Construction and Building Materials 74 (2015) 259–267

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier .com/locate /conbui ldmat

Characterization of a stabilized earth concrete and the effect ofincorporation of aggregates of cork on its thermo-mechanical properties:Experimental study and modeling

http://dx.doi.org/10.1016/j.conbuildmat.2014.09.1060950-0618/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: Civil Engineering Department, University of Djelfa,17000 Djelfa, Algeria.

E-mail addresses: bachargc@gmail.com (M. Bachar), l.azzouz@mail.lagh-univ.dz(L. Azzouz), rahmoh_m@yahoo.fr (M. Rabehi), mezghichebm@yahoo.fr(B. Mezghiche).

Mohamed Bachar a, Lakhdar Azzouz b, Mohamed Rabehi a,c,⇑, Bouzidi Mezghiche c

a Civil Engineering Department, University of Djelfa, 17000 Djelfa, Algeriab Civil Engineering Department, University of Laghouat, 03000 Laghouat, Algeriac Research Laboratory Civil Engineering, University of Biskra, 07000 Biskra, Algeria

h i g h l i g h t s

� Effect of aggregates of cork on thethermo-mechanical properties of CEB.� Mechanical strength decreases but

acceptable with increasing content ofcork.� Some analytical models are used for

comparison of experimental results.� The energy consumption for thermal

comfort can be reduced.

g r a p h i c a l a b s t r a c t

a r t i c l e i n f o

Article history:Received 16 May 2014Received in revised form 26 August 2014Accepted 23 September 2014Available online 12 November 2014

Keywords:Compression strengthTensile strengthMaximum dry densityThermal conductivitySoil stabilizationAggregates of corkThermal insulation

a b s t r a c t

In this paper, mechanical and thermal properties of compressed earth blocks stabilized (soil–sand dune–cement) with and without aggregates of cork have been studied, with the use of some models that predictthermal conductivity for comparison of experimental results. The first part highlights the influence of thepercentage by weight of cement and of sand dune on the maximum dry density, optimum moisture con-tent and mechanical resistance. The results showed that mass content of 30% sand dune and 12% cementsignificantly improves these properties more in the wet conditions than dry, and therefore gives the opti-mal mixture (58% of soil–30% of sand dune –12% of cement). However, the composite materials used forbuilding must present sufficient mechanical strength to be suitable for constructions. For that the optimalcomposition has undergone a static compaction (2.5–5–7.5–10 MPa) thus showing further improvementon the same properties. Incorporating the aggregates of cork (3/8) in the optimal mixture, improves sig-nificantly the thermal performance with little influence of compaction, while remaining within the rangeof acceptable strength. Lastly the study was extended to the search for theory capable of predicting theeffective thermal conductivity of a dry blend considered like a two-phase system, which shows that lin-ear model and some theoretical models give the best concordance with experimental results.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Environmental concerns can discover the qualities of earthmaterial. Its use is not involved in the depletion of resources andincreased pollution (water, air, soil) and waste, biological changes.

Table 1Characteristics of the soil used.

PropertyAtterberg’s limits Liquid limit WL 70.4

Plasticity index Ip 43.2

Grain size distribution Gravel (>4.75 mm) (%) 2.3Sand (0.074–4.75 mm) (%) 6.4Clay and silt (<0.074 mm) (%) 91.3

Chemical characteristics Iron oxide-alumina (%) (Fe2O3–Al2O3) 14.6Carbonate CaCO3 (%) 33.0Chloride NaCl (%) 0.15Sulfates CaSO4 (%) 0.20Insoluble residue IR (%) 42.2

Normalized Proctor test Optimum water content (%) 9.0Maximum dry density (kg/m3) 1680

