LNCINEER - Vol. XXXVIII, No. 03, pp. 23-30, 2005O The Institution of Engineers, Sri Lanka

Structural Properties of Cement StabilizedRammed Earth

C. Jayasinghe and N. Kamaladasa

Abstract: Earth is gradually gaining popularity as a structural material due to the desirable propertiesthat can be obtained with various stabilization techniques. Cement stabilized rammed earth is one suchapplication which can offer many advantages such as walls that will not need any mortar or plaster. Thewider application of this material for walls of single and two storey houses will need strength anddurability characteristics. The strength parameters relevant to compressive and flexural behaviourswore found using a detailed experimental study using wall panels. The durability characteristics weretested with panels soaked in water for 24 hours. The applications of above parameters for single andIWD storey houses were highlighted.

Keywords: Rammed earth, Structural properties

I. Introduction

Tin1 prevailing scarcity of conventional buildingmaterial based on natural resources indicates theimportance of developing alternative buildingmaterials. The continuous use of burnt clay bricksor rrment sand blocks as walling materials hasli'd to the over exploitation of many naturalresources with the associated environmentalproblems. The problems associated with( • M t ' s s i v i 1 clay and sand mining can behighlighted as examples.

Due to the scarcity of conventional building1 1 1 . M I - i i , i l - . , t h e awareness a n d interest o n earth a s•1 building material has grown very much all overl l u - world [I], [2], [3], [4]. Already, cement

i . i l ' i l i / , e d soil blocks have become a viableAlternative to bricks and cement sand blocks inSi i L a n k a . Another alternative is cementMitl'ili/.ed rammed earth. This can offer manyadvantages in the light of the severe sand shortageprevailing in Sri Lanka since the walls will notturn-any mortar joints. Since it can be constructedto give a reasonably straight and smooth surface,tin- need to apply plaster can be minimised. Thus,the consideration of rammed earth as and l t i - rna t i ve walling material is extremelyworthwhile. The wider applications will need the•.I H -i i);t 11 characteristics so that proper engineeringi l i - . i g n methods can be used with variousapplications. This paper highlights the findingsol a detailed research carried out using wallpanels made with different soils and cementi ontenls tested for compressive and flexuralproperties. Further, rammed earth constructionrequir ing unskilled labor comes as a solution to

the scarcity and eventual high price of skilledmasons, making it advantageous than any typeof block construction.

2. Objectives

The main objective was to assess the followingmain design parameters.

a. Compressive strengths and associatedparameters

b. Flexural strengths

c. Wet strengths and durability aspects

3. Methodology

In order to achieve the objectives, a number ofwall panels were cast with different types of soilswhich were tested under compression andflexure. Wall panels immersed in water for 24hours were also tested under compression. Theseresults were compared with results obtained invarious other countries to determine thesuitability of laterite soils available in Sri Lankafor rammed earth construction.

Eng. (Dr.)Mrs. C. Jayasinghe, PhD, MEng, B.Sc Eng.(Hons), C. Eng.,MIE(SL), Senior Lecturer, Department of Civil Engineering, UniversityofMoratuwa.Eng N.Kamaladasa, C.Eng, MIE(SL), B.Sc Eng (Hons), Director, CHPBSri Lanka.

23 ENGINEER

4. Manufacturing process of rammedearth panels

Rammed earth walls are formed by compactingthe prepared soil inside a set of temporary forms.The soil is mixed with the chemical stabilizingagent (cement) [4], [5], [6] and a very smallamount of water if the soil is too dry. The size ofthe moulds used in the experimental programmeare 2'-0" (0.6 m), 4'-0" (1.2 m) and 6'-0" (1.8 m)feet long which can form 12'-0" (3.6 m), 10'-0"(3.0m), 8'-0" (2.4 m), 6'-0" (1.8 m), 4'-0" (1.2m) and2'-0" (0.6m) long walls by combining two or threemoulds. The soil is placed in layers of 100 mm to150 mm thick and each layer is compacted withthe manually operated hammer (Figure 1). Oncethe soil is compacted well, the f ormwork is movedupwards for the next lift of the wall.

