Materials and Design 72 (2015) 51–62

Contents lists available at ScienceDirect

Materials and Design

journal homepage: www.elsevier .com/locate /matdes

The mix design for self-compacting high performance concretecontaining various mineral admixtures

http://dx.doi.org/10.1016/j.matdes.2015.01.0060261-3069/� 2015 Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: F.A. Finger-Institute for Building Materials Science,Faculty of Civil Engineering, Bauhaus-University Weimar, Germany. Tel.: +49 3643584765; fax: +49 3643 584759.

E-mail addresses: lehautc@daad-alumni.de (H.T. Le), matthias.mueller@uni-weimar.de (M. Müller), karsten.siewert@uni-weimar.de (K. Siewert),horst-michael.ludwig@uni-weimar.de (H.-M. Ludwig).

1 Tel.: +49 3643 584807, fax: +49 3643 584759.2 Tel.: +49 3643 584725, fax: +49 3643 584759.3 Tel.: +49 3643 584761, fax: +49 3643 584759.

Ha Thanh Le a,b,⇑, Matthias Müller a,1, Karsten Siewert a,2, Horst-Michael Ludwig a,3

a F.A. Finger-Institute for Building Materials Science, Faculty of Civil Engineering, Bauhaus-University, Weimar, Germanyb Institute of Construction Engineering, University of Transport and Communications, Hanoi, Viet Nam

a r t i c l e i n f o a b s t r a c t

Article history:Received 1 August 2014Accepted 21 January 2015Available online 20 February 2015

Keywords:Self-compacting high performance concreteMix design methodRice husk ashEfficiency factorSelf-compactabilityCompressive strength

This paper is an effort towards presenting a new mix design method for self-compacting high perfor-mance concrete (SCHPC) containing various mineral admixtures (MA). In the proposed method, the con-stituent materials were calculated by using the absolute volume method. The packing theory of Funk andDinger with the exponent q = 0.25 was adopted to determine the grading of aggregate. The primary pastevolume for filling capacity was computed from the void content of compacted aggregate. The superplas-ticizer dosage for the concrete was set on the basis of the superplasticizer saturation dosage of the cor-responding mortar. Efficiency factors were used to express the effect of MAs on compressive strength ofconcrete. The results show that the method was adequate to proportion SCHPC mixtures containing tern-ary binders, i.e. cement and two different MAs (rice husk ash (RHA), silica fume, fly ash, and lime stonepowder), satisfying the self-compactability requirements and compressive strength class in the range ofC60/75–C90/105. With 5–20 wt.% cement replacement, RHA was very effective in improving compressivestrength of SCHPC. The efficiency factor for RHA, i.e., 2.7–1.8, which is the first time applied, is onlymarginally lower as compared to that of silica fume.

� 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Self-compacting concrete is a concrete that can flow and con-solidate under its own weight, pass through the spaces betweenthe reinforcement bars to completely fill the formwork, and simul-taneously maintain its stable composition [3–5]. Self-compactinghigh performance concrete (SCHPC) is defined as a new generationof concretes on the basis of the concepts of self-compacting con-crete (SCC) and of high performance concrete (HPC). A methodfor proportioning SCHPC aims at fulfilling the self-compactabilityrequirements of SCC (filling ability, passing ability, and segregationresistance), as presented in Table 1, and of high compressivestrength and good durability of HPC [6]. To realize this goal, a highvolume of Portland cement, a very high dosage of chemicaladmixtures, i.e. super plasticizer (SP) and viscosity modifying

admixtures, and reactive mineral admixtures (MA), e.g. silica fume(SF), are used [5–7]. Hence, high costs and environmental impactconstitute the main disadvantage of SCHPC. The performance ofSCHPC is highly improved by using SF however it is expensivedue to the limited availability especially in developing countries.Rice husk ash (RHA), with its high amorphous silica content, is avery good replacement for SF with regards to compressive strengthand durability of concrete [6,8–12]. The reactive RHA is the residueof under proper conditions completely incinerated rice husk. Ricehusk, the outer covering of a rice kernel, is an agricultural wastefrom the milling process of paddy. Rice husk is abundant in manyrice cultivating countries, e.g. Vietnam, India and China. Normally,rice husk from paddy rice mills is disposed directly into the envi-ronment or sometimes is dumped or burnt in open piles on thefields. This results in serious environmental pollution, especiallywhen it is disintegrated in wet conditions. Substitution ofless-expensive RHA for SF as a partial cement replacement, notonly improves the sustainability of SCHPC, but also reduces envi-ronmental pollution from the disposal of rice husk.

It is well known that mix design is of major importance for theconcrete production process. The mix design can be understood ascombining optimum proportions of the constituent materials tofulfill the requirements of fresh and hardened concrete for a par-ticular application [13]. Generally, in the mix proportioning of

Nomenclature

A the designed air content (vol.%)Ai the content of the air-dry aggregate i (kg/m3)ai the ratio of aggregate i to aggregate blend (wt.%)AB the content of aggregate blend (kg/m3)B the binder content (cement plus MA) (kg/m3)C0 the cement content in the MA-blended concrete (kg/m3)Dmin, Dmax the minimum and maximum particle sizes in the

aggregate blend (mm)fc,cube the characteristic minimum cube compressive strength

(MPa)fc,dry,cube the average cube compressive strength cured in dry

conditions (MPa)ki the efficiency factor of MAi

MAithe moisture content of aggregate i (wt.%)

n the number of MA usedP(D) the weight percentage of aggregate passing the sieve

with size D (wt.%)PMAi

the content of MAi (kg/m3)pi the percentage of cement replaced by MAi (wt.%)RS,A the coefficient related to the shape (S) and the angu-

larity (A) of aggregate in the range of 1–5 [1]

SP the dosage of SP (wt.%)Ssp the solid content of SP (wt.%)VAB the absolute volume of the aggregate blend (m3)VAi

the absolute volume of aggregate i (m3)Vexp the excess paste volume (vol.%)Vp the primary paste volume required for filling ability

(vol.%)Vvoid the void volume of the compacted aggregate blend in

concrete (vol.%)Voids the void content in compacted aggregate blend (vol.%)WAAi

the water absorption of aggregate i (wt.%)Wad the adjusted water content (kg/m3)cAB the bulk density of dry aggregate blend (kg/m3)Dfc the allowance for compressive strength (6–12 MPa)

regulated in DIN EN 206-1 [2]qAi

the density of aggregate i (kg/m3)qAB the density of dry aggregate blend (kg/m3)qc the density of cement (kg/m3)qMAi

the density of MAi (kg/m3)