Sand equivalent By piston test (%) 14.8By sight (%) 17.6

260 M. Bachar et al. / Construction and Building Materials 74 (2015) 259–267

This material has become more and more economically competi-tive because of raw material availability and simplicity in the pro-duction process. Soil stabilization with hydraulic binders wasstarted in 1917 and that many researchers focused their researchin this direction [1–13]. The hardening of the soil is generallyaffected by the cement hydrates in the presence of water to formcomplex carbohydrates. The cement induces the consolidation(creation of a skeleton) that coats the grains and opposes themovement of the material. The main reactions taking place in theconsolidation of the cement itself between the stabilizer and thesand fraction of the earth. However, side reactions are observedbetween the stabilizer and the clay fraction of soil. The clay actson the effectiveness of the stabilization process and modifies themechanical behavior of the earth. Several studies have investigatedthe behavior of stabilized earth blocks facing a number of physicalconstraints. And the effect of the type of cement that presents alevel of reactive silica, plays a vital role in reducing the porosityand consuming a portlandite generated by the selected cement,This is confirmed by [14–16], different hydrates responsible forthe hardening of the cement that bind the clay particles to forma solid block [1,17,19], and also the mechanical strength increaseswith the gradual addition of cement [7,8]. A study by Guettala et al.[20] shows the effect of the mass of sand and a content of 30%affects positively the mechanical strength for 30% of sand in drymedium and a 36% in wet. Venkatarama Reddy et al. [21] also showthat the resistance changes significantly with the dry density, andthat it operates independently of the content of water and cement;But many authors restrict the dry density in the range of 1500–2000 kg/m3 and determined according to the standard procedureas ASTM C 140 and BS 1924-2 (1990) standard and others[1,15,22,23]. The compressive strength and the dry density, changewith the increase of the static compacting [1,17,18]. One of themajor concerns of a designer is to provide a well-insulated buildingproviding comfort at the least cost to the user both in winter andsummer. Bahar et al. [1] proclaimed in a study that the additionof cement and sand may slightly decrease the conductivity of thebrick; however; moisture increases the thermal conductivity ofthe sample relative to its dry state [1,17]. In the same way Meukamet al. [24,25] show that the thermal conductivity significantlyincreases with the content of the water in mixture. Indeed,increased the wetting of materials, results a replacement the poresprogressive of air by the water. Several studies have been made onthe integration aggregates of cork in cement and plaster to obtain abest thermally insulating [26–30]. Based on the experiencesreported by Castro et al. [31,32], the concrete blocks containingcork exhibit a reduction in the thermal conductivity rate of 45% rel-ative to the blocks without cork. Panesar et al. [33,34], also showthat the thermal conductivity is affected by the bulk density ofthe composite of cork that provides additional volume in void.The apparent thermal conductivity of porous material dependson many parameters; the thermal conductivities of the solid andfluid, the degree of porosity, the size, shape and distribution ofthe pores. The thermal study of a sample of porous materialsrequires good knowledge of the microstructure on the one handand on the other hand has the analytical or numerical tools for pro-cessing data required. There are analytical models considered pre-dictions tools developed to understand the heat transfermechanisms in the multiphasiques middles. These should allowthe incorporation of the microstructure (particle shape, the contactareas) and the microstructure (boundary conditions, porosity).Many analytical models are used to calculate the apparent thermalconductivity of materials biphasic according to the solid middleand fluid of the two present phases and also of their porosity andtheir compactness. The actual thermal conductivity of the mediumand whatever the model used, is always between two extreme val-ues, the lower bound (series model) is an environment where the

vector density of heat flow is perpendicular to the layers and theupper bound (parallel model) corresponds to an arrangement oflayers parallel to the direction of the density of heat flow. [35].Bensenouci et al. [36] adopt a thermal model of a concrete pozzo-lan on two approaches by comparing the experimental results withthose obtained by theoretical calculation. Some prediction modelsare used by many authors and generally are Hashin–Shtrikman[37] and Krischer and Kroll [38]. Willy and Soutwik [39] and Max-well [40]; by homogenization self-consistent [41] and others. Thispaper therefore aimed to study the influence of the addition ofaggregates of cork on the mechanical and physical properties afteroptimization of the mixture (soil–sand–cement) and their effectson the thermal behavior; a comparison of experimental resultswith some prediction models will be well studied by observingthe correlation between the results.

2. Experimental program

2.1. Materials

The soil used in this investigation was sourced from the Djelfa region which wasfirst passed through a 5 mm sieve before being characterized. Table 1 summarizesthe characteristics of the soil used. Fig. 1 gives the grading curve of the soil used.Composite cement (CEM II/B) class 42.5 MPa with 35% to limestone fillers was usedfor the chemical stabilization of soil. The clinker is from the cement factory ofM’sila. The chemical analysis of clinker shows that it is in conformity with standardNFP 15-301. The chemical and mineralogical compositions of clinker are presentedin Tables 2 and 3, respectively. The sand used was fine sand dune passing a 0.63 mmsieve from the Djelfa region. Table 4 also shows some physical characteristics of thestudied sand dune. The grading curve of sand dune is given in Fig. 1. Cork granulesfrom waste in sawing plate’s compressed cork at the factory located in Jijel region,these wastes are then separated according to the different sizes using sieves. Table 5shows some characteristics of the studied cork granules. The grading curve of thecork granules 3/8 mm is given in Fig. 1. The water is drinking water that containslittle sulfate and having a temperature of 20 ± 2 �C. Its quality conforms to therequirements of NFP 18-404 standard.

2.2. Testing method, proportions of mixtures and the specimens preparation

An experimental program was carried out to studying in parallel the effect ofdifferent percentages of sand dune 0%, 10%, 20%, 30%, 40%, 50%, 60% and 70% bymass relative to soil and the influence of the chemical stabilization by cementaddition 0%, 2%, 4%, 6%, 8%, 10% and 12% by mass relative to soil on the phy-sic-mechanical properties of the various mixtures. We have prepared fifty-six for-mulations without compaction and without the incorporation of the aggregates ofcork. The details of the mixtures proportions are given in Table 6. After this wehave determined the optimal composition (soil–sand dune–cement). Subse-quently, we studied the influence of the mechanical stabilization by static com-paction for four levels of applied stresses 2.5, 5, 7.5 and 10MPa on themechanical properties, maximum dry density and thermal conductivity the mix-ture already optimized in the first part, well as the effect of the curing meth-ods(dry-humid) on the mechanical properties and thermal conductivity ofoptimized mixture. Finally, we evaluated the effect of the incorporation of themass contents of the cork granules 2%, 4%, 6%, 8%, 10% and 12% relative to soil

1E-4 1E-3 0,01 0,1 1 10 1000

20

40

60

80

100

Sieve size (mm)

SoilDune sand (0/0.63)Cork aggregates (3/8)

Perc

enta

ge p

assi

ng (

%)

Fig. 1. Grading curves of soil, the sand dune and the cork granules.