The optimum water content is about 9.5 - 11.0%and the maximum dry density is about 20 kN/m3

for compaction levels. Mechanical stabilization bydynamic compaction seems to give better resultsas compared to static or vibro -static compaction[1]. The optimum moisture content (OMC) forrammed earth soils is critical in order to achievemaximum dry density through dynamiccompaction, which will directly influence thestrength and durability of the material [2].

The moisture content during the soil mixing stagewas controlled by a simple test called drop test. Aball is made in the palm using a small sample ofsoil and then dropped on the floor from about 1.0m height. If it breaks into 4 or 5 pieces, the moisturecontent is satisfactory. If it crumbles away, the soilis too dry or if it drops as one pat, it is too wet [7].

Not all soils are suitable for rammed earthconstruction. The Soil should be reasonably wellgraded between gravel to clay size particles. Themain soil constituent for rammed earth is sand (40-70%). Clay content from 10-20% is generallysufficient for binding as higher clay contents mayresult in excessive shrinkage [8]. The researchcarried out at the Department of Civil Engineering,University of Moratuwa in 1999 by Jayasinghe &Perera [9] has shown that a clay content less than30% is most suited for compressed earth blockswhich are stabilized with cement. Therefore, inthis study, three different laterite soil types wereselected which can be identified as sandy, gravellyand clayey soils.

Unlike the other building materials, generalsuitability of soil composition for construction isnot readily standardized because of its inherentnatural variability. However, it is considered asan ideal replacement to reduce the environmentalimpacts, green house gas emissions andembodied energy [10]. Rammed earth technologyis also suitable for soils with a high percentage oflarge grain size particles [1].

The three types of soils used for the experimentalprogramme have material properties given inTable 1.

The results given in Table 1 were obtained fromthe sieve analysis. The silt and clay contents wereconsidered together as this cannot be separatedat the site conditions. At site, these proportionscan be determined by conducting a jar test. Asthis technology will be popularized in the ruralareas, where there can be difficulties in testingsamples at laboratories, the jar test is an idealmeasure to determine the soil type. In this simpletest, a soil sample is placed in a bottle to about I/3 the volume, the rest is filled up with water,shaken it well and left for about 24 hours. Thecomposition of the soil sample can be seen aslayers to approximately identify the constituentmaterial [7].

Another important parameter that is of significantimportance to strength and durability is cement.However, the cost of construction is very muchaffected by the cement content. Therefore, theoptimum cement content should be determinedto meet the strength requirements of single andtwo storey load bearing constructions.

Cement stabilization of soil increases the elasticmodulus of the wall. It changed from 1.89 GPafor unstabilised soil to 2.51 GPa for 10% cementstabilized soil [1]. Since the wall has no joints, itcan be adversely affected by shrinkage. Thecement can control the shrinkage once the wallgains sufficient strength. It was found thatshrinkage of cement stabilized soil increasesrapidly during the first four days and at a laterage the increase is very slow. Hence, curingduring the first four days is very important inreducing drying shrinkage and cracking. Sandparticles reduce the shrinkage as it opposes theshrinkage movement [1].

ENGINEER 24

5. Wall panel testing for compressivestrength

Wall panel dimensions were determined in sucha way that the slenderness effects will not bepredominant. The panels were constructed withllm-e soil types mentioned in Section 4. A panelbring tested is shown in Figure 2. In addition tot in 1 compressive strength characteristics, the loadill-formation curve was also obtained. The loadversus deflection values were used to find out theload deformation behaviour of rammed earthpanels.

Two panels were constructed and tested for eachparameter as stated in BS 5628: Part 1:1992. Theul t imate load given in Table 2 is the average ofthe two values obtained for the identical panels.Tho summary of the results are given in Table 2.Thr average strength of wall panels is given inTable 3.