52 H.T. Le et al. / Materials and Design 72 (2015) 51–62

ordinary concrete, the required compressive strength is the primecriterion. For SCHPC, however, self-compactability, compressivestrength and durability are equally taken into account in propor-tioning mixtures [6]. Regarding the properties of SCC/SCHPC, thereexist two main mix design approaches to proportioning mixtures.One approach emphasizes on self-compactability, and ignores ordoes not give equal importance to compressive strength and dura-bility as presented in [1,4,14,15]. This approach does not renderproperly controlling the compressive strength, nor reaching highcompressive strength levels due to the high W/B ratio determinedfrom the water demand of the binder. The other approach consid-ers self-compactability as well as compressive strength as the tar-gets of mix design for SCC/SCHPC. Compressive strength ofSCC/SCHPC designed by these methods ranges mostly from 30 to90 MPa [16–18]. This kind of method does not take into consid-eration the effect of MAs on self-compactability and compressivestrength of SCC/SCHPC. Regarding the utilization of MAs inSCC/SCHPC, several individual MAs such as FA, RHA, SF, and groundgranulated blast-furnace slag (GGBFS), are taken into considerationin the mix design to satisfy adequate self-compactability andrequired compressive strength [6,19–22]. Each MA has its ownadvantages and disadvantages. Combination of various MAs canexploit their advantages and increase the cement replacement lev-el. In order to take into account the effect of MAs on the mechanicalproperties of concrete, the concept of ‘‘efficiency factor’’ has beendeveloped. The efficiency factor is empirically determined for thegiven content of materials and the exposure conditions [23–25].The efficiency of RHA and other MAs will be presented in the nextsection. The efforts to relate the efficiency factor of MAs and com-pressive strength in mix-design for SCC/SCHPC are scarce in the lit-erature, especially when more than one MA is used as cementreplacement to achieve a particular high compressive strengthexceeding 90 MPa.

In this study, a simple mix design method for SCHPC wasdeveloped on the basis of the cementitious efficiency of mineraladmixtures and of the requirements for ordinary concreteproportions laid down in DIN EN 206-1 [2] and DIN EN 1045-2[26]. The proposed method was applied to design mixtureproportions of SCHPC containing ternary binders, i.e. cement andtwo different MAs, i.e. RHA, SF, FA, and LSP, at various low W/Bratios.

2. Cementitious efficiency of RHA and other mineral admixtures

All kinds of MAs, i.e. nearly inert, pozzolanic and latent hydrau-lic MAs, have been used to produce SCHPC. Where LSP is nearlyinert or low reactive. SF, RHA, metakaolin, and low calcium classF-FA (according to ASTM: C618) are pozzolanic. And groundgranulated blast-furnace slag and high calcium class C-FA (accord-ing to ASTM: C618) are both latent hydraulic and pozzolanic MAs[6,8,13]. Each type of MA exerts different effects on properties ofboth fresh and hardened concrete depending on its characteristicand replacement levels. In terms of compressive strength, theeffect of MA can be expressed as an efficiency factor (k-value).The efficiency factor is defined as the portion of MA, which canbe considered as equivalent to Portland cement in aMA-containing concrete. A k-value of a MA equal to 1 indicatesthat the MA is equivalent to cement. On the contrary, a k-value lessthan 1 implies that the MA underscores cement as to its effect oncompressive strength. The content of a MA can be multiplied bythe k-value to convert to the equivalent cement content [23].

Recently, the efficiency factors for compressive strength of cal-cined kaolin and SF have been determined by the procedure pro-posed by Wong and Abdul Razak [25]. The k-value of a MA isobtained from the ratio of compressive strength of theMA-blended mixture to the control mixture (containing 100%OPC). It is generally concluded that the k factors increase withage but decrease with higher pozzolanic content. It was alsoobserved that changes in W/C ratio from 0.33 to 0.27 did not sig-nificantly affect the resultant efficiency factors. The fundamentalprinciple of Abram’s rule is applied in the method. The compressivestrength of the MA-blended mixture is inversely proportional tothe water to equivalent cement content ratio (W/Ceq), where theequivalent cement content is C0 + kPMA. The compressive strengthof Portland cement concrete, fc, can be expressed by:

f c ¼ K11

W=C

� �ð1Þ

The compressive strength of concrete containing a MA, fMA, canin analogy with Eq. (1) be expressed by:

f MA ¼ K21

W=ðC 0 þ kPMAÞ

� �ð2Þ

Table 1The requirements for the properties of fresh SCC/SCHPC according to DIN EN206-9 [49].

Slump-flow Viscosity Passing ability Segregation resistance

Class Slump-flow (mm)a Class T500 (s)b V-funnel (s) Class J-ring step height (mm) Class Sieve segregation (%)

SLF1 550–650 VS1/VF1 <2 <9c PJ1 610 with 12 bars SR1 620SLF2 660–750 VS2/VF2 P2 9–25d PJ2 610 with 16 bars SR2 615SLF3 760–850 – – – – – –

Permissible deviation:a ±50 mm.b ±1 s.c ±3 s.d ±5 s.

fMA = -0.027p2 + 1.049p + 63.207R² = 0.97

40

50

60

70

80

Com

pres

sive

stre

ngth

at 2

8 da

ys (M

Pa)

RHA content (wt.%)0 5 10 15 20 25 30 35

Fig. 1. Compressive strength vs. percentage replacement by RHA.

k28 = 0.0007p2 - 0.076p + 3.056R² = 0.99

0

1

2

3

4

0 5 10 15 20 25 30 35

Effic

ienc

y fa

ctor

at 2

8 da

ys

RHA content (wt.%)

Fig. 2. Efficiency factor vs. percentage replacement by RHA.

Table 2Efficiency factors of RHA and other mineral admixtures.

Mineral admixtures LSP [50] FA [50] SF [25] RHAa

Content (wt.%) 15–55 20–60 5–15 5–20k-value 0.29 0.56 3.1–2.1 2.7–1.8

a Proposed by the authors.

H.T. Le et al. / Materials and Design 72 (2015) 51–62 53

in which the k-value relevant for the specific MA is used to estimatethe equivalent cement content. Further, K1 and K2 are proportion-ality constants. In Eqs. (1) and (2), C is cement content in the controlmixture (kg/m3); C0 is cement content in the MA-blended mixture(kg/m3); and PMA is MA content (kg/m3).

It is assumed that K1 is equal to K2, because the mixture propor-tions, W/C ratio, curing history and testing conditions for the con-trol and the MA-blended mixtures are similar. Therefore, k can beassessed for a specific MA by dividing Eq. (2) by Eq. (1), yielding:

k ¼ ðf MA=f cÞC � C 0

PMAð3Þ

When p is the percentage cement replacement by the MA, Eq.(3) transforms into:

k ¼ 1þ ðf MA=f cÞ � 1p

ð4Þ

In this study, the strength efficiency factor of RHA was deter-mined by the procedure above. The control and RHA-blended mor-tars with cement–sand–water proportions of (1/3/0.5) wereprepared. For the RHA-blended mortars, Portland cement (CEM I52.5 R) was partially replaced by 5, 10, 15, 20, 25, 30 wt.% RHA (aver-age grain size of 7.7 lm), and superplasticizer was introduced tomeet the flow of the control mortar determined in accordance withDIN EN 1015-3 [27]. Mortar specimens having dimensions of40 � 40 � 160 mm3 were cast and cured in molds at temperatureof 20 �C and 95% relative humidity for one day. After demoldingthe specimens were stored in water at 20 �C until testing at 28 days.Compressive strength was tested according to DIN EN 196-1 [28]. Sixspecimens of each mixture were tested and the average values werereported. The average 28 day-compressive strength of mortars andefficiency factor at different percentage replacements by RHA areshown in Figs. 1 and 2 respectively.