Table 2Chemical composition of clinker (%).

SiO2 Al2O3 Fe2O3 CaO MgO SO3 Na2O K2O Cl LOI

17.49 4.51 3.02 62.78 2.15 2.38 0.05 0.64 0.02 8.10

LOI: Loss on ignition.

Table 3Mineralogical composition of clinker (%).

C3S C2S C3A C4AF

55.41 13.65 2.25 14.83

M. Bachar et al. / Construction and Building Materials 74 (2015) 259–267 261

on the mechanical properties, maximum dry density and the thermal conductiv-ity, as well the effect of the curing methods (dry-humid) on the mechanical prop-erties and thermal conductivity of optimized mixture. In this study, the

Table 4Physical characteristics of sand dune (0/0.63 mm).

Finenessmodulus

Apparent density(kg/m3)

Specific density(kg/m3)

1.13 1491.13 2517

Table 5Characteristics of aggregates of cork (3/8 mm).

Thermal conductivity (W/m K) Apparent density (kg/m3)

0.036 71

Table 6Mixtures proportions.

Mixture (soil–sand dune–cement) Water (%)

p = 0% cement 2%

(100-p)% S + 0% DS + p% C 16.20 16.10(90-p)% S + 10% DS + p% C 15.80 16.40(80-p)% S + 20% DS + p% C 15.40 16.02(70-p)% S + 30% DS + p% C 14.60 15.08(60-p)% S + 40% DS + p% C 15.20 15.60(50-p)% S + 50% DS + p% C 16.05 15.88(40-p)% S + 60% DS + p% C 16.58 16.65(30-p)% S + 70% DS + p% C 16.80 16.88

specimens were demolded 1 min after the completion of the compaction andwere stored in two environments (in the free air and wet wrapped in thin plasticfilm). We have thought to the treatment of aggregates of cork before addition inthe earth concrete stabilized and compressed. The treatment is a superficial coverto the cement selected from the literature [42]. The aggregates of cork are coatedwith a cement milk (cement + water) and allowed to dry completely before theiruse so that the aggregates of cork do not absorb water during incorporation inthe cement-stabilized compressed earth blocks.

2.3. Tests conducted

2.3.1. Proctor compaction testThe Proctor test is used to determine the optimal water content that leads to the

maximum density. This test uses a soil sample with less than 4.76 mm (sieve no. 4)to which an increasing water content is added. The soil sample is compressed intothree layers with 25 blows per layer. Proctor standard compaction test, according toBS 1377-1990: Part 4 [43] was applied to determine the maximum dry density(MDD) and the optimal water content (OWC) of soil.

2.3.2. Static compaction testStatic compaction is obtained by applying a static pressure. The oedometric

mold enables the production of the cylindrical specimens under an axial stresswhich can vary from 2.5 to 10 MPa. It consists of two half-shells fixed to an outershaft and two lower pistons (fixed) and upper (mobile). This mold 100 � 100 mmis movable along its axis independently of the two pistons to ensure compactionsymmetrical relative to the median plane and horizontal spread between the fric-tion material and the walls along the entire height of the sample.

2.3.3. Compressive strength testThe compressive strengths were measured by testing on the cylindrical speci-

mens of 100 mm in diameter and 100 mm in height in accordance with the ASTMD 1633-00 [44]. The compressive strength was determined at the ages of 7, 14,28 and 90 days.

2.3.4. Tensile strength testThe tensile strength by splitting (Brazilian test) was measured by testing on the

cylindrical specimens with 100 mm in diameter and 100 mm in height in accor-dance with the ASTM 496-96 [45]. The tensile strength by splitting was determinedat the ages of 7, 14, 28 and 90 days.

2.3.5. Determination of porosityTo measure the porosity as a percentage of different compositions, the Eq. (1)

has been used that expresses the difference in weight between the oven-dry andthe water-saturated conditions.

Porosity(%)

Sand equivalent

By sight (%) By piston test (%)

41 97.58 96.58

Specific density (kg/m3) Porosity (%) Absorption (%)

145 51 2.3

4% 6% 8% 10% 12%

17.40 17.60 18.20 18.60 18.7016.50 16.90 17.50 17.80 17.9016.05 16.08 16.50 16.90 17.0515.10 15.80 16.14 16.23 16.8015.90 16.02 16.25 16.32 16.4216.20 16.26 16.36 16.62 16.7416.74 16.88 17.12 17.62 17.8717.05 17.25 17.32 17.48 17.54

Heat flux

Fig. 3. Parallel model [35].

Heat flux

Fig. 4. Series model [35].

262 M. Bachar et al. / Construction and Building Materials 74 (2015) 259–267

p ¼ wssd �wd

wssd �ww� 100 ð1Þ

where p is the porosity (100%), wssd is the specimen weight in the saturated surfacedry (SSD) condition (g), wd is the specimen dry weight after 24 h in oven (g), and ww

is the weight of saturated specimen (g). This method has been used to measure theporosity of the cement-based materials successfully [46–49].

