Properties of concrete with pumice powder and fly ash as cement replacement materials

Construction and Building Materials 85 (2015) 1–8

Contents lists available at ScienceDirect

Construction and Building Materials

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

Properties of concrete with pumice powder and fly ash as cementreplacement materials

http://dx.doi.org/10.1016/j.conbuildmat.2015.03.0260950-0618/� 2015 Published by Elsevier Ltd.

⇑ Corresponding author. Tel.: +90 212 3835245; fax: +90 212 3835133.E-mail address: [email protected] (N. Kabay).

Nihat Kabay ⇑, M. Mansur Tufekci, Ahmet B. Kizilkanat, Didem OktayYıldız Technical University, Department of Civil Engineering, Construction Materials Division, Davutpasa Campus, 34210 Esenler-Istanbul, Turkey

h i g h l i g h t s

� The effect of pumice powder and fly ash on concrete properties was investigated.� Pumice powder and fly ash improved physical properties of concrete.� Strength of concretes with pumice powder and fly ash were comparable to reference.� Pumice powder and fly ash contributed to sulfate resistance of concrete.� Pumice powder can be used in concrete where sulfate resistance is desired.

a r t i c l e i n f o

Article history:Received 20 June 2014Received in revised form 23 February 2015Accepted 4 March 2015Available online 28 March 2015

Keywords:Pumice powderFly ashStrength developmentPhysical propertiesMagnesium sulfate

a b s t r a c t

Turkey is rich in natural pozzolan and pumice is abundantly found in several regions of the country. Inthis study, pumice powder (PP) and fly ash (FA) were used as cement replacement materials and theeffect of partial replacement of PP, FA and their blends by cement on physical, mechanical and durabilityproperties of concrete was investigated. Test results showed both PP and FA addition resulted in lowermechanical strength at early ages, but comparable strength at later ages compared to the referenceconcrete. Replacement of cement with PP, FA and their blends resulted in concretes with decreased waterabsorption, sorptivity and void content and higher magnesium sulfate resistance compared to the refer-ence concrete. Since pumice is abundantly found in Turkey, this material might be used as an additive inconcrete applications or as a precaution against magnesium sulfate attack.

� 2015 Published by Elsevier Ltd.

1. Introduction

Environmentally friendly cement-based materials is a topic ofinterest and cement replacement materials play an important rolein the construction industry considering economical, technologicaland ecological points of view [1,2]. Therefore, the search foralternative binders or cement replacement materials has beenthe subject of many publications. Concrete materials should notonly possess good workability, excellent mechanical propertiesand durability, but also offer environmental and economic benefits[3]. Besides cost reduction and enhancement of workability offresh concrete, the use of pozzolans might help improve thedurability of concrete such as resistance to thermal cracking,alkali-aggregate expansion, and sulfate attack [4].

A pozzolana is defined as a natural or artificial material whichcontains reactive silica [5]. A more detailed definition is given in

Ref. [6] and pozzolans are defined as materials that have little orno cementitious value by themselves, however, when finelydivided and in the presence of moisture they will chemically reactwith alkalis to form cementing compounds. The silica in a poz-zolana has to be amorphous, or glassy, to be reactive [6]. Naturalpozzolans are generally derived from volcanic rocks and minerals[4]. Turkey is rich in natural pozzolan, where almost 20% of thecountry is covered by Tertiary and Quaternary-age volcanic rocks[1], and pumice can be found in several regions of the country.Pumice is a natural lightweight material of volcanic origin pro-duced by the release of gases during the solidification of lava.The cellular structure of pumice is created by the formation of bub-bles or air voids when gases contained in the molten lava flowingfrom volcanoes become trapped on cooling [7]. When lightweightaggregates of pozzolanic materials were ground to very finepowder, they could possess certain cementitious properties. Atthe same time, when they mix with a certain amount of cementand lime, their binding property increases [8]. Therefore the useof pumice powder (PP) as cement replacement material has been

Table 2Particle size distribution of aggregates.

Sieve opening (mm) Cumulative passing (%)

Fine aggregate Coarse aggregate

16 100 10011.2 100 878 100 554 100 02 75 01 50 00.5 35 00.25 15 0

Table 3Chemical and physical properties of cement, pumice powder and fly ash.

