of fine particles required to produce a homogeneous and stable mix, mixing time compared with normal concrete is increased. In addition, the need for an accurate total moisture content of the mix requires good knowledge of the properties of the materials being used. Consistency of moisture content and particle size distribution from the material supplier is critical. Sand grading and moisture content are particularly important.

SCC has been produced in different types and sizes of batching plants. Evidence from the UK suggests that dry batching is perfectly satisfactory for producing SCC.

SCC is more sensitive compared with normal concrete and if the concrete has not been sufficiently mixed before transportation, slumpflow can be increased due to further dispersion of the superplasticiser through the concrete.

These factors need to be considered after successful trial mixes have been established, owing to the controlled nature of the laboratory environment.

Placement

No special equipment is needed to place SCC; standard pumps and skips can be used.

However, because SCC can be used to reduce construction time, there will be no real advantage in skipping the concrete into place as this time is restricted to the amount of concrete the skip can hold. Generally, while the skip is returned to be filled up, compaction is carried out by poker vibrators for conventional concrete.

This leads to SCC being pumped into place as the main option to save construction time. As with normal concrete, a well designed SCC can be pumped considerable distances without any problems. Pumping from the base of structures is feasible.

Formwork

In order to achieve the benefits of a reduction in construction time, SCC needs to be placed quicker. With no need for vibration of the concrete this can be achieved.

This assumes increasing the rate of rise of the concrete within the structure, which will lead to an increase in hydrostatic pressure on the formwork. This in turn could necessitate the need for formwork redesign to accommodate the theoretical increase in pressure.

One study, however, confirms that the properties of SCC actually give lower form pressures when compared to normal vibrated concrete at the same rate of rise. This is because once the kinetic energy of the fresh concrete has dissipated, the concrete stiffens in a thixotropic manner and so no longer acts as a liquid¹⁰. More research is, however, required on this subject.

In the meantime, it is sensible to design formwork assuming full hydrostatic pressure.

Surface finish

In the UK, surface finish is one of the perceived benefits of SCC leading to a myriad of architectural possibilities. There are several factors, however, which give rise to the final surface finish, including:

• mix design
• workability
• formwork configuration
• formwork material
• mould release agent
• rate of rise
• method of placement.

A series of trials were undertaken at RMC’s technical centre to examine the effects of different formwork materials together with different categories of mould release agent for the same SCC mix. The mix was designed with a total cementitious content of 500kg/m³, a free water/cement ratio of 0.36 and a polycarboxylate superplasticiser and VMA, at a slumpflow of 700mm.

Units constructed as detailed in figure 3 were used to compare eight different combinations of formwork and release agent (fig. 4). Steel and plywood were used as the formwork materials in conjunction with several categories of release agent.

The results of the trials are summarized in table 1. It gives ratings (somewhat subjectively) of the combinations of types of release agent and formwork material, based on the general appearance and the number and size of voids present in an area of 0.06m².

As would be expected, plywood provides a better surface finish than steel. It should also be noted that the type of mould release agent also plays an important role in the finished surface. Surprisingly, the release agents based on vegetable oil gave the poorest results.

Mix design optimization

Moving SCC to mainstream construction

Since the infancy of SCC, a total cementitious content of approximately 500–600kg/m³ has been used, typically achieving strengths in excess of 70N/mm². Usually such high strengths have not been a structural requirement.

One of the main drawbacks to mix designs in current use, however, is their increased cost, attributable in part to the elevated cement contents required and the use of state-of-the-art admixture technology; newer admixtures themselves also contribute, to some degree, to concerns about the specification.

The ability to reduce the total cementitious contents of mixes and to incorporate additions would lower the strength and, more importantly, lower the cost, making SCC a more attractive and competitive proposition for mainstream construction.

In work undertaken by RMC Readymix in conjunction with BRE¹¹, a series of laboratory trial mixes were carried out over a cement content range of 360–500kg/m³, with blend levels of 30% and 50% of limestone filler (gravel), initially with only a superplasticiser. Figure 5 summarizes the performance of the different mixes. It should be stressed that the trials were investigating high-performance SCCs, with realistic slumpflows of 700mm.