2.3.6. Principle of measuring the thermal conductivity by the hot wire methodThe hot wire method is a standard transient dynamic technique based on the

measurement of the temperature rise in a defined distance from a linear heat source(hot wire) embedded in the test material; the device used for measuring the ther-mal conductivity is illustrated in Fig. 2. If the heat source is assumed to have a con-stant and uniform output along the length of test sample, the thermal conductivitycan be derived directly from the resulting change in the temperature over a knowntime interval [50]. The hot wire probe method utilizes the principle of the transienthot wire method. Here the heating wire as well as the temperature sensor (thermo-couple) is encapsulated in a probe that electrically insulates the hot wire and thetemperature sensor from the test material [51]. The ideal mathematical model ofthe method is based on the assumption that the hot wire is an ideal, infinite thinand long line heat source, which is in an infinite surrounding from homogeneousand isotropic material with constant initial temperature. If q is the constant quan-tity of heat production per unit time and per unit length of the heating wire(W m�1), initiated at the time t = 0, the radial heat flow around the wire occurs.Then the temperature rise DT(r,t) at radial position r from the heat source conformsto the simplified formula:

DTðr; tÞ ¼ q4pk

In4atr2C

ð2Þ

where k is the thermal conductivity (W m�1 K�1), a thermal diffusivity (m2 s�1)(a ¼ k

qCp, with r is the density (kg m�3) and Cp the heat capacity (J kg�1 K�1) of the test

material and C = Exp (c), c = 0.5772157 is the Euler’s constant. Eq. (2) is valid onlywhen r2

4at� 1 is fulfilled, i.e. for a sufficiently long time t larger than certain minimumtime tmin and for a small distance r. Thus the measurement of temperature riseDT(r,t) as a function of time is employed to determine the thermal conductivity k,by calculating the slope A of the linear portion of temperature relative to natural log-arithm of the time (lnt) evolution from the following:

k ¼ q4pA

ð3Þ

2.4. Some theoretical models for predicting the thermal conductivity

From the resolution of the energy equation, many calculation models of thermalconductivity have been developed. The interest of this section is to expose somemodels in the literature that can give good results for the calculation of the appar-ent thermal conductivity of heterogeneous media. It seems necessary to clarify thatthe models we propose, are the two-phase media. The solid phase consists of a sin-gle phase represented by the solid particles. Moreover, the fluid is representedexclusively by pores filled with air.

2.4.1. Series and parallel model [35]The first-order model is a model which assumes that the temperature gradient

and heat flow are isotropic and homogeneous. In 1912, Wiener [35] proposed equa-tions giving a value of the effective thermal conductivity of the two media and thiswas done by applying the electric analogy to a series or parallel circuit in the case ofa problem of heat transfer in a heterogeneous material. The two phases areassumed to arrange parallel to one another Figs. 3 and 4. The effective thermal con-ductivity is then bounded by the two values Wiener upper bound and Wiener lowerbound respectively.

kseries ¼1

ekfþ a

ks

ð4Þ

Fig. 2. Device for measuring t

kparallel ¼ a � ks þ e � kf ð5Þ

where kseries; kparallel is the effective thermal conductivity in parallel and in seriesrespectively; kf is the thermal conductivity of the fluid phase (W m�1 K�1); ks isthe thermal conductivity of the solid phase (W m�1 K�1); e is the porosity of themedium (volume fraction of the fluid phase); and a is the compactness of the middle(volume fraction of the solid phase).

2.4.2. Hashin and Shtrikman [37]Hashin and Shtrikman [37] give a model that frames for the values of thermal

conductivity of multiphase materials. In general, the lower limit of the conductivitycorresponds to the case where the inclusion, has higher thermal properties thanmatrix (conversely, for the upper bound). These terminals are valid regardless ofthe morphology of the inclusion phase of the material. Where kHSþ; kHS� the effectivethermal conductivity in upper and lower respectively.

kHSþ

kf¼ 1þ a

1kskf�1þ e

3

ð6Þ

kHS� ¼ ks

kfþ �

11�ks

kf

þ a3�ks

kf

ð7Þ

kseries < kHS� < k < kHSþ < kparallel ð8Þ

2.4.3. Model of Krischer and Kroll [38]Model of Krischer and Kroll [38] is illustrated in Fig. 5: In this model, the mate-

rial consists of three layers: two ensembles of thermal conductivity plates one inseries and one in parallel.

kapp ¼kserieskparallel

nkrkseries þ ð1� nkrÞkparallelð9Þ

he thermal conductivity.

Heat flux

1- )

(1- )

Fig. 5. Schematic representation of a porous medium according to Krischer [38].

1(1)

2(2)

3(3)

Fig. 7. Modeling of by auto-coherent homogenization of three phases [41].