Based on the experimental data in this study, a model for theefficiency factor at 28 days (k28) at different percentage replace-ments by RHA is given by:

k28 ¼ 0:0007p2 � 0:076pþ 3:056 ð5Þ

According to this model, the k-values for RHA with 5–20 wt.%cement replacement decline from 2.7 to 1.8. In Table 2, k-valuesfor RHA and other MAs at concrete age of 28 days are presented.

3. Proposed mixture design procedure

The proposed method considers SCHPC to consist of two phas-es: aggregate and paste. The grading of the aggregate blend isdetermined in accordance with the packing theory of Funk andDinger with the exponent q = 0.25. The paste volume is determinedon the basis of a void content in the compacted aggregate blend.Adequate self-compactability is mainly governed by the paste vol-ume, the type and content of MA, and by the SP dosage. The W/Bratio and the cementitious efficiency of MA are mainly taken into

account to achieve the required compressive strength. The pro-posed procedure for proportioning SCHPC containing variousMAs is schematically presented in Fig. 3 and follows the basicsteps, listed thereafter.

Determine the cement, mineral admixtures, and water contents

Determine the content of aggregate components

Conduct trials to test fresh properties

Determine the SP saturation dosage of mortar

Determine voids of the compacted aggregate blend

Determine the primary paste volume

Determine the water-binder ratio

Is expected compressive strength fulfilled?

Conduct trials to test hardened properties

Designed mixture proportions

Adjust w/b

Are fresh properties fulfilled?

Adjust SP

Adjust Vp

Select and test the constituent materials

Inputs on expected performance (fc, SLF, VF, SR, PJ)

Yes

No

Yes

No

Fig. 3. Procedure of mix design method for proportioning SCHPC.

0

20

40

60

80

100

0.10 1.00 10.00

Pass

ing

(%)

Sieve size (mm)

Real aggregateFunk & Dinger C16A16

Fig. 4. Aggregate grading ranges for SCHPC.

54 H.T. Le et al. / Materials and Design 72 (2015) 51–62

3.1. Step 1: determination of the void content of compacted aggregateblend

The ratio of aggregate components are computed on the basis ofFunk and Dinger theory with exponent q = 0.25 [29], which followsEq. (6). The optimum ratio of aggregate results in the particle sizedistribution of aggregate, which follows the ideal curve with theminimum deviation.

PðDÞ ¼ Dq � Dqmin

Dqmax � Dq

min

ð6Þ

The ideal grading curve of the Funk and Dinger theory withq = 0.25, Dmax = 16 mm and Dmin = 0.63 mm is applied. Thegrading curve of the aggregate used in this study and the requiredrange of aggregate grains for ordinary concrete (A16 and C16) com-plying with DIN EN 206-1 [2] and DIN 1045-2 [26] are shown inFig. 4.

The bulk density of compacted aggregate blend is determinedby experiment with the determined ratio of the aggregate compo-nents. The mixed aggregate blend was filled into a 8 l container andvibrated in two times of 2 min each. The bulk density of the

compacted aggregate was calculated from the weight and volumeof aggregate in the container. Next, the void content of the com-pacted aggregate blend can be calculated by:

H.T. Le et al. / Materials and Design 72 (2015) 51–62 55

Voids ¼ 1� cAB

qAB

� �� 100 ð7Þ

qAB ¼Xn

i¼1

VAi

VAB� qAi

ð8Þ

Fig. 5. Relationship between compressive strength and W/C ratio; Walz curve [52].

3.2. Step 2: determination of the primary paste volume

The paste volume in SCHPC is significantly larger than that inordinary concrete. It is required to fill the void volume betweenaggregate particles, and make sufficient lubricating layers on thesurface of aggregate particles. The lubricating layers reduce thefriction between the aggregate particles, and hence increase theflowability of concrete [6]. The primary paste volume can be calcu-lated via the equations given by Koehler and Fowler [1]. So,

Vp ¼ Vexp þ Vvoid ð9Þ

Vp ¼ Vexp þ Voids � ð100� VexpÞ100

ð10Þ

Vexp ¼ 8þ 16� 84

� �ðRS;A � 1Þ ð11Þ

The paste volume should range from 30% to 42% by volume ofthe concrete according to [1,4,30].

3.3. Step 3: determination of the water–binder ratio

Established relationships between components of normal con-crete, such as W/C vs. compressive strength, can also be appliedto SCC/SCHPC [13,17]. As a consequence, the average compressivestrength can be determined in accordance with DIN EN 206-1 [2]and DIN 1045-2 [26] from:

f c;dry;cube ¼f c;cube þ Df c

0:95ð12Þ

Thereupon, the equivalent W/Ceq ratio can be calculated fromthe relationship between compressive strength and W/C ratio(Walz curve) presented in Fig. 5. For the durability, the W/C ratioshould satisfy the limit values regulated in DIN EN 206-1 [2] andDIN 1045-2 [26].

When the type and content of MA replacing the cement areknown, the W/B ratio can be calculated via Eq. (13).

WB¼ W

Ceq� 1�

Xn

i¼1

pi

!þXn

i¼1

pi � ki

( )ð13Þ

The W/B ratio of HPC is generally in the range of 0.25–0.4 [9,31],while the W/B ratio of SCC ranges from 0.26 to 0.48 [30]. Therefore,the appropriate W/B ratio for SCHPC should be between 0.25 and0.4. A W/B ratio below 0.25 can be used when SP and paste compo-sition can yield a low enough plastic viscosity.

3.4. Step 4: determination of cement, mineral admixture, and watercontents

The content of the cement, of each MA and of the water canbe determined on the basis of the known paste volume (Vp), W/Bratio, and the cement content replaced by each MA. By solvingEq. (14) for B, the binder content is obtained. Then thecement, both mineral admixtures and water contents can be calcu-lated by:

B �1�

Pni¼1pi

� �qc

þ B �Xn

i¼1

pi

qMAi

!þ B � W

B

� �¼ Vp ð14Þ

C 0 ¼ 1�Xn

i¼1

pi

!� B ð15Þ

PMAi¼ pi � B ð16Þ

W ¼ WB

� �� B ð17Þ

The binder contents should be in the range from 425 to625 kg/m3 [30], while the cement content should not be lower thanthe limit values regulated for durability in DIN EN 206-1 [2] andDIN 1045-2 [26]. The water content should be lower than200 kg/m3 [15].

3.5. Step 5: determination of aggregate content

Based on the known paste volume and the ratio of aggregatecomponents, the content of aggregate blend and then individualaggregate components can be calculated by using the followingequations:

AB �Xn

1

ai

qAi

!þ Vp þ A ¼ 1000 ð18Þ

Ai ¼ ai � AB � ð1þMAiÞ ð19Þ

The coarse aggregate content should vary from 28 to 38 vol.% ofthe concrete [30].