The results illustrated that true SCC could not be produced with just the addition of a super- plasticiser below a cement content of 440kg/m³. Although the mixes were highly fluid they became segregated. Instability was created by the excess water needed to achieve the desired workability in combination with the insufficient fines needed to maintain the viscosity. Good strength reductions were achieved as expected.

Further trials were undertaken using a viscosity modifier at cement contents of 400 and 360kg/m³. The results showed that SCC could be achieved in the laboratory with a total cementitious content of around 370kg/m³, using limestone filler. This is also illustrated in figure 5.

SCC in the city

Within structural design there is a general trend towards slimmer elements¹², particularly in building structures where the advantages are chiefly increased useable space and reduced self-weight, which thus requires high-strength concrete.

Slimmer elements can lead to difficulty in vibration of the concrete because of congested reinforcement. This offers a great market opportunity for utilizing SCC. Figure 6a shows congested reinforcement with SCC being placed, while a core taken from the structure shows good aggregate distribution (fig. 6b).

Within a city location environmental issues are very important. The use of SCC leads to a reduction in noise levels for site neighbours owing to the elimination of vibration equipment, which in turn reduces energy consumption. Materials consumption will also be reduced due to less spillage and reduced cement consumption. Energy consumption and CO₂ emissions will also be reduced¹³.

Health and safety is an important factor on any site but even more so within a city environment with its congested ground areas. Thus, without the need to move pump hoses or handle vibrator equipment, the use of SCC will significantly improve the working environment. And, without the use of hand-held pokers, which can cause blood circulation problems, there should be a reduction in injuries.

Quality of construction work is also vitally important and, with the noticeable reduction in the number of skilled workers, SCC reduces the dependency of concrete quality on the workforce. SCC has already been used on several sites in city locations including HM Treasury and Guys Hospital in London, and the Millennium Tower and Ares Tower in Vienna.

References

1. OKAMURA, H.: ‘Self-compacting high-performance concrete’, Concrete International, 1997 vol. 19, no. 7, pp 50–54.
2. KHAYAT, K.: ‘Workability, testing and performance of self-consolidating concrete’, ACI materials journal, 1999, vol. 96, no. 3, pp 346–353.
3. OKAMURA, H. and M. OUCHI: ‘Self-compacting concrete – development, present uses and future’, proceedings of first RILEM international symposium on self-compacting concrete, Sept 1999, Stockholm.
4. OZAWA, K.; MAEKAWA, K. and H. OKAMURA: ‘High-performance concrete with high filling capacity’, proceedings of RILEM international symposium on admixtres for concrete: improvement of properties, May 1990, Barcelona.
5. PETERSSON, Ö. and P. VAN BK. BILLBERG: ‘A model for self-compacting concrete’, proceedings of RILEM international conference on production methods and workability of fresh concrete, June 1996, Paisley.
6. SEDRAN, T. and F. DE LARRARD: ‘Self-compacting concrete – a rheological approach’, proceedings of RILEM international workshop on self-compacting concrete, Aug 1998, Japan.
7. SAAK, A.; JENNINGS, H. and S. SHAH: ‘New methodology for designing self-compacting concrete’, ACI materials journal, 2001, vol. 98, no. 6.
8. SKARENDAHL, A.: ‘State-of-the-art of self-compacting concrete’, proceedings of seminar on self-compacting concrete, Malmö, Nov 2000, pp 10–14.
9. GIBBS, J. and W. ZHU: ‘Strength of hardened self-compacting concrete’, proceedings of first RILEM international symposium on self-compacting concrete, Sept 1999, Stockholm.
10. PETERSSON, Ö.: ‘Design of self-compacting concrete, properties of the fresh concrete’, proceedings of seminar on self-compacting concrete, Malmö, Nov 2000, pp 16–20.
11. Building Research Establishment, Practical Guide for Engineers using SCC.
12. MARSH, B., Ove Arup, personal communication, Feb 2002.
13. GLAVIND, M.: ‘How does self-compacting concrete contribute to implementation of sustainable/clean technologies in the construction industry?’, proceedings of seminar on self-compacting concrete, Malmö, Nov 2000, pp57–61.

The authors, Rob Gaimster and Noel Dixon, are divisional technical services manager and concrete analyst, respectively, with RMC Readymix UK Ltd. Their paper was first published in the Institute of Concrete Technology Yearbook 2002-2003 and is reprinted here by kind permission of the ICT.