M. Bachar et al. / Construction and Building Materials 74 (2015) 259–267 263

2.4.4. Model of Willy and Souwik [39] is illustrated in Fig. 6The model will be formed of two sets of plates in parallel and in series with

respect to the direction of heat flow. The first set of thermal conductivity, consistsof plates in series, the second in parallel. Where kapp is the apparent thermal con-ductivity and the share of series arrangement is denoted as n_w and that of the par-allel arrangement (1 – n_w).

kapp ¼nwks

e kskfþ aþ ð1� nwÞðaks þ ekf Þ ð10Þ

2.4.5. Model of Maxwell [40]Maxwell’s approach was initially associated with a problem of electrical con-

duction in a heterogeneous medium consisting of spheres that have conductivitylambda dispersed in solid medium continuous for a conductivity d; which resultedin the equivalent conductivity of a homogeneous medium, and whose spheres donot interact thermally:

kapp ¼ekf ð2kf þ ksÞ þ 3akskf

eð2kf þ ksÞ þ 3akfð11Þ

2.4.6. Modeling by auto-coherent homogenization [41]Modeling by auto-coherent homogenization [41] allows estimating thermal

conductivity of heterogeneous materials on the basis of knowing the conductivityof each component and its concentration. This approach was developed for themechanical characterization (elasticity and elastoplasticity) of heterogeneousmaterials and was used by Arnaud et al. [41] on hemp in bulk, hemp concreteand hemp wools. The auto-coherent model is applied to the composite with 3 com-ponents, comprising a spherical cavity of air ‘‘1,’’ with vacuum contained in corkand matrix), surrounded by a concentric cork shell ‘‘2,’’ and moreover surroundedby an additional shell of matrix ‘‘3’’ (optimal mixture: soil (58%)–sand dune(30%)–cement (12%)). where R1, R2, R3 are the rays of the spheres having conductiv-ities k1, k2, k3 respectively and b is the volume concentration of phase 1 (see Fig. 7).

kapp ¼ k3 1þ b

1�b3 þ

1þ

k1k2�1

� �d

3

k1k3�1�

dk1k2�1

� �2k2k3�1

� �3

266666666664

377777777775

ð12Þ

Heat flux

(1- ) (1- )

Fig. 6. Schematic representation of a porous medium according to Willy andSoutwich [39].

b ¼ V2

V3¼ R2

R3

� �3

ð13Þ

d ¼ 1� R1

R2

� �3

ð14Þ

3. Test results and discussion

3.1. Mixture optimization (soil–sand dune–cement)

Fig. 8 shows the different variations in the maximum dry den-sity and optimal water content as a function of cement contentof 0–12% for each mass percentage of sand dune ranging from 0%to 70%, and we find that for a content of 0% sand dune the maxi-mum dry density decreases from 1730 kg/m3 for 0% of cement to1670 kg/m3 for 12% of cement corresponding to a decrease of theorder of 3.59%. This effect is explained by the cement demand towater for hydration affecting the decline in maximum dry densityand increased the water content of a rate of the order of 15.43%.However an optimal percentage of 30% of sand dune improvesthe maximum dry density for 0% of cement a rate of the order of9.82% and approximately 4.79% with 12% of cement, by the optimalwater content decreased a rate in the order of 10.96% with 0% ofcement and approximately 11.31% with 12% of cement. Beyond30% to 70% of sand dune, it is also observed for 12% of cement alessening of the maximum dry density and the compressivestrength coming respectively at a rate in the order of 15.13% and114.18% and an increase of the optimal water content in the orderof 4.40%. The results illustrated in Fig. 9 show that whatever thecement content an increase in mechanical strength at 28 dayswas observed from 0% of sand dune up to 30%. Therefore the sanddune plays a beneficial role up to the contents of about 30%. Bycons, beyond we observed a decrease in mechanical strength evenwith the increase in the cement content, more we are adding ofsand dune to the soil more we lose in quantity of fines (binder)and therefore cohesion. This effect is explained by the loss of cohe-sion and entanglement of the granular skeleton. The compressivestrength for 12% of cement and 30% of sand dune increases a ratein the order of 112.59% compared with the soil without sand dune,and this is due to the increase of the compactness and simulta-neously with the hydration of the cement thereby forming C–S–H responsible for hardening of the mixture. So we take as optimalmixture (58% of soil–30% of sand dune–12% of cement). Also themechanical strength curves have peaks to the right of the masscontent of sand dune which is in the order of 30%; this confirmsthat the mechanical strength is closely related to the dry density,which is in accord with the study conducted by [20].

3.2. Influence of static compaction and the curing method on theproperties of the optimal mixture

In this part, the impact of the increase in pressure static com-paction from 2.5 to 10 MPa and the curing method (dry-humid)on the properties of optimal mixture has been studied. The resultspresented in Fig. 10 show an increase in compressive strength

Fig. 8. Effects of sand dune and cement content on the maximum dry density and optimum water content of the mixtures.

Fig. 9. Variation in the compression strength and the tensile strength by splitting as a function of cement content for each percentage of sand.

Fig. 10. Effects of static compaction energy and the humidity degree on the compressive strength and the tensile strength by splitting of optimal mixture.

264 M. Bachar et al. / Construction and Building Materials 74 (2015) 259–267

more remarkable at early-age up to a relative stabilization at28 days, which is in accord with the study conducted by [1,17].However, the compressive strength at 28 days increases in a dryenvironment of the order of 46.18% whereas in humid environ-ment approximately 77.91% when the compacting pressure passesfrom 2.5 to 10 MPa. This also manifests at the tensile strength bysplitting which increases in a dry environment a rate of the orderof 40.62% and in environment humid a rate of the order of79.10%. These effects are explained by the cement hydration andthe C–S–H formation, subsequently improving compactness

thereby increasing the rigidity of cement-stabilized compressedearth blocks, which is in accord with the study conducted by[1,14–20].