3.6. Step 6: determination of the dosage of superplasticizer

The superplasticizer saturation dosage (SSD) of the correspond-ing mortar is used as the primary SP demand for SCHPC. The SSD isdetermined by testing the mini-cone slump flow of mortar, asdescribed in EFNARC [15] (Figs. 6 and 7). The SSD is defined as a

56 H.T. Le et al. / Materials and Design 72 (2015) 51–62

SP dosage beyond which the mini-slump flow no longer increasesor reaches a maximum mini-slump flow [6].

3.7. Step 7: adjustment of the water content

The water absorption of aggregates and the water contributionof SP should be considered to adjust the content of mixing water.Therefore, the adjusted water content can be determined by:

Wad ¼W þXn

i

Ai

1þMAi

� ðWAAi�MAi

Þ � SP � Bð1� SspÞ ð20Þ

Fig. 6. Mini-cone slump flow and T250 time test.

0

5

10

15

20

25

100

150

200

250

300

350

1 1.5 2 2.5 3 3.5

T250

tim

e (s

ec)

Min

i-slu

mp

flow

(mm

)

Superplasticizer dosage in mortar

Mini-slump flow

T250 time

Fig. 7. SSD of mortar formulated from SCHPC mixture ‘‘30LSP20R10’’ determinedby mini-cone slump flow.

Table 3Chemical composition (wt.%) and physical properties of cement and MAs.

Chemical analyses Cement RHA USF DSF FA LSP

SiO2 19.4 87.0 96.6 96.2 56.6 10.9Al2O3 5.3 0.8 0.7 0.7 25.8 4.2Fe2O3 2.5 0.4 0.2 0.3 6.4 1.3CaO 61.2 1.2 0.3 0.0 2.5 46.8MgO 1.2 0.6 0.4 0.1 1.3 1.2SO3 3.2 0.4 0.1 0.1 0.6 0.6Na2O 0.07 0.4 0.16 0.06 0.62 0.3K2O 0.61 2.63 0.65 0.37 2.08 1.02LOI 4.9 3.7 0.9 1.6 2.9 34.0Density (kg/m3) 3090 2270 2260 2260 2270 2740MPS (lm) 7.07 5.7; 7.7 0.35 0.29 16.39 7.88Specific surface area

BET [Blain](103 m2/kg)

[0.595] 25.21(5.7 lm) 18.09 26.43 2.14 6.8823.05(7.7 lm)

Water demand byPuntke method[51] (10�3 m3/kg)

0.34 0.57 (5.7 lm) 0.53 0.53 0.33 0.370.61 (7.7 lm)

Pore volume by BJHmethod (10�3 m3/kg)

– 0.08 (5.7 lm) – – – –0.106 (7.7 lm)

LOI-loss on ignition, DSF-densified SF, USF-undensified SF.

3.8. Step 8: trial mixtures and adjustments

Trial batches are prepared using the proportions calculatedabove. To examine whether the mixture proportions designed bythe proposed method could meet the self-compactability and com-pressive strength requirements, the quality control tests should becarried out.

When the results of the quality control tests do not fulfill therequirements of the fresh concrete (Table 1), adjustments will bemade until all properties of SCHPC meet the requirements expect-ed in the design. For example, when the filling and passing abilitiesare poor, the SP dosage should be increased first. When the fillingand passing abilities cannot be achieved by adjusting SP dosage,increase in paste volume leads to higher filling and passing abilitiesfor a given SP dosage.

4. Validation of the proposed method

4.1. Materials

The materials used in this study were Portland cement (CEM I52.5 R), RHA, undensified SF (USF), densified SF (DSF), FA, LSP, nat-ural sand (0–2 mm), and crushed basalt stone (2–5 mm, 5–8 mm,8–11 mm, 11–16 mm). Rice husk from Vietnam was burnt underappropriate temperature conditions, and the obtained ash wasthereupon ground in a ball mill. In the present study, RHA wasused with two different mean particle sizes of 5.7 lm (RHA5.7)and 7.7 lm (RHA7.7). The first type was used in mixtures withW/B ratio of 0.26 and the second one in the mixtures with W/Bratio of 0.30 and 0.34. The physical properties and the chemicalcomposition of cement and MAs are summarized in Table 3.Their particle size distributions are shown in Fig. 8. In Fig. 9, SEMimages show the morphology of RHA and of other MA particles.Obviously, the ground RHA is a porous material with macro pores(>50 nm) and meso pores (2–50 nm) on the surface and inside theparticles. The pore structure of RHA has been analyzed in detail inthe previous studies [32,33]. The physical properties of aggregateare presented in Table 4, and its particle size distribution is shownin Fig. 4. In addition, a polycarboxylate-based SP with density of1080 kg/m3 and 40 wt.% solid content was used.

4.2. Mixture proportions

Nine SCHPC mixtures were designed to validate the proposedmethod. They contained ternary binders, i.e. cement and two dif-ferent MAs, such as RHA, SF, FA and LSP. The aim of the designwas to fulfill slump flow class SLF2, viscosity class VF2, passingability class PJ2, segregation resistance class SR2 (Table 1) andcompressive strength classes C60/75, C70/85, C80/95, andC90/105 as regulated in standards DIN EN 206-1 [2] and DIN1045-2 [26]. The ratio of aggregate components was computedon the basis of Funk and Dinger with the minimum deviation,yielding: natural sand (0–2 mm): 45.0 wt.%, basalt stone

0

20

40

60

80

100

0.01 1 100

Cum

ulat

ive

pass

ing

(vol

.%)

Particle size (µm)

USFDSFRHA5.7RHA7.7CementLSPFA

Fig. 8. Particle size distribution of cement and MAs used in this study.

Fig. 9. SEM images of undensified SF (a), densified SF (b), porous structu

H.T. Le et al. / Materials and Design 72 (2015) 51–62 57

(2–5 mm): 28.0 wt.%, (5–8 mm): 6.2 wt.%, (8–11 mm): 11.2 wt.%,(11–16 mm): 9.6 wt.%. The particle size distribution of theaggregate blend is illustrated in Fig. 4. The bulk density of com-pacted aggregate blend was determined by experiment,cAB = 2170 (kg/m3), while the density of aggregate blend was calcu-lated from Eq. (8), qAB = 2930 (kg/m3). Next, the void content ofcompacted aggregate blend was calculated by using Eq. (7),Voids = 26 vol.%. The proportions of constituent materials weredetermined in accordance with the procedure, as mentioned inSection 3. Summarizing, the volume fractions of paste, fine andcoarse aggregates are 0.365, 0.298 and 0.317 (m3), respectively.The air content was set at 2 vol.% (0.02 m3) for non-air entrainedconcrete. The mixture proportions are presented in Table 5. Themixture types were designated on the basis of W/B ratio, typeand percentage of MA replacing cement by weight. For instance,in the mixture ‘‘30LSP20R10’’, 30 wt.% cement content wasreplaced by 20 wt.% LSP, and 10 wt.% RHA, and the W/B ratio was0.30.

re of RHA7.7 (c), FA (d), porous structure of RHA5.7 (e) and LSP (f).

Table 4Physical properties of the fine and coarse aggregate.

Properties Natural sand Basalt stone

Fineness modulus 2.32 6.14Density (kg/m3) 2650 3050Water absorption (wt.%) 0.08 0.8Moisture content (wt.%) 0.0 0.2

1 min 8 min2 min2 minCement

MAs80 % Water

20 % Water

SP

Endof

mixing

Fine, coarse

aggregate

Fig. 10. Mixing procedure for SCHPC.