Chemical compound (%)and physical properties

Cement Pumicepowder

Flyash

ASTM C 618requirements

Naturalpozzolana

F typeflyash

SiO2 20.23 77.52 60.13 – –Al2O3 5.14 12.99 19.00 – –Fe2O3 3.87 1.5 8.98 – –SiO2 + Al2O3 + Fe2O3 29.24 92.01 88.11 Min 70% Min

70%CaO 63.14 0.1 1.90 – –MgO 1.25 0.4 4.77 – –SO3 2.89 0.52 0.95 Max 4% Max

5%Loss on ignition 1.55 5.42 1.69 Max 10% Max

6%Cl� 0.0435 0.0096 – – –Na2O/K2O 0.27/

0.900.12/0.95

– – –

Free CaO 1.2 – – – –Specific gravity 3.14 2.32 2.21 – –Specific surface (cm2/g) 3780 4400 3545 – –

68

1012141618

m io

n co

ncen

tratio

n(m

mol

/L)

PP

FAThe curve of calcium ion saturation concentration [15]

2 N. Kabay et al. / Construction and Building Materials 85 (2015) 1–8

the topic of a number of studies in the literature [1,2,7–10]. Asreported in a previous paper [11], the importance of using naturalpozzolans in the cement industry requires a complete evaluation oftheir effects on concrete and therefore there is still a need for fur-ther studies on the use of such natural pozzolans as cementreplacement materials, and their effects on concrete propertiesshould be investigated. In this study pumice is chosen as a naturalpozzolan because of its availability in Turkey and that it can beeasily grinded to obtain its powder without requiring high amountof energy.

The most common artificial pozzolana is the fly ash (FA) whichis precipitated electrostatically or mechanically from the exhaustgases of coal-fired power stations [5]. ASTM C 618 [12] classifiesFA into two groups as Class F and Class C, where Class F FA has poz-zolanic properties and Class C FA in addition to having pozzolanicproperties, also has some cementitious properties.

ASTM C 618 [12] presents chemical and physical requirementsfor FA and natural pozzolana for use in concrete. ASTM 311 [13]defines strength activity index to determine whether artificial(fly ash) or natural pozzolan results in an acceptable level ofstrength development when used with hydraulic cement in con-crete. There is also a pozzolanic activity index with lime, whichdetermines the total activity of pozzolana [5,14]. EN 196-5 [15]defines a direct test for determining the pozzolanicity of poz-zolanic cements which is also known in the literature as theFrattini test [16,17]. This test is based on chemical titration andcan accurately define the pozzolanic activity of blended Portlandcements measuring the CH consumption released during PC hydra-tion [17].

This study presents the results of the research conducted toassess the effect of partial replacement of PP, FA and their blendsby cement on physical, mechanical and durability properties ofconcrete. Therefore, seven concrete mixtures with various poz-zolana contents up to 20% were cast and tested in order to evaluatethe effect of PP, FA and their blends on standard consistency andsetting times of cement pastes and workability, void content, waterabsorption, sorptivity, compressive strength, splitting tensilestrength and magnesium sulfate resistance of concrete mixtures.It should also be noted that chemical composition and propertiesof pumice can vary place to place and might have different effectson concrete properties; therefore the results reported here repre-sents the materials with particular properties.

2. Experimental study

2.1. Materials and mix proportions

The materials used in the research consist of limestone coarse and fine aggre-gates, cement, chemical admixture, PP and FA. Physical properties and particle sizedistribution of the aggregates are presented in Tables 1 and 2, respectively. The typeof cement used throughout the study was CEM I 42.5 R. PP was obtained by initiallyoven drying the coarse pumice aggregates at around 100 �C to eliminate the freewater and then grinding them by a laboratory type disc grinder. The powder wasthen sieved from 63 l and the passing material was used throughout the research.Physical properties of pumice aggregates used in this study can be found in Ref.[18]. F type FA was also used in some mixtures as an artificial pozzolana.Chemical and physical properties of cement, PP and FA are presented in Table 3.

It can be clearly seen from the chemical analysis (Table 3) that the main com-ponent of PP and FA is SiO2, where the main component of cement is CaO. ASTM C618 [12] requires that the sum of SiO2, Al2O3 and Fe2O3 be more than 70% for natu-ral pozzolans and F type FA. Table 3 shows that both PP and FA used in this studyconfirm this and that the sum of the cementitious compounds is much above than

Table 1Physical properties of aggregates.

Aggregate Particle size (mm) Particle density (kg/dm3)

Fine aggregate 0–4 2.76Coarse aggregate 2–11.2 2.81

the limit value of 70%. Pozzolanic activity of PP and FA were also investigatedaccording to the procedure explained in EN 196-5 [15]. Test results verified thatboth materials show pozzolanic activity (Fig. 1).