3.3. Influence of aggregates of cork on the physical and mechanicalproperties of the optimal mixture

In the following section, the incorporation of aggregates of corksignificantly affects the mechanical properties for each level ofcompression. The results illustrated in Fig. 11 clearly show degres-

Fig. 12. Effect of incorporation of aggregates of cork on the maximum dry density ofthe optimal mixture for the different energy levels of static compaction.

2,00 Static compaction=sc sc=2.5 MPa

M. Bachar et al. / Construction and Building Materials 74 (2015) 259–267 265

sion increasing mechanical strengths both in a dry environmentand at wet environment, which translates effectively by creatinga porous network. However we note an evolution of the mechani-cal strength with increasing the compaction. The results of thecurve in a dry environment show that for 2% of the aggregates ofcork, the compressive strength increases in the order of 89.64%when the compacting pressure passes from 2.5 to 10 MPa whilewith 12% of aggregates of cork increase the compressive strengthof about 140.91%. This effect is explained by the fact that the aggre-gates of cork show largely the crush energy and subsequently theincrease in the plasticity zone. This is also evident in tensile bysplitting. For a static compacting of 10 MPa and 12% of the aggre-gates of cork, the compressive strength in the wet state decreasesa rate of 7.41% at 28 days relative to optimized mixture withoutthe incorporation of the cork.

It is also noted a more remarkable decrease in the maximumdry density as a function of aggregates of cork for each level of sta-tic compaction (Fig. 12). The experimental measurements showthat for a static compacting of 2.5 MPa maximum dry densitydecreases in the order of 29.36% when the percentage of corkincreases from 2% to 12% and for a static compacting of 10 MPadecreases in the order of 16.68%, which results in the increasedporosity that provides the aggregates of cork.

0 10 20 30 40 50 60 70 800,00

0,25

0,50

0,75

1,00

1,25

1,50

1,75 sc=5 MPa sc=7.5 MPa sc=10 MPaModel λ = − 0.017.ε + 1.5 : R

2=0.98

Model of Hashin and Shtrikman (Max) [35] Series model [35]Model of Hashin and Shtrikman (Min) [37]Model of Krischer [38]: Model of Maxwell [40]Modeling by auto-coherent homogenization [41]Parallel model [35]

Ther

mal

con

duct

ivity

in

Porosity of optimal mixture for different teneur of agreggates of cork (%)

dry

stat

e λ

(w/m

.k)

Fig. 13. Effects of aggregates of cork and static compaction on thermal conductivityand comparison with some models.

3.4. Influence of aggregates of cork and static compaction on thermalconductivity and comparison with some models

The effective thermal conductivity of porous material dependson many parameters: the thermal conductivities of the solid andgas, the degree of porosity, the size, shape and distribution of thepores. To predict the value of the effective thermal conductivityfor a given simple requires one the hand to have as much informa-tion on the microstructure and on the other hand to have analyticalor numerical tools that take into account the information. Geomet-ric simplifications are the starting point for all models. The conceptof this approach, considers that the material consists of a solid orcombined with a fluid phase (air) matrix. In this case, the thermalconductivity will be apparent according to the thermal conductiv-ities of the solid phase and the fluid phase (air). In this study it waspossible to measure the approximate value of the conductivity ofthe solid phase (soil–cement–sand dune) using the same measur-ing device (hot wire) and approaching a value ks = 1.5 W/m K andthe conductivity of the gaseous phase (area) at room temperatureequal kaire = 0.026 W/m K. In the study of the influence of porosityand conductivity of the two phases, the thermal conductivity wasconducted by comparison between the experimental results and

Fig. 11. Effects of incorporation of aggregates of cork and the humidity degree on the mcompaction.

the predictions, by analytical calculations using theoretical modelsand this is well represented in Fig. 13. In the same figure, we note aslight influence of compaction on thermal conductivity in the dry

echanical strengths of the optimal mixture for the different energy levels of static

266 M. Bachar et al. / Construction and Building Materials 74 (2015) 259–267

state. According to the results, we see that for zero aggregates ofcork, thermal conductivity increases at a rate of 3.75%, when thecompacting pressure passes from 2.5 to 10 MPa. The Maxwellmodel [40] is far from a prediction model, however, those ofKrischer and Kroll [38] and self-consistent homogenization [41],and the new model of linear, (see Fig. 13) perfectly correlate withexperimental results on all the range of porosity, except that thevalues of this last model diverge with experimental values whenthe porosity exceeds 80%.

4. Conclusion

The main objective of this study was to evaluate through exper-iments the influence of the addition of aggregates of cork on themechanical properties, the maximum dry density and thermal con-ductivity of compressed earth blocks stabilized cement, and thus acomparison of experimental measurements with some models pre-diction was made in order to choose the best of these models.Based on the results of this experimental study, the following con-clusions can be drawn.