58 H.T. Le et al. / Materials and Design 72 (2015) 51–62

4.3. Experimental methods

All SCHPC mixtures were prepared in a Pemat ZK30 mixer withtotal mixing time of 13 min. The mixing procedure is shown inFig. 10.

SSD of mortar was determined by testing the mini-cone slumpflow as described in EFNARC [15]. Flow rate of mortar was alsoevaluated by T250 time, which is the spread time for a diameterof 250 mm, corresponding to T500 time measured in SCC/SCHPC[15,34] (Figs. 6 and 7).

To examine whether the mixture proportions designed by theproposed method could meet the self-compactability and com-pressive strength requirements, the following tests were carriedout.

Slump flow test: test was carried out in accordance with DIN EN12350-8 [35] to measure filling ability and flow rate T500, whichindicate plastic viscosity of the fresh concrete.

V-funnel test: test was carried out in accordance with DIN EN12350-9 [36] to measure the flow time through the V-funnel.The plastic viscosity of fresh concrete can be evaluated based onthe V-funnel flow time, and the arching effect of aggregate canbe detected as well.

J-ring test: test was carried out in accordance with DIN EN12350-12 [37] to assess filling ability and passing ability of freshconcrete between reinforcement bars.

Sieve segregation test: test was carried out in accordance withDIN EN12350-11 [38]. The test aims at investigating the resistanceof SCHPC to segregation by measuring the portion of the freshSCHPC sample passing through a 5 mm sieve. If the SCHPC has poorresistance to segregation, the paste or mortar can easily pass thesieve. Therefore, the sieved portion indicates whether the SCHPCis stable or not.

Compression test: cubic specimens of 150 � 150 � 150 mm3 forcompressive strength were cast without vibration and compaction.After 1 day, the specimens were demolded, stored in water at20 ± 2 �C for further 6 days, and then cured in a room under con-trolled temperature (20 ± 2 �C) and humidity (65 ± 5%) conditionsuntil testing at 7 and 28 days according to DIN EN 12390-2 [39].Compressive strength of concrete was determined under DIN EN12390-3 [40]. Three specimens of each mixture were tested, andthe average values are reported.

Table 5Mixture proportions of SCHPC.

Mixtures W/Ceq

W/B Water (kg/m3)

Cement (kg/m3)

LSP (kg/m3)

FA (kg/m3)

SF (kgm3)

34LSP20R10 0.34 0.34 182 374 107 0 034FA20R10 0.32 0.34 178 366 0 104 034FA20R15 0.32 0.34 176 336 0 104 034FA20R20 0.32 0.34 175 308 0 103 030LSP20R10 0.30 0.30 168 397 113 0 030FA20R10 0.28 0.30 170 388 0 111 026FA20R10 0.24 0.26 155 413 0 118 026FA20DSF10 0.24 0.26 155 413 0 118 5926FA20USF10 0.24 0.26 155 413 0 118 59

a Percentage ratio of SP demand for concrete to SP saturation dosage of mortar.

4.4. Self-compactability of SCHPC

The experimental results of self-compactability of fresh con-crete are presented in Table 6. The slump flow of the SCHPCs ran-ged from 730 to 780 mm, without signs of bleeding andsegregation. The V-funnel flow times and T500 values were in therange of 9.2–22.3 s and of 2.9–7.0 s, respectively. The J-ring stepheight is equal to 10 mm in most cases. The sieve segregation indexranged from 4.7 to 13.6 wt.%. This indicates that the SCHPCs hadexcellent filling ability, good plastic viscosity, adequate passingability and good segregation resistance. All of the SCHPC mixturesmeet the requirements of slump flow class SLF2, viscosity classVF2, passing ability class PJ2, segregation resistance class SR2, asshown in Table 1.

It is well known that the self-compactability of SCHPC is gov-erned by paste composition and SP dosage when the paste volumeand aggregate volume are fixed. All trial batches of concretes weremade with the SSDs of the corresponding mortars. Interestingly,the SP demand of SCHPC to meet self-compactability requirementswas similar to the SSD of the corresponding mortar, as shown inTable 5 and Fig. 11. Mixtures proportioned with W/B ratio of 0.26satisfied the requirements of self-compactability with the SSDs ofthe corresponding mortars. For mixtures proportioned with W/Bratio of 0.30 and 0.34, SP demand for the requirements ofself-compactability of concrete was slightly adjusted in some casescompared with SSD of the corresponding mortars. For instance, themixture ‘‘34LSP20R10’’ made with SSD had a slump flow of650 mm, T500 value of 2.8 s, and sieve segregation index of3.8 wt.%. However, the J-ring step height was 20 mm significantlyhigher than the standardized value of 10 mm (Table 1). After SPdosage was increased from the SSD of 2.0 to 2.25 wt.%, all require-ments for self-compactability were fulfilled (Table 6).

The results of properties of fresh SCHPC show that the mixturesincorporating LSP need a much higher SP dosage than the mixturesincorporating FA to meet the required self-compactability(Table 5). It is caused by LSP having angular particles with roughsurface and smaller size, and hence with a larger specific surfacearea. On the contrary, FA has spherical particles with larger sizeand hence has a lower specific surface area and water demand(Table 3 and Fig. 9). As a result, the incorporation of LSP into themixtures leads to an increase in viscosity of paste and hence adecrease in slump flows of concrete. To reach the required flowingability, a higher SP dosage must be used in the mixture containingLSP, compared to the mixture containing FA.

/ RHA (kg/m3)

Sand (kg/m3)

Basalt stone (kg/m3)

SP(wt.%)

SP/SSDa

(%)Strengthclass

53 790 968 2.25 113 C60/7552 790 968 1.25 100 C70/8578 790 968 1.50 100 C70/85

103 790 968 1.65 94 C70/8557 790 968 3.0 109 C70/8555 790 968 1.5 92 C80/9559 790 968 2.0 100 C90/105

0 790 968 2.0 100 C90/1050 790 968 1.75 100 C90/105

Table 6Test results of self-compactability of investigated SCHPC.

Mixtures Slumpflow(mm)

T500

(s)V-funnel(s)

J-ring stepheight(mm)

Sievesegregation(%)

Aircontent(vol.%)

34LSP20R10 770 2.9 10.2 10 12.7 1.934FA20R10 780 2.9 10.5 10 12.5 1.434FA20R15 770 3.3 12.5 10 10.5 1.334FA20R20 760 3.7 16.4 10 8.3 1.430LSP20R10 730 4.8 11.0 10 8.3 2.730FA20R10 770 3.3 18.5 10 11.0 1.126FA20R10 760 6.7 19.7 9.5 4.7 1.026FA20DSF10 750 7.0 22.3 10 5.4 1.326FA20USF10 770 4.0 9.2 10 13.6 2.5

H.T. Le et al. / Materials and Design 72 (2015) 51–62 59

At W/B of 0.34, the SP demand of concrete increased with ahigher content of RHA7.7. The increase in RHA7.7 contentdecreased the filling ability (slump flow), and increased the plasticviscosity (V-funnel time) and hence segregation resistance ofSCHPC (sieve segregation). As concluded in the previous studies[32,33], RHA is a macro-mesoporous material (Fig. 9c). The partialreplacement of cement by RHA results in the increase in specificsurface area of binder and water demand of mixture due to thelarge amount of water absorbed into macro and meso pores ofthe RHA particles. With 10, 15, and 20 wt.% cement replacementsby RHA7.7, the amount of water needed to fully fill its pore volumeis about 5.5, 8.2 and 10.9 l/m3 concrete and about 3.1, 4.7 and6.2 wt.% mixing water, respectively.