Polycarboxylic ether based superplasticizer was used in all mixtures at a con-stant amount of 1.5% of the binder by weight and the slump values of concretes ran-ged between 18 and 25 cm.

Table 4 summarizes the mix proportions and fresh concrete properties. Mixingwas performed in a 45 L capacity pan mixer with a vertical rotation axis and freshconcrete properties were determined after the mixing. A total of seven concretemixtures were prepared with constant water to binder ratio and binder contentof 0.42 and 500 kg/m3, respectively. The mixtures were coded according to the poz-zolanic material addition and content, where ‘‘R’’ defines the reference concretewith no pozzolana, ‘‘P’’ and ‘‘F’’ defines PP and FA inclusion, respectively, and thenumbers 5, 10 and 20 after P and F define the substitution ratio of regarding poz-zolana by cement by weight.

024

35 40 45 50 55 60 65 70 75 80 85 90 95 100

Cal

ciu

Hydroxil ion concentration (mmol/L)

Fig. 1. Diagram for assessing pozzolanicity of PP and FA.

Table 4Mix proportions and fresh properties of concrete.

Mixture code Cement (kg) Pumicepowder (kg)

Fly ash (kg) Water (kg) Superplasticizer (kg)

Fineaggregate (kg)

Coarseaggregate (kg)

Slump (cm) Fresh density(kg/m3)

R 500 0 0 25 2472P10 450 50 0 21 2444P20 400 100 0 18 2406F10 450 0 50 210 7.5 944.1 787.9 23 2456F20 400 0 100 24 2428P5–F5 450 25 25 18 2444P10–F10 400 50 50 20 2406

N. Kabay et al. / Construction and Building Materials 85 (2015) 1–8 3

2.2. Sample preparation and testing procedure

The effect of substituting cement by PP, FA and their blends on concrete proper-ties was evaluated by several physical, mechanical and durability tests. Workabilityof concretes was assessed by slump test. Initial slump values were taken at around15 min (this time includes the duration of mixing) and the test was repeated at15 min intervals until 60 min to determine slump loss. Setting time and standardconsistence tests were also determined on cement paste specimens that containthe same amount of PP and FA as in the concrete mixtures according to EN 196-3[19]. The water to binder ratio of the pastes was adjusted for each mixture to getthe standard consistency. In order to determine the initial and final setting timeof the pastes, Vicat mould was filled with the corresponding paste and the needlewas lowered gently until it was in contact with the paste and then the moving partswere released quickly and the needle was allowed to penetrate vertically into thepaste. The scale readings were recorded at 10 min intervals and the time at whichthe distance between the needle and the base-plate is 6 ± 3 mm was considered asthe initial setting time of the cement. The time at which the needle first penetratesonly 0.5 mm into the specimen was considered as the final setting time.

Physical tests consisted of determining sorptivity, bulk dry density, waterabsorption and void content of concretes and the mechanical tests were compres-sion test and splitting tensile test. Physical and mechanical properties were deter-mined at 7, 28, 90 and 180 days. The specimens were cured in water until the testdate. Durability of concretes was evaluated by exposing the concrete specimens tomagnesium sulfate solution.

Sorptivity test was performed on two disc specimens with dimensions of Ø100/50 mm according to ASTM C 1585 [20]. The side surfaces of each specimen weresealed with paraffin and the top of the specimen not contacting with water wassealed by using a loosely attached plastic sheet. The mass of the sealed specimenwas measured and then placed in a pan with supports. The pan was filled withwater so that the water level is 1 to 3 mm above the top of the support. The massof the specimen was measured at 1, 5, 10, 20, 30, 60, 120, 180, 240, 300 and 360 minand the sorptivity was calculated according to ASTM C 1585 [20].

Bulk dry density, water absorption and void content of concretes were deter-mined according to ASTM C 642 [21] on two disc specimens with dimensions ofØ100/50 mm. The specimens were initially oven dried at 105 ± 5 �C for at least24 h, cooled to room temperature in a desiccator and weighed (A). Afterwards theywere placed in a suitable container, covered with tap water and boiled for 5 h. Thespecimens were then kept in the container to naturally cool and the soaked, boiled,surface-dried mass (C) was recorded. After immersion and boiling, the apparentmass of the specimens in water were recorded (D). Bulk dry density (Dd, kg/m3),water absorption (Wa, %) and void content (Vc, %) of the concretes were determinedby Eqs. (1), (2) and (3), respectively.

Dd ¼A

ðC � DÞ ð1Þ

Wa ¼ðC � AÞ

A� 100 ð2Þ

Table 5Standard consistency and setting times of cement pastes.