� Composition of (58% soil–30% sand dune–12% cement) assuresmaximum dry density of 1750 kg/m3 with a compressivestrength in the dry state of 2.87 MPa at 28 days.� Compressive strength at 28 days increases in a dry environment

of the order of 46.18% whereas in humid environment approx-imately 77.91% when the compacting pressure passes from2.5 to 10 MPa. The use of static compaction is in the aim ofimproving the mechanical strengths by cons it increases thethermal conductivity grace to densifying the blocs this is whywe thought to incorporate aggregates of cork which has twoeffects: the lightening and the reduction in thermalconductivity.� Compressive strength and tensile of cement-stabilized com-

pressed earth blocks decrease with the increasing content ofaggregates of cork. Nevertheless, with minimal mass contentswe obtained acceptable strengths.� Regardless of the level of static compaction there is a notewor-

thy decrease in the maximum dry density with increasing con-tent of the aggregates of cork.� The values of the thermal conductivity of compacted composite

(matrix + aggregates of cork) were compared with some theo-retical models of prediction. Only models of Krischer and upperbound of Hashin and Shtrikman and that of auto-coherenthomogenization, give results near to those obtained experimen-tally. The new linear model remains meanwhile, the theoreticalmodel that estimates the best values of the thermal conductiv-ity, with some divergence when the porosity exceeds 80%.

References

[1] Bahar R, Benazzoug M, Kenai S. Performance of compacted cement-stabilisedsoil. Cem Concr Compos 2004;26(7):811–20.

[2] Al-Amoudi OSB, Khan K, Al-Kahtani NS. Stabilization of a Saudi calcareous marlsoil. Constr Build Mater 2010;24(10):1848–54.

[3] Basha EA, Hashim R, Mahmud HB, Muntohar AS. Stabilization of residual soilwith rice husk ash and cement. Constr Build Mater 2005;19(6):448–53.

[4] Horpibulsuk S, Rachan R, Chinkulkijniwat A, Raksachon Y, Suddeepong A.Analysis of strength development in cement-stabilized silty clay frommicrostructural considerations. Constr Build Mater 2010;24(10):2011–21.

[5] Horpibulsuk S, Katkan W, Naramitkornburee A. Modified Ohio’s curves: a rapidestimation of compaction curves for coarse- and fine-grained soils. GeotechTest J ASTM 2009;32(1):64–75.

[6] Horpibulsuk S, Katkan W, Apichatvullop A. An approach for assessment ofcompaction curves of fine-grained soils at various energies using a one pointtest. Soils Found 2008;48(1):115–25.

[7] Horpibulsuk S, Katkan W, Sirilerdwattana W, Rachan R. Strength developmentin cement stabilized low plasticity and coarse grained soils: laboratory andfield study. Soils Found 2006;46(3):351–66.

[8] Horpibulsuk S, Miura N, Nagaraj TS. Assessment of strength development incement-admixed high water content clays with Abrams’ law as a basis.Geotechnique 2003;53(4):439–44.

[9] Houssain KMA. Development of stabilized soils for construction applications.Ground Improv J 2010;163(3):173–85.

[10] Houssain KMA. Stabilized soils for construction applications incorporatingnatural resources of Papua New Guinea. Resour Conserv Recy2007;51(4):711–31.

[11] Maslehuddin M, Al-Amoudi OSB, Shameem M, Rehman MK, Ibrahim M. Usageof cement kiln dust in cement products–research review and preliminaryinvestigations. Constr Build Mater 2008;22(12):2369–75.

[12] Sariosseiri F, Muhunthan B. Effect of cement treatment on geotechnicalproperties of some Washington state soils. Eng Geol 2009;104(1–2):119–25.

[13] Goodary R, Lecomte-Nana GL, Petit C, Smith DS. Investigation of the strengthdevelopment in cement-stabilised soils of volcanic origin. Constr Build Mater2012;28(1):592–8.

[14] Attoh-Okine NO. Lime treatment of laterite soils and gravels-revisited. ConstrBuild Mater 1995;9(5):283–7.

[15] Rabehi M, Guettala S, Mezghiche B. The open porosity of concrete covers:correlation between the resistance to compression and initial absorption. Eur JEnviron Civil Eng 2012;16(6):730–43.

[16] Rabehi M, Mezghiche B, Guettala S. Correlation between initial absorption ofthe cover concrete the compressive strength and carbonation depth. ConstrBuild Mater 2013;45:123–9.

[17] Kenai S, Bahar R, Benazzoug M. Experimental analysis of the effect of somecompaction methods on mechanical properties and durability of cementstabilized soil. J Mater Sci 2006;41(21):6956–64.

[18] Taallah B, Guettala A, Guettala S, Kriker S. Mechanical properties andhygroscopicity behavior of compressed earth block filled by date palmfibers. Constr Build Mater 2014;59:161–8.

[19] Bastian G. Determination of thermophysical characteristics of buildingmaterials by the method of the plane source in asymptotic transientregimes. Rev Phys Appl (Paris) 1987;22(6):431–44.

[20] Guettala A, Houari H, Mezghiche B, Chebili R. Durability of lime stabilizedearth blocks. Courrier du Savoir 2002;2:61–6. Université Mohamed KhiderBiskra, Algérie.

[21] Venkatarama Reddy BV, Prasanna Kumar P. Cement stabilised rammed earth.Part A: compaction characteristics and physical properties of compactedcement stabilised soils. Mater Struct 2011;44(3):681–93.

[22] Morel JC, Pkla A, Walker P. Compressive strength testing of compressed earthblocks. Constr Build Mater 2005;21(2):303–9.

[23] Walker PJ. Bond characteristic of earth block masonry. J Mater Civil Eng1999;11(3):249–56.

[24] Meukam P, Noumowe A, Jannot Y, Duval R. Thermophysical and mechanicalcharacterization of stabilized clay bricks for building thermal insulation. MaterStruct 2003;36(7):453–60.