Thr characteristic compressive strength, fk of anymasonry is determined by tests on wall specimensusing the following equation (BS 5628: Part 1:I«W2)

P. F /A1 • IB *

/1.2u ' .(1)

where,I is Ilir mean of the maximum loads carried byllu- two test panels.

A is the cross sectional area of each panel

M')n is the reduction factor for strength of mortar

4'(| is the unit reduction factor for samplestructural strength

Tin1 f ac to r 1.2 is introduced to relate thet liaraclcristic value to the mean value. Both T

m

and M'u are taken as 1.0 since no mortar is used.

According to the results obtained for the threedifferent soil types, sandy soil gave better resultslor the characteristic strength. This confirms thelimlings made at the University of Bath, UnitedKingdom [8].

Thi'si1 strengths indicate that a characteristicI 1 impressive strength of 1.5 N/mm2 could be usedlor rammed earth walls. This is more thanhull icient for single storey construction and hencetvonomy can be achieved by using a walli I n , kncss of about 140 -160mm. For two storey

load bearing wall houses, wall thickness of 240mm may be a possibility as recommended byPerera and Jayasinghe (2003) [9], with cementstabilized soil blocks. However, it is advisable tohave more testing on thicker walls prior to actualadoption of this for two storey houses.

The weight of a panel varies from 190 kg to 220kg. The density of rammed earth is in the rangeof 1800 to 2000 kg/m3.

Wet strength of wall panels was determined aftersoaking the panels in water for 24 hours. Wetstrength results are given in Table 4. For this, thecomplete panel was immersed in a water bath for24 hours. Due to the difficulties of doing animmersed test, only the panels with gravely andclayey soils were used. These would be moresusceptible to the strength decrease due to water.

Heathcote (1995) [3] has stated that the ratio ofwet to dry strength be used as an indicator of thedurability of earth wall components. The ratio ofwet to dry strength of 0.33 - 0.5 may be regardedas suitable depending on the severity of rainfall.

Table 4 shows that panels made with evengravelly or clayey soils can give a ratio more than0.33. Therefore, the strength of rammed earthwalls under adverse conditions will be adequate.

According to the experimental results obtainedfor load deformation characteristics, rammedearth walls behave in a similar manner to theother masonry walls.

However, all three soil types gave loaddeformation characteristics with little warningbefore failure. As most of the panels did not showsignificant cracking before the ultimate crushingfailure, a higher factor of safety is recommendedin structural design of rammed earth walls. Theavailability of a compressive strength in excessof 1.5 N/mm2 can give a very high factor of safetyfor single sotrey construction. New ZealandStandards [12] considers design strength of 0.5N/mm2 as adequate for single storeyconstruction. This includes a factor of safety ofabout 5. Thus, rammed earth walls withcommonly available laterite soils of low claycontent can give a factor of safety close to 15 andhence can be used with confidence for singlestorey construction.

25 ENGINEER

6. Flexural properties of rammedearth

Flexural properties are of less significance thancompressive strengths for load bearing masonryconstruction. New Zealand Standard [12] takesthe characteristic bending strength equal to 10%of the characteristic compressive strength or about0.1 N/mm2 when no other data is available.Australian earth building hand bookrecommends ignoring any material strength inbending in the absence of test data [6].

In order to determine the flexural strength oframmed earth, a special test method was adopted.A wall panel was constructed over a concretelintel of 75 mm thick and 160 mm wide. Arammed earth panel was constructed with a725mm height. On this, again a concrete lintelwas placed. The length of the panel was 1900mm and it was loaded to failure by applying loadsas shown in Figure 3. A panel used for flexuraltesting is shown in Figure 4. The failure occurredwith the formation of diagonal cracks whichinitiated at the supports. It should be noted thatthe concrete beams will be of very little influenceto the failure load since the section predominatelyconsists of cement stabilized rammed earth. Sincethe shear failure occurs due to direct tensilestresses induced in the wall, the failure stress canbe used as an indication of tensile strength of thematerial. For the two panels tested, the shearstresses at first crack were 0.239 N/mm2and 0.31N/mm2 respectively. These give an average valueof 0.275 N/mm2. Therefore a value of 0.1 N/mm2

recommended in New Zealand Standard [12] canbe used for walls made with cement stabilizedrammed laterite soils. This can be used to checkthe lateral stability of the walls such as freestanding or simply supported at the top.