Mix code C C–P10

C–P20

C–F10

C–F20

C–P5–F5

C–P10–F10

Water content(% of binder)

26 28 31 26 26 28 29

Initial settingtime (min)

70 100 140 130 120 130 120

Final setting time(min)

180 190 240 230 210 200 220

Vc ¼ðC � AÞðC � DÞ � 100 ð3Þ

Compressive strength and splitting tensile strength were determined on three150 mm cube and Ø100/100 mm cylinder specimens according to EN 12390-3[22] and EN 12390-6 [23], respectively.

The resistance of concrete mixtures to magnesium sulfate attack was evaluatedby visual examination, mass change, and compressive strength change. After28 days of curing in water, the concrete specimens with dimensions of 70 � 70 �70 mm were divided into two groups. One group was continuously cured in waterwhile the second group was placed in containers where magnesium sulfate concen-tration was 10%. Magnesium sulfate solution in the containers was renewed every2 weeks in the first 3 months and every month for the remaining time of period.Photographs of the specimens were taken before exposure to magnesium sulfatesolution, and then after 180 and 360 days of exposure for visual examinations. Inorder to determine the mass changes, the specimens were weighed before they

were placed in sulfate solution, and further weighed at 7, 14, 21, 28, 90, 180, 300and 360 days. Compressive strength of the specimens exposed to magnesium sul-fate solution and their references that were kept in water are determined at 180and 360 days.

3. Results and discussion

3.1. Setting time and workability

The variation of standard consistency and setting times ofcement pastes with different percentage of PP, FA and their blendsare presented in Table 5. Paste specimens were coded as C, C–P10,C–P20, C–F10, C–F20, C–P5–F5 and C–P10–F10 where C representsthe reference paste whose binder is cement only, P and F repre-sents PP and FA addition, respectively, and numbers 5, 10 and 20represents the substitution ratio (%) of pozzolana by cement byweight. The results show that PP addition results in an increasein the water content to achieve the standard consistency, whichcould be due to the rough and porous structure of PP (Fig. 2).Similar results were also observed on C–P5–F5 and C–P10–F10pastes where cement was replaced by a blend of PP and FA. Onthe other hand FA addition did not have an adverse effect on theconsistency of the pastes due to its spherical shape (Fig. 2); inC–F10 and C–F20 paste mixes, standard consistency was obtainedwith the same water content as in reference paste (C).

Effect of pozzolana addition on setting times of pastes is shownin Table 5. Pozzolana addition increased both the initial and finalsetting times of the pastes, indicating a slower hydration processcompared to the reference paste. As indicated by Hossain [2], theslow hydration means low rate of heat development which is ofgreat importance in mass concrete construction. Therefore thereplacement of PP, FA and their blends with cement at specificratios might be useful in such mass concretes, besides other gen-eral use.

Table 6 presents the variation of slump values of the concretemixtures by time and Fig. 3 visually presents the slump forms ofthe mixtures. Since the superplasticizer content, binder contentand water to binder ratios were constant in the mixtures, Table 6and Fig. 3 provides information about how PP, FA and their blendsaffect the workability and cohesiveness of concrete. Fig. 3 showsthat the test yielded a true slump [24] in all mixtures and thatthe mixtures preserved their plasticity and cohesiveness even at60 min.

50

60

70

80

90

100

R P10 P20 F10 F20 P5-F5 P10-F10

Rel

ativ

e sl

ump

(%)

Concrete mixtures

15 min30 min45 min60 min

Fig. 4. Relative slump of concretes.

(a) (b)

Fig. 2. SEM image of fly ash (a) and pumice powder (b) particles.

Table 6Slump test results.

Test time (min) R P10 P20 F10 F20 P5–F5 P10–F10

15 (initial) 25 21 18 23 24 18 2030 21 19 12 20 19 15 1545 18 17 11 17 17 14 1360 17 16 10 16 15 13 12

4 N. Kabay et al. / Construction and Building Materials 85 (2015) 1–8

Reference concrete (R) possessed the highest initial slump valueof 25 cm and substitution of PP with cement resulted in a decreasein the slump values. This effect increased with the increase in thePP content where P20 mix possessed the lowest slump value of18 cm. On the other hand, only a slight decrease occurred in theslump of concretes when FA was used. This could be due to thespherical shape of the FA particles (Fig. 2). Since PP particles exhi-bit a porous structure and irregular shape (Fig. 2), this might haveresulted in a decrease in the slump values of concretes.