At W/B of 0.26, the mixture containing RHA5.7 had a higher SPdemand than the mixture containing USF, and similar to the mix-ture containing DSF. The incorporation of RHA5.7 reduced the fill-ing ability, and significantly increased plastic viscosity and hencesegregation resistance of SCHPC. The effect of RHA5.7 is strongerthan USF and similar to DSF (Table 6). Different from RHA, USF par-ticles are spherical and dense (Fig. 9a) and hence increased fillingability and decreased viscosity of SCHPC due to ‘‘ball bearingeffect’’ and lower water demand. The difference in effect of RHAand SF has been fully explained in the previous studies [32,33].In the case of DSF, very fine SF particles were compacted as largeSF particles (Fig. 9b). The experimental results show that the mix-ture containing DSF had lower filling ability, higher plastic vis-cosity and hence higher segregation resistance compared to themixture containing USF. That indicates that the compacted parti-cles could not be separated completely during mixing. The poresformed between the fine SF particles in the large compacted SFparticles might absorb an amount of water resulting in the

R² = 0.96

1.0

1.5

2.0

2.5

3.0

1.0 1.5 2.0 2.5 3.0 3.5

SP d

eman

d fo

r SC

HPC

(wt.%

)

SSD of mortar (wt.%)

Fig. 11. SP demand of SCHPC vs. SSD of mortar formulated from the SCHPC.

reduction in free water in mixture, hence decreased the filling abil-ity and increased viscosity of mixture.

4.5. Compressive strength of SCHPC

Compressive strength results of SCHPC are presented in Figs. 12and 13. It can be seen that the 28-day compressive strengths of allSCHPC mixtures reached over 90 MPa and met the designed com-pressive strength classes. The relationship between expected andexperimental compressive strength at 28 days of all SCHPC mix-tures (Table 5) is presented in Fig. 12. These results in this studyindicate that application of the efficiency factor (k-value) and therelationship between compressive strength and W/C ratio for ordi-nary concrete (Walz curve) are suitable for estimating compressivestrength in the mix design for SCHPC.

The development of compressive strength of SCHPC is present-ed in Fig. 13. At W/B of 0.34, increasing RHA7.7 content decreasedcompressive strength of concrete at 3 and 7 days howeverincreased compressive strength at 28 days. RHA is a very reactiveporous MA. The additional C–S–H formed by the pozzolanic reac-tion of RHA refines the pore structure of the cement matrix, andimproves the interface transition zone resulting in the increase incompressive strength [6,9,41,42]. Furthermore, the internal watercuring effect of RHA can also exert a positive effect on compressivestrength at 28 days, especially at later ages. The amount of waterabsorbed in the pores of RHA particles will be released to promotefurther hydration of the cement, especially when the relativehumidity of the paste considerably decreases [41,43]. At the earlyages (3 and 7 days), the incorporation of 20 wt.% RHA7.7 led to thereduction in compressive strength. This result has also beenobtained in a previous study [41] and other studies on ultra highperformance concrete [12]. 20 wt.% cement replacements byRHA7.7 resulted in a significantly lower cement content in thepaste and hence reduced compressive strength due to the dilutingeffect. Additionally, the larger amount of water absorbed in thepores of RHA particles causes a lack of available water for cementhydration at the early ages. RHA particles with porous structure arethemselves the weakest points in the cement matrix [32]. Thismight be another negative effect of RHA on compressive strengthat the early ages, especially at high RHA content in concrete.

At W/B of 0.30 and 0.34, the compressive strength of the mix-tures containing FA is slightly higher than that of the mixtures con-taining LSP, regardless of ages, especially at the lower W/B ratio.LSP is considered as a nearly inert MA, whereas FA is a pozzolanicMA with k-value of 0.59 (Section 2). Consequently, pozzolanicactivity of FA might contribute the higher compressive strengthof the mixture containing FA. At W/B of 0.26, the compressivestrength of the mixture containing RHA was similar to that of themixture containing SF, irrespective of ages. As mentioned inSection 2, RHA is a highly reactive MA comparable with SF[8,9,32,42]. This result is also consistent with the results in a pre-vious study and of other studies [6,9,12,41,42].

4.6. Discussion

Most proportioning design methods for SCC/SCHPC are basedon empirical tests and considerably different from those for ordi-nary concrete. To compare with the proposed method, severalwell-known mix design methods for SCC/SCHPC are summarizedand analyzed in what follows:

In the method of Okamura and its modified versions [3,4,14,15],generally, the coarse aggregate volume is fixed at 50% of the solidvolume in the concrete, and the fine aggregate volume is fixed at40% of the mortar volume. The W/B ratio is first determined fromthe slump flow tests on the paste. Then the W/B ratio and SP con-tent are determined from slump flow tests on mortar. The final

R² = 0.95

80

90

100

110

120

130

80 90 100 110 120 130

Expe

cted

com

pres

sive

stre

ngth

(MPa

)

Experimental compressive strength (MPa)

Fig. 12. Expected vs. experimental compressive strength data at 28 days.

Table 7The mixture proportions in this study, by EFNARC and in reported studies.

Results inthis study

By EFNARC[4]

From reportedstudies [30]

Coarse aggregate (vol.%) 31.7 27.0–36.0 28.1–42.3Fine/total aggregate (wt.%) 45.0 48.0–55.0 38.8–52.9Paste (vol.%) 36.5 30.0–38.0 29.6–40.4Powder (kg/m3) 514–590 380–600 410–607Mixing water (kg/m3) 155–182 150–210 160–200W/B ratio (by wt./vol.) 0.26–0.34 0.28–0.36/

0.85–1.150.26–0.48

60 H.T. Le et al. / Materials and Design 72 (2015) 51–62

W/B ratio and SP content are determined by trials on concrete so asto ensure self-compactability. The procedure developed byPetersson et al. [44,45], the so called CBI method, aims findingthe maximum content of aggregate that ensures the flow of con-crete through the reinforcement without causing blockage. Thecontent of the filler, water and SP are determined on concrete testswith coaxial rheometer. More recently, the mix design methodproposed by Su et al. [17,46] is developed on the basis of the pack-ing factor of aggregate which is the mass ratio of tightly to looselypacked aggregate in SCC. Then coarse and fine aggregate contentsare calculated with the assumption of volume ratio of fine to coarseaggregate (50–57%). The cement content is determined by therequired compressive strength (0.11–0.14 MPa/kg cement). MA,e.g. FA and GGBFS, is used to obtain the high paste volume ensur-ing the required self-compacting properties. The water demand ofMA is considered. The SP content is chosen on the basis of engi-neering experience, finally determined from experimental testson concrete to meet the required self-compacting properties.