Vertical bending moment capacity of the wall isgiven by [6]:

cy ~ ( * t d ) X ̂ (2)

Where, f( characteristic design bending strengthof rammed earth which can be taken as 0.1 N/mm2.

f d design compressive stress at the crosssection

Z - Section modulus of the cross sectionunder consideration

7. The uses of the strength properties

The strength properties obtained from thisdetailed study can be used for a number ofapplications in the structural design of single andtwo sotrey houses. Some of them are highlightedbelow.

1. Design of load bearing walls subjected tovertical loads in single and two storeyhouses. This can be carried out using thedesign guidelines given in BS 5628: Part1:1992, [11] and with a characteristicstrength of 1.5 N/mm2.

2. The flexural properties can be used topredict the lateral load carrying capacityof free standing walls as presented in Cl.36.4.3 of BS 5628: Part 1:1992 [11].

3. The compressive strength can be used todetermine the adequacy of internal returnwalls to carry both in-plane loads causedby lateral loads on outer walls and verticalloads that occur in two storey houses.

4. The shear capacity under flexural loads canbe used to determine the reinforcementrequirements for foundation strengtheningneeded in weak soil conditions [13].

It can be seen that the design parameters obtainedon experimental basis will provide valuabledesign data that can be confidently used by designengineers for structural design purposes. This willassist in widening the applications of cementstabilized rammed earth and also making themsafer.

8. Conclusions

The development of alternative buildingmaterials is of considerable importance in thecontext of various environmental problemsassociated with traditional building materials.Cement stabilized rammed earth is one suchalternative. Whenever a new building material isintroduced, determination of design parameterscan enhance the range of applications.

As a result of the detailed experimentalprogramme presented in this study, it is nowpossible to use rammed earth as a strong anddurable material for wall construction. Thedesign strength indicates that characteristic wallstrengths in excess of 1.5 N/mm2 can be obtainedwith cement stabilized rammed earth using

ENGINEER 26

laterite soils. The sandy laterite soils have shownthe best behaviour. Thus, it is ideal for singlestroey construction. It can become a potentialbuilding material even for the load bearingground floor walls of two storey houses

Table 1: Particle size distribution of soil types

Soil type

Sandy

Gravelly

Clayey

Particle size>19mm

4.3

17.9

6.4

Gravel°/ocontent

32.2

56

50.5

Sand %content

59.4

29.6

30.4

Fines% content(clay and silt)

8.4

14.4

19.1

Table 2: Results for the compression testing of panels

Soil type

Sandy

Gravelly

Clayey

Cement %

6%

8%

10%

6%

8%

10%

6%

8%

10%

Load atfirst crack(tonnes)

*

*

45.0*

*

44.0*

*

30.0

Dimensions(mmxmmxmm)

1045 x!60 x633

1050 x 155x640

1050 x!60 x645

1040x160 x 650

1020 x!60 x630

1035 xl50 x640

1040 x!63 x640

1030 x!65 x630

1050 xl60 x698

Average load(tonnes)

41.65

58.20

61.25

33.70

31.75

67.45

31.10

34.50

37.60*- no cracks formed before failure

Table 3: Average strength of wall panels

Soil

Sandy

Gravelly

Clayey

Cement

6%

8%

10%

6%

8%

10%

6%

8%

10%

Average strength(N/mm2)

2.47

3.525

3.71

2.03

1.97

4.34

1.82

2.06

2.30

fk (N/mm2)

2.06

2.94

3.09

1.69

1.64

3.62

1.52

1.72

1.92

27 ENGINEER

Table 4: Wet strength results

Soil type

Gravelly

Clay

Cement %

6%

6%

Load atfirst crack

12

Ult.load(tonnes)