[25] Meukam P, Jannot Y, Noumowe A, Kofane TC. Thermo physical characteristicsof economical building materials. Constr Build Mater 2004;18(6):437–43.

[26] Pintor Ariana MA, Ferreira Catarina IA, Pereira Joana C, Correia Patrícia, SilvaSusana P, Vilar Vítor JP, et al. Use of cork powder and granules for theadsorption of pollutants: a review. Water Res 2012;46(10):3152–66.

[27] Miled K, Sab K, Le Roy R. Scale effect in the lightweight concrete withexpanded polystyrene. Colloque Microstructure et Propriétés des Matériaux.ENPC; 2005. p. 189–94.

[28] Aziz MA, Murphy CK, Ramaswamy SD. Lightweight concrete using corkgranules. Int J Lightweight Concr 1979;1(1):29–33.

[29] Hernandez-Olivares F, Bollati MR, Del Rio M, Parga-Landa B. Development ofcork–gypsum composites for building applications. Constr Build Mater1999;13(4):179–86.

[30] El-Bakkouri A, Ezbakhe H, Ajzoul T, El-Bouardi A. Thermomechanical study oflightweight concrete with cork and concrete lightened with olive pomace.12èmes Journées Internationales de Thermique. Tanger, Maroc, 15–17Novembre 2005. p. 307–10.

[31] Castro I, Simoes N, Tadeu A, Branco FG. Acoustic and thermal behaviour ofconcrete building blocks with cork. In: 6th Dubrovnik conference onsustainable development of energy water and environment systems.Dubrovnik Croatia; 2011. 8p.

[32] Gibson LJ, Ashby MF. Cellular solids: structure and properties. 2nd ed.(Paperback). Cambridge: Cambridge University Press; 1999. p. 453–67.

[33] Panesar DK, Shindman B. The mechanical, transport and thermal properties ofmortar and concrete containing waste cork. Cem Concr Compos 2012;34(9):982–92.

[34] Brás Ana, Leal Márcio, Faria Paulina. Cement–cork mortars for thermal bridgescorrection. Comparison with cement–EPS mortars performance. Constr BuildMater 2013;49(12):315–27.

[35] Wiener O. Abh Math Phys Klasse Königl Sächs Ges Wiss 1912;32:509.[36] Bessenouci MZ, BibiTriki EN, Khelladiet S, Abene A. The apparent thermal

conductivity of pozzolana concrete. Phys Proc 2011;21:59–66.[37] Hashin Z, Shtrikman S. A variational method of the theory of effective

magnetic permeability of multiphases materials. J Appl Phys 1962;33:3125–31.

[38] Krischer D, Kroll K. Technique de séchage, centre technique des industriesaéraulitiques et termiques, traduction du. Berlin: Springer-verlag; 1963.

[39] Willey MRJ, Soutwick AR. J Petrol Technol 1954;6:44.[40] Maxwell JC. A treatise on electricity and magnetism, vol. 1. 2nd ed. Dover, New

York; 1954 [Chapter 9 article 314].

M. Bachar et al. / Construction and Building Materials 74 (2015) 259–267 267

[41] Arnaud L, Monnet H, Cordier C, Sallet F. Modélisation par homogénéisationauto cohérente de la conductivité thermique du béton et laines de chanvre. In:Proceedings of the Congre‘s franc�ais de thermique, 15–17 mai 2000, ElsevierEd. A. Lallemand et J.F. Leone; 2000. p. 543–8.

[42] Ledhem A. Contribution to the study of a wood concrete: development of amethod for minimizing dimensional changes of a composite clay-cement-wood. Thèse de Doctorat de l’INSA de Lyon, France; 1997.

[43] British Standards Institution BS1377-4. Methods of tests for soils for civilengineering purposes. Part 4: compaction related tests; 1990. 70p.

[44] ASTM D1633-00. Standard test methods for compressive strength of moldedsoil–cement cylinders; 2000.

[45] ASTM 496-96. Standard test method for splitting tensile strength of cylindricalconcrete specimens; 1996.

[46] Day RL, Marsh BK. Measurement of porosity in blended cement pastes. CemConcr Res 1988;18:63–73.

[47] Matusinovic T, Sipusic J, Vrbos N. Porosity–strength relation in calciumaluminate cement pastes. Cem Concr Res 2003;33:1801–6.

[48] Papayianni I, Stefanidou M. Strength–porosity relationships in lime-pozzolanmortars. Constr Build Mater 2006;20:700–5.

[49] Chindaprasirt P, Rukzon S. Strength, porosity and corrosion resistance ofternary blend Portland cement, rice husk ash and fly ash mortar. Constr BuildMater 2008;22:1601.

[50] ASTM C1113-99(2004). Standard test Method for thermal conductivity ofrefractories by Hot Wire (platinum resistance thermometer technique).Annual book of ASTM Standards; vol. 15.01, West conchohoken: ASTMInternational; 2004.

[51] Wechsler AE. The probe method for measurement of thermal conductivity. In:Maglic KD, Cezairliyan A, Peletsky VE, editors. Compendium of thermophysicalproperty measurement methods. Vol 2. Recommended measurementtechniques and practices. New York, London: Plenum Press; 1992. p. 281.