This survey of common mix design methods reveals their com-plexity for practical implementation. They use as prime para-meters the content of coarse and fine aggregate and of the SP, aswell as the W/B ratio. Hence, grading and packing of the aggregateare ignored, although well-known to influence paste content andworkability of the SCC. Viscosity of the concrete declines at tighterpacking of the aggregate because more paste will be available aslubricant between the particles for a given paste content. This issimilar to viscosity of slurry produced by solids and water [29].Next, W/B ratio is determined to meet the self-compactability

0

20

40

60

80

100

120

140

Com

pres

sive

stre

ngth

(MPa

)

3 days 7days 28days

Fig. 13. Compressive strength of investigated SCHPC.

requirements, so that compressive strength cannot be controlled.Actually, many mixtures composed by the Okamura method havea higher compressive strength than required due to a relativelyhigh paste content [47]. Su et al. [17,46] control the compressivestrength via the strength efficiency factor of cement. However, thisis a valid approach for type I PC (CEM I) only and not for a blendedcement, such as CEM III/B 42.5 [47]. For all mix design methods, SPcontent and W/B ratio are derived from experiments, i.e., by slumpflow test and/or rheometer, a procedure that is laborious,time-consuming and expensive. Furthermore, the effect of MA orfiller on rheological properties/self-compactability and compres-sive strength is not taken into consideration either.

The proposed method in this study is developed on the basis ofthe cementing efficiency of MA and the requirements for propor-tioning ordinary concrete. In doing so, all major parameters aretaken into consideration, as demonstrated herein. Further, due toits similarity with the methods for normal concrete, the presentset up is as simple and handy for practical applications. The mix-ture proportions designed by the proposed method are consistentwith the regulations for SCC/SCHPC proportioning as recommend-ed by EFNARC [4] and with reported studies on successful mixturesin the period 1993–2003, as presented in Table 7.

It is very difficult to produce SCHPC with very high compressivestrength as obtained for ultra high performance concrete becauseof the high entrapped air content at very low W/B ratio. Yet, inthe herein reported tests that were based on mixture proportion-ing by the proposed method, guaranteeing goodself-compactability, we have demonstrated that a compressivestrength level of 120 MPa can be obtained, exceeding the reportedstrength levels in earlier reports [6,13,17,30,48].

5. Conclusions

The aim of this work was to develop a new mix design methodfor SCHPC on the basis of the cementitious efficiency of mineraladmixture and the requirements for ordinary concrete proportionsregulated in DIN EN 206-1 [2] and DIN 1045-2 [26]. From theexperimental results, the following conclusions can be drawn:

(1) The strength efficiency factor of RHA was assessed.Increasing percentage replacement of cement by RHAdecreased the efficiency factor of RHA.

(2) A simple mix design method was proposed for SCHPC con-taining various mineral admixtures. It is based on thecementitious efficiency of mineral admixture and therequirements for ordinary concrete proportions regulatedin DIN EN 206-1 [2] and DIN 1045-2 [26].Self-compactability can be achieved with a few trials. Aproper compressive strength level as in ordinary concretecan be obtained for a particular W/B ratio, a given contentand type of mineral admixture.

(3) The procedure of the proposed method for SCC/SCHPC issimilar to that of ordinary concrete, so is as simple for mixdesign of SCC/SCHPC and for practical applications. The

H.T. Le et al. / Materials and Design 72 (2015) 51–62 61

packing theory of Funk and Dinger with an exponent q = 0.25was adopted to determine the grading of aggregate. The pri-mary paste volume for filling ability was computed from thevoid content of compacted aggregate. The superplasticizerdosage for the concrete was set on the basis of the super-plasticizer saturation dosage of the corresponding mortar.W/B ratio was determined on the basis of the required com-pressive strength. Efficiency factors were used to expresseffect of MAs on compressive strength of concrete.

(4) In the range of 5–20 wt.% cement replacement, RHA wasvery effective in improving compressive strength of SCHPC.The underlying design of the SCHPC could be successfullybased on the proposed value range of the efficiency factorfor RHA. i.e., 2.7–1.8, which is only marginally lower as com-pared to that of SF.

(5) Using the proposed method, SCHPC was developed withternary binders, i.e. cement and two different MAs fromRHA, SF, FA, and LSP, having good self-compactability andhigh compressive strength in the range of over 90 MPa. Itwas demonstrated that even 120 MPa is attainable at28 days with the hybrid blended mixtures of SCHPC, whilestill manifesting good self-compactability.

(6) It is possible to use common MAs (LSP, FA) in combinationwith RHA to produce SCHPC with good self-compactabilityand very high compressive strength. The combination of10 wt.% RHA and 20 wt.% FA is superior over 10 wt.% RHAand 20 wt.% LSP with respect to superplasticizer demandand compressive strength of SCHPC.

Acknowledgements

The authors would like to express thanks to Ministry ofEducation and Training of Viet Nam, F.A. Finger-Institute forBuilding Material Science (FIB) – Bauhaus University Weimar andGerman Academic Exchange Service (DAAD) for financial support.The authors are also grateful to Dr. Bui, D.D.; Dipl.-Ing. Flohr, A.;Dipl.-Ing. Ehrhardt, D.; Dipl.-Ing. Giese, A. for helpful discussions.

References

[1] Koehler E, Fowler D. Aggregate in self-consolidating concrete. In: ICAR project108. The University of Texas at Austin: International Center for AggregatesResearch; 2007. p. 362.

[2] DIN EN 206-1. Specification, performance, production and conformity. Berlin:Beuth Verlag GmbH; 2001. p. 77.

[3] Okamura H, Ouchi M. Self-compacting high performance concrete. Prog StructEng Mater 1998;1:378–83.

[4] The European guidelines for self-compacting concrete, specification,production and use. Norfolk: The Self-Compacting Concrete European ProjectGroup; 2005. p. 63.

[5] Guidelines for viscosity modifying admixtures for concrete. Norfolk: TheEuropean Federation of Specialist Construction Chemicals and ConcreteSystems (EFNARC); 2006. p. 12.

[6] Safiuddin M. Development of self-consolidating high performance concreteincorporating rice husk ash [PhD thesis]: University of Waterloo, Canada;2008.

[7] Gesoglu M, Güneyisi E, Özbay E. Properties of self-compacting concretes madewith binary, ternary, and quaternary cementitious blends of fly ash, blastfurnace slag, and silica fume. Construct Build Mater 2009;23:1847–54.

[8] Mehta PK. Rice husk ash – a unique supplementary cementing material.Advances in concrete technology. Ottawa: Center for Mineral and EnergyTechnology; 1994. p. 419–44.

[9] Bui DD. Rice husk ash as a mineral admixture for high performance concrete[PhD Thesis]. The Netherlands: Delft University of Technology; 2001.

[10] Rodriguez de Sensale G. Effect of rice-husk ash on durability of cementitiousmaterials. Cem Concr Compos 2010;32:718–25.