Fig. 4 presents the relative slump of concrete mixtures. It can benoticed that concrete mixtures P10, P5–F5 and F10 preserve 76%,

Time (min) R P10 P20

15

30

45

60

Fig. 3. Slump photographs of the concret

72% and 70% of their initial slump value at 60 min, respectively,which are slightly higher than that of the reference (68%).Highest slump losses occurred in the mixtures P20, P10–F10 andF20 as 42%, 40% and 37%, respectively, which were also higher than

F10 F20 P5-F5 P10-F10

e mixtures at 15, 30, 45 and 60 min.

5

6

7

tion

(Wa,

%) 7 days

28 days90 days

N. Kabay et al. / Construction and Building Materials 85 (2015) 1–8 5

that of the reference (32%). This result shows that the replacementof PP, FA and their blends by 20% of cement by weight has a nega-tive effect on long term (60 min) concrete workability, and thisshould be well considered when using such materials where alonger time of slump preservation is desired.

3

4

R P10 P20 F10 F20 P5-F5 P10-F10

Wat

er a

bsor

p

Concrete mixtures

180 days

Fig. 6. Water absorption of concretes.

2200

2250

2300

2350

2400

R P10 P20 F10 F20 P5-F5 P10-F10

Bul

k dr

y de

nsity

(Dd,

kg/m

3 )

Concrete mixtures

7 days28 days90 days180 days

Fig. 7. Bulk dry density of concretes.

020406080

100120140160180200

R P10 P20 F10 F20 P5-F5 P10-F10

Sorp

tivity

(x10

-4 m

m/s

0.5 )

Concrete mixtures

7 days28 days90 days180 days

Fig. 8. Sorptivity of concretes.

3.2. Physical properties

Figs. 5–7 present the void content, water absorption and bulkdry density of concretes at 7, 28, 90 and 180 days, respectively.Figs. 5 and 6 indicate that the void content and water absorptionof concretes gradually decreases with an increase in concreteage, while the bulk dry density of concretes increase with concreteage (Fig. 7). Void content of the concretes varied between 13.4%and 15.6% at 7 days, where P20 mixture possessed the lowestvalue. At this early age (7 days), both the replacement of PP andFA with cement provided slightly lower void contents comparedto the reference. This could be explained by the filling up of poresand voids by PP and FA. Void content of the mixtures continued todecrease at later ages and P10–F10, P20 and P5–F5 mixtures hadclearly lower void content values than the other concrete mixturesat 180 days as 8.2%, 8.7% and 8.8%, respectively, where this valuewas 11.2% for the reference concrete. This can be attributed tothe pozzolanic reactions which results in more dense structure inconcrete. In general, it can be noted that the effect of PP and FAreplacement on the void content of concrete was more significantat 180 days where the void content of pozzolana added concreteswere 4.1–26.5% lower than the reference concrete.

The water absorption results showed a similar trend with voidcontent, where the absorption values of concretes decreased bycuring time. Test results (Fig. 6) show that the major effect of PPand FA addition on absorption takes place at 180 days, where thelowest values were observed. At this age P10–F10 mix possessedthe lowest water absorption value as 3.4%, where the highest valuebelongs to the reference concrete as 4.7%. At 7 days, P20, F10 andF20 mixes had lower absorption values than the reference con-crete, P20 mix possessing the lowest as 5.8%.

Fig. 7 shows that the density of concretes increased with curingtime where the highest values are obtained at 180 days. At 7 daysthe density of PP and FA added concretes showed slightly lower den-sity values than the reference. At the other ages, however, densityvalues of concretes were similar and in some cases PP and FA addedconcretes possessed higher density values than the reference.

Sorptivity test results of the concrete mixtures are plotted inFig. 8, where it can be noticed that the sorptivity of concretesdecrease by curing time and the lowest values are observed at180 days. Pozzolana addition resulted in a decrease in sorptivityof concretes at 7 days and 90 days when compared to the refer-ence, where the lowest values belong to P10 at 7 days as136 � 10�4 mm/s0.5 and P20 at 90 days as 77 � 10�4 mm/s0.5. At

8

10

12

14

16

R P10 P20 F10 F20 P5-F5 P10-F10

Void

con

tent

(Vc,

%)

Concrete mixtures

7 days28 days90 days180 days

Fig. 5. Void content of concretes.

180 days, sorptivity of F10 and F20 mixtures yielded the highestvalues as 87 � 10�4 and 80 � 10�4 mm/s0.5, respectively, howeversorptivity of the other mixtures including the reference weresimilar and P10–F10 mixture possessed the lowest value as59 � 10�4 mm/s0.5.