22.5

14

Dimensions(mmxmmxmm)

1080 x!60x630

1030 x!60 x640

Weight

(kg)

218

216

Strength(N/mm2)

1.30

0.85

Table 5: Ratio of wet/ dry strength of rammed earth panels

Soil type

Gravelly

Clay

Cement %

6%

6%

Dry strength

2.03

1.82

Wet strength

1.30

0.85

Ratio

0.64

0.46

Figure 1: Moulds and hammer used to construct the rammed earth walls

Figure 2: Panel being tested in the compression testing machine

ENGINEER 28

h2)

loading concrete beam

diagonal crackdue to shear

Masonry wall

supports

concrete beam

Figure 3: Testing of reinforced masonry beams in two point loading

Figure 4: Wall panel used for flexural testing

which should be further investigated withadditional laboratory experiments on thestructural behaviour. The characteristic flexuralstrength obtained is more than 0.1 N/mm2.Although it appears as a low value, this can givea significant lateral load carrying capacity for athick wall. Since the weight of the structure willact as an advantage, it is possible to obtainsuitable housing layouts that will have adequatevertical and lateral load resistances. Thus,rammed earth could be considered as a materialof comparable strength to good quality bricksavailable in Sri Lanka.

Acknowledgements

The authors wish to thank the engineering staffof Center for Housing Planning and Building(CHPB) Mr. Chandradasa, Ms. Kandambi and Mr.Pushpakumara for their fullest cooperation in theexperimental programme. The final yearundergraduates, Messers ALP Silva, RSMallawarachchi and KK Madurawala supportedthis experimental programme with muchenthusiasm. The support given by the laboratorystaff, Messers SP Madanayake, SL Kapuruge, HPNandaweera of The Department of CivilEngineering, University of Moratuwa for thetesting programme is gratefully acknowledged.

29 ENGINEER

References

1. Bahar R., Benazzoung M., Kenai S., (2004),"Performance of compacted cement stabilized soil",Cement & Concrete composites, 26, pp 811-820.

2. Hall M., Djerbib Y., (2004), "Rammed earth sampleproduction: context, recommendations andconsistency", Construction and Building Materials, 18,pp 281-286.

3. Heathcote K.A., (1995), "Durability of earth wall

buildings", Construction and Building Materials, Vol:9, No. 3, pp 185-189.

4. Walker P., Maniatidis V.,(2003), "A review oframmed earth construction", May 2003, NaturalBuilding Technology group, Department ofArchitecture and Civil Engineering, University ofBath, United Kingdom, 109 p.

5. New Mexico Earthen building material code, (2003).

6. Standards Australia 2002, The Australian Buildinghandbook, Standards Australia, Sydney, Australia.

7. Houben H. and Guillaud H., (1994), "Earth

Construction", A comprehensive guide, CRTerre -BAG, Intermediate Technology Publications.

8. www.bath.ac.uk. Developing rammed earthwalling for UK housing construction.

9. Perera AADAJ, Jayasinghe C, (2003), "Strengthcharacteristics and structural design methods forcompressed earth block walls", MasonryInternational, Vol 16, No.l, January 2003, pp 34-38.

10. Morel JC, Meshah A, Oggero M, Walker P (2001),"Building houses with local material: means to

drastically reduce the environmental impact ofconstruction", Building and Environment, 36 1119-1126

11. BS 5628: Part 1: 1992, Code of Practice for Use ofMasonry, British Standards Institute, UnitedKingdom

12. NZS 4297: 1998, New Zealand Standards,

Engineering design of earth buildings, StandardNew Zealand, Wellington, New Zealand

13. Jayasinghe MTR, (1999), "Foundation improvementtechniques for brick wall structures" Engineer,Journal of Institute of Engineers, Sri Lanka, Vol:xxx,No3, pp 41-50.

14. Reddy B.V., Jagadish J.S. (1989), "Properties of soil-cement block masonry", Masonry International,Vol: 3, No.2, October pp 80 -84.

ENGINEER 30