[11] Chao-Lung H, Anh-Tuan BL, Chun-Tsun C. Effect of rice husk ash on thestrength and durability characteristics of concrete. Construct Build Mater2011;25:3768–72.

[12] Nguyen VT, Ye G, van Breugel K, Fraaij ALA, Bui DD. The study of using ricehusk ash to produce ultra high performance concrete. Construct Build Mater.2011;25:2030–5.

[13] De Schutter G, Bartos P, Domone P, Gibbs J. Self-compactingconcrete. Caithness, Scotland, UK: Whittles Publishing; 2008.

[14] Okamura H, Ozawa K. Mix design for self-compacting concrete. Concr libr JSCE1995;25:107–20.

[15] Specification and guidelines for self-compacting concrete. Norfolk: TheEuropean Federation of Specialist Construction Chemicals and ConcreteSystems (EFNARC); 2002. p. 32.

[16] Saak A, Jennings H, Shah S. New methodology for designing self-compactingconcrete. ACI Mater J 2001;98:429–39.

[17] Su N, Hsu K-C, Chai H-W. A simple mix design method for self-compactingconcrete. Cem Concr Res 2001;31:1799–807.

[18] Kheder G, AlJadiri R. New method for proportioning self-compacting concretebased on compressive strength requirements. ACI Mater J 2010;107:490–7.

[19] Kwan A. Use of condensed silica fume for making high strength,self-consolidating concrete. Can J Civ Eng 2000;27:620–7.

[20] Chowdhury S, Basu P. New methodology to proportion self-sonsolidatingconcrete with high-volume fly ash. ACI Mater J 2010;107:222–30.

[21] Xie Y, Liu B, Yin J, Zhou S. Optimum mix parameters of high-strengthself-compacting concrete with ultrapulverized fly ash. Cem Concr Res2002;32:477–80.

[22] Dinakar P, Sethy K, Sahoo U. Design of self-compacting concrete with groundgranulated blast furnace slag. Mater Des 2013;43:161–9.

[23] Papadakis VG, Antiohos S, Tsimas S. Supplementary cementing materials inconcrete: Part II: a fundamental estimation of the efficiency factor. Cem ConcrRes 2002;32:1533–8.

[24] Papadakis VG, Tsimas S. Supplementary cementing materials in concrete: PartI: efficiency and design. Cem Concr Res 2002;32:1525–32.

[25] Wong HS, Abdul Razak H. Efficiency of calcined kaolin and silica fume ascement replacement material for strength performance. Cem Concr Res2005;35:696–702.

[26] DIN 1045-2. Concrete specification, properties, production andconformity-application rules for DIN EN 206-1. Berlin: Beuth Verlag GmbH;2008. p. 62.

[27] DIN EN 1015-3. Determination of consistence of fresh mortar (by flow table).Berlin: Beuth Verlag GmbH; 2007. p. 12.

[28] DIN EN 196-1. Determination of strength. Berlin: Beuth Verlag GmbH; 2005. p.33.

[29] Funk JE, Dinger DR. Predictive process control of crowded particulatesuspension, applied to ceramic manufacturing. New York: Kluwer AcademicPress; 1994.

[30] Domone PL. Self-compacting concrete: an analysis of 11 years of case studies.Cem Concr Compos 2006;28:197–208.

[31] Neville A, Aïtcin P-C. High performance concrete—an overview. Mater Struct1998;31:111–7.

[32] Le HT, Rößler C, Siewert K, Ludwig HM. Rice husk ash as a pozzolanic viscositymodifying admixture for self-compacting high performance mortar. In:Proceedings of the 18th international conference on building materials(Ibausil 18). Weimar, Germany: F.A. Finger-Institut für Baustoffkunde; 2012.p. 0538–45.

[33] Le HT, Siewert K, Ludwig HM. Rheological behaviour of fresh mortarformulated from self-compacting high performance concrete incorporatingrice husk ash. In: Proceedings of the 1th RILEM international conference onrheology and processing of construction materials. Paris, France. RILEMpublications S.A.R.L., 2013. p. 131–8.

[34] Rizwan SA. High performance mortars and concretes using secondary rawmaterials [PhD thesis]. Freiberg, Germany: Technischen UniversitätBergakademie Freiberg; 2006.

[35] DIN EN 12350-8. Self-compacting concrete – slump-flow test. Berlin: BeuthVerlag GmbH; 2010. p. 12.

[36] DIN EN 12350-9. Self-compacting concrete – V-funnel test. Beuth VerlagGmbH; 2010. p. 10.

[37] DIN EN 12350-12. Self-compacting concrete – J-ring test. Berlin: Beuth VerlagGmbH; 2010. p. 13.

[38] DIN EN 12350-11. Self-compacting concrete – sieve segregation test. Berlin:Beuth Verlag GmbH; 2010. p. 11.

[39] DIN EN 12390-2. Making and curing specimens for strength tests. Berlin:Beuth Verlag GmbH; 2009. p. 10.

[40] DIN EN 12390-3. Compressive strength of test specimens. Berlin: Beuth VerlagGmbH; 2009. p. 19.

[41] Le HT, Nguyen ST, Ludwig HM. A study on high performance fine-grainedconcrete containing rice husk ash. Int J Concr Struct Mater 2014;8:301–7.

[42] Nguyen VT. Rice husk ash as a mineral admixture for ultra high performanceconcrete [PhD thesis]. The Netherlands: Delft University of Technology; 2011.

[43] Nguyen VT, Ye G, van Breugel K, Copuroglu O. Hydration and microstructure ofultra high performance concrete incorporating rice husk ash. Cem Concr Res2011;41:1104–11.

[44] Petersson Ö, Billberg P, Van BK. A model for self-compacting concrete. In:Bartos PJM, Cleand DJ, editors. RILEM conference product methods workabilityconcrete. E & FN Spon; 1996. p. 483–92.

[45] Petersson Ö, Billberg P. Investigation on blocking of self-compacting concretewith different maximum aggregate size and use of viscosity agent instead offiller. In: Petersson ÅSaÖ, editor. First international RILEM symposium onself-compacting concrete. Paris: RILEM Publications SARL; 1999.

[46] Su N, Miao B. A new method for the mix design of medium strength flowingconcrete with low cement content. Cem Concr Compos 2003;25:215–22.

62 H.T. Le et al. / Materials and Design 72 (2015) 51–62

[47] Brouwers HJH, Radix HJ. Self-compacting concrete: theoretical andexperimental study. Cem Concr Res 2005;35:2116–36.

[48] Bouzoubaâ N, Lachemi M. Self-compacting concrete incorporating highvolumes of class F fly ash: preliminary results. Cem Concr Res2001;31:413–20.

[49] DIN EN 206-9. Additional rules for self-compacting concrete. Berlin: BeuthVerlag GmbH; 2010. p. 29.

[50] Domone PL. A review of the hardened mechanical properties ofself-compacting concrete. Cem Concr Compos 2007;29:1–12.

[51] Puntke W. Wasseranspruch von feinen Kornhauf-werken. Beton2002;52:242–8.

[52] Zement-Taschenbuch. 51. Auflage ed. Düsseldorf: Verein DeutscherZementwerke; 2008.