The results of physical properties indicate that the use of bothPP and FA, in most cases, results in a denser microstructure withlower porosity and absorption characteristics and therefore isexpected to contribute to the durability of concrete by preventingthe absorption of detrimental chemical solutions. The enhance-ment in the physical properties can be attributed to the pore fillingeffect of fine PP and FA particles.

3.3. Mechanical properties

The average of compressive and splitting tensile strength ofconcretes is presented in Figs. 9 and 10, respectively, where it

4550556065707580859095

R P10 P20 F10 F20 P5-F5 P10-F10

Com

pres

sive

stre

ngth

(MPa

)

Concrete mixture

7 days

28 days

90 days

180 days

Fig. 9. Compressive strength of concretes.

8090

100110120130140150160

R P10 P20 F10 F20 P5-F5 P10-F10

Rel

ativ

e co

mpr

essi

ve st

reng

th (%

)

Concrete mixtures

7 days28 days90 days180 days

Fig. 11. Relative compressive strength development of concretes.

3.0

3.5

4.0

4.5

5.0

5.5

R P10 P20 F10 F20 P5-F5 P10-F10

Split

ting

tens

ile st

reng

th (M

Pa)

Concrete mixture

7 days28 days90 days180 days

Fig. 10. Splitting tensile strength of concretes.

6 N. Kabay et al. / Construction and Building Materials 85 (2015) 1–8

can be seen that the strength of concretes with PP, FA and theirblends possessed lower values at 7 days when compared to thereference. This could be explained by the substitution of cementby a relatively slow reacting material (PP and/or FA). Hossain [7]reported that compressive strength of cement mortars withvolcanic pumice powder replaced by cement by various contents(up to 25% by weight) resulted in a decrease at 1, 3, 7 and 28 days.The author also pointed out that the strength reduction isdecreased with the increase of age. In this study, a gradual increaseby curing time was observed for each mixture and strength valueswere comparable to the reference at later ages.

The compressive strength of the mixtures containing PP, FA andtheir blends decreased when the pozzolanic material content wasincreased at all ages, the exception was however P20 mixture,where a slight increase was observed at 180 days when comparedto P10.

Fig. 11 presents the relative compressive strength developmentof concrete mixtures by curing time. Strength development wascalculated separately for each mixture by dividing their strengthat 28, 90 and 180 days by their strength at 7 days. The results show

that the relative strength development is higher in mixtures whichcontain PP and/or FA when compared to that of the reference, at28, 90 and 180 days. At 180 days P20, F20 and F10 mixturespossessed more than 50% of their strength at 7 days as 56%, 54%and 51%, respectively, where this value was only 36% in thereference concrete. Test results showed that after possessing alower strength at 7 days (Figs. 9 and 10), the continued pozzolanicactivity of PP and FA contributed to increased strength gains atlater ages.

Although PP had a higher pozzolanic activity than FA (Fig. 1), thecompressive strength of concretes with PP were found to be lowerthan those with FA at early and later ages. This can be due to thefact that; at early ages, the pH of the Portland cement system isabout 12.5 and not enough alkalinity is available for the dissolutionof the PP and FA particles, therefore these particles might be con-sidered as relatively inert to the hydration mechanism and wouldonly contribute to the physical properties such as particle packingof the structure. F10 mix possessed the highest final compressivestrength amongst all mixes. Relatively considering PP and FA; theFA has spherical particles (Fig. 2) which contribute to ball bearingeffect that gives a good workability and a denser packing at earlyages which on hydration gives a denser microstructure with C–S–H. Thus FA gains more strength than PP at early and later ages.

3.4. Sulfate resistance

Sulfate resistance of concrete is of great importance and hasbeen the topic of many studies in the literature [1,25–36]. Mehtaand Monteiro [4] reported that due to magnesium sulfate attack,the conversion of calcium hydroxide to gypsum (Eq. (4)) is accom-panied by the simultaneous formation of magnesium hydroxide,and that in the absence of hydroxyl ions in the solution, C–S–H isno longer stable and is also attacked by the sulfate solution(Eq. (5)). A similar explanation was given by [35] and it is notedby [37] that magnesium sulfate attack on the cement matrixresults with the disintegration of the C–S–H gel to a non-cementitious M–S–H gel.

MgSO4 þ CaðOHÞ2 þ 2H2O! CaSO4 � 2H2OþMgðOHÞ2 ð4Þ

3MgSO4 þ 3CaO � 2SiO2 � 3H2Oþ 8H2O

! 3ðCaSO4 � 2H2OÞ þ 3MgðOHÞ2 þ 2SiO2 �H2O ð5Þ

In this study, magnesium sulfate resistance of concretemixtures was evaluated by visual observations, mass change mea-surements and compressive strength variation of the specimens.

3.4.1. Visual inspectionPhotographs of the specimens were taken after 0, 180 and

360 days of magnesium sulfate exposure for visual evaluations(Fig. 12). In all concrete series the first sign of magnesium sulfateeffect was the deterioration of the corners of the specimens. Theeffect was more obvious at 360 days where in some cases deteri-oration of the edges and the surfaces of the specimens werenoticed. Dehwah [35] reported that due to sulfate attack, formationof gypsum would show signs of etching of the concrete skin andthat this would ultimately lead to the exposure of aggregates.This type of deterioration was clearly observed after 360 days ofexposure on the reference specimens (R), where Fig. 12 shows thatat the edges of the reference specimen, the mortar layer wassoftened and destroyed and the aggregates were exposed. Similardegradation was also observed on other specimens, especiallywhich contain FA (F10 and F20) but the degree of the degradationwas lower than the reference. On the other hand, after 360 daysof exposure, specimens belonging to P10 and P20 mixturesperformed best, showing only a slight degradation. Visual

50

60

70

80

90

100

R P10 P20 F10 F20 P5-F5 P10-F10Rel

ativ

e co

mpr

essi

ve st

reng

th (%

)

Concrete mixture

180 days360 days

Fig. 14. Relative compressive strength of mixtures exposed to magnesium sulfate.

Time of exposure R P10 P20 F10 F20 P5-F5 P10-F10

0

180

360

Fig. 12. Visual appearance of specimens before and after exposed to magnesium sulfate.

N. Kabay et al. / Construction and Building Materials 85 (2015) 1–8 7

observations showed that when exposed to magnesium sulfateattack, the use of FA had no significant contribution to concretecompared to those with PP, however, in concrete mixtures wherepart of the cement was replaced with a blend of FA and PP(P5–F5 and P10–F10), degradation of the concretes was observedto be lower than that of the concretes only with FA (F10 and F20).

3.4.2. Mass changeFig. 13 presents the mass change of the specimens exposed to

magnesium sulfate solution, where it shows that the mass of allspecimens increase until 90 days. The increase in mass can beattributed to the filling up of pores and voids through the ingressof sulfate ions, by crystals of gypsum and/or ettringite [33,38].After 90 days, a decrease in the mass of reference concrete (R)can be noticed while the other mixtures showed none. Howeverafter 180 days mass loss is obvious in all mixtures and at 360 daysF20 mixture possessed the highest mass loss followed by R, P10and F10, while P20, P5–F5 and P10–F10 had the lowest mass lossvalues.

3.4.3. Compressive strength changeFig. 14 presents the average relative compressive strength

values of concretes exposed to magnesium sulfate solution at180 days and 360 days, respectively. Relative compressive strengthvalues were obtained by dividing the compressive strength of con-cretes exposed to magnesium sulfate solution by the compressivestrength of the specimens cured in water. It is clear from the figurethat magnesium sulfate exposure results in a decrease in the com-pressive strength of concretes and its effect increases with the

-0.6-0.4-0.20.00.20.40.60.81.01.2

0 90 180 270 360

Mas

s cha

nge

(%)

Days

R

P10

P20

F10

F20

P5-F5

P10-F10

Fig. 13. Mass change of specimens exposed to magnesium sulfate.

exposure time. At both ages, the detrimental effect of sulfate solu-tion was mostly observed on the reference concrete (R) specimens.On the other hand pozzolana addition provided a significant con-tribution, and P20 and P10–F10 mixtures showed the loweststrength losses as 23% and 22% at 360 days, respectively, wherethis value was 44% in the reference specimens. The literaturepresents conflicting results, where some report that pozzolanaaddition increases the magnesium sulfate resistance of concrete[25,30–32,34,36,39,40], and others report the opposite [9,26,35,37]. In this study, concrete mixtures with the addition of PP,FA and their blends performed well against magnesium sulfateattack and the increase in the replacement ratio of pozzolans from10% to 20% contributed to the relative compressive strength of con-cretes after 180 and 360 days of exposure to magnesium sulfate.

The overall results obtained from magnesium sulfate exposureshowed that the replacement of Portland cement with pozzolanicmaterials had beneficial effects on concrete. This enhancementcan be attributed to the reduction in the cement content. Morespecifically, the replacement of part of the cement by a pozzolanresults in a reduction in the C3A content and therefore thealuminate-bearing phases reduce and the formation of ettringiteis mitigated. The secondary C–S–H formed by the pozzolanic reac-tions create a film on the alumina-rich and other reactive phasesthereby hindering the formation of ettringite which also resultsin the densification of the hardened cement paste since it is depos-ited in the pores thereby making blended cements impermeableand, therefore, sulfate ions cannot easily diffuse through the con-crete matrix [41].

8 N. Kabay et al. / Construction and Building Materials 85 (2015) 1–8

4. Conclusions

This study reports the test results on the use of PP, FA and theirblends in concrete as cement replacement materials. The mainconclusions of the study may be summarized as follows:

1. The addition of PP decreased the slump value of concrete, whileFA addition had no significant effect. On the other hand theinclusion of PP, FA and their blends, in some cases (at 10%replacement ratio), contributed to the preservation of slumpvalues of concrete up to 60 min.

2. Replacement of PP, FA and their blends with cement resulted inconcrete with lower water absorption, sorptivity and voidcontent values compared to the reference.

3. The addition of PP, FA and their blends in replacement withcement, as expected, decreased the early age compressive andsplitting tensile strength of concretes, however at later ages(28, 90 and 180 days) the strength values were comparablewith that of the reference concrete.

4. Magnesium sulfate resistance of all mixtures containing PP, FAand their blends were higher than that of the reference concreteafter 180 and 360 days of exposure considering the compressivestrength change. P20 and P10–F10 mixtures turned out to bethe ones performing best against magnesium sulfate attackafter 360 days of exposure to magnesium sulfate.

5. In the scope of this study, PP addition provided contributions toconcrete properties considering physical, mechanical and dura-bility test results. Since pumice is abundantly found in manyparts of Turkey and can easily be grinded to obtain its powder,Turkish cement and concrete industry might consider using thismaterial as additive in general concrete applications or as aprecaution against magnesium sulfate attack.

Acknowledgements

This research was carried out in the Faculty of Civil Engineeringat Yıldız Technical University. The authors would like to acknowl-edge the laboratory staff for their helps.

References

[1] Binici H, Kapur S, Arocena J, Kaplan H. The sulphate resistance of cementscontaining red brick dust and ground basaltic pumice with sub-microscopicevidence of intra-pore gypsum and ettringite as strengtheners. Cem ConcrComp 2012;34:279–87.

[2] Hossain KMA. Blended cement using volcanic ash and pumice. Cem Concr Res2003;33:1601–5.

[3] Long G, Wang X, Xie Y. Very-high-performance concrete with ultrafinepowders. Cem Concr Res 2002;32:601–5.

[4] Mehta PK, Monteiro PJM. Concrete microstructure, properties and materials.3rd ed. New York: McGraw Hill; 2006.

[5] Neville AM. Properties of concrete, Fourth and Final edition. Pearson EducationLtd.; 2000.

[6] Newman J, Choo BS. Advanced concrete technology. Oxford: An imprint ofElsevier; 2003.

[7] Hossain KMA. Properties of volcanic pumice based cement and lightweightconcrete. Cem Concr Res 2004;34:283–91.

[8] Sahin S, Orung I, Okuroglu M, Karadutlu Y. Properties of prefabricated buildingmaterials produced from ground pumice aggregate and binders. Constr BuildMater 2008;22:989–92.

[9] Hossain KMA, Lachemi M. Performance of volcanic ash and pumice basedblended cement concrete in mixed sulfate environment. Cem Concr Res2006;36:1123–33.

[10] Binici H, Aksogan O, Cagatay IH, Tokyay M, Emsen E. The effect of particle sizedistribution on the properties of blended cements incorporating GGBFS andnatural pozzolan (NP). Powder Technol 2007;177:140–7.

[11] Rodrıguez-Camacho RE, Uribe-Afif R. Importance of using the naturalpozzolans on concrete durability. Cem Concr Res 2002;32:1851–8.

[12] ASTM C618. Standard specification for coal fly ash and raw or calcined naturalpozzolan for used in concrete; 2012.

[13] ASTM 311. Standard test methods for sampling and testing fly ash or naturalpozzolans for use in Portland-cement concrete; 2011.

[14] Turkish Standard, TS 25 Natural pozzolan (Trass) for use in cement andconcrete-Definitions, requirements and conformity criteria; 2008.

[15] European Standard, EN 196-5. Methods for testing cement-Part 5:puzzolanicity test for pozzolanic cement; 2011.

[16] Donatello S, Tyrer M, Cheeseman CR. Comparison of test methods to assesspozzolanic activity. Cem Concr Comp 2010;32:121–7.

[17] Tironi A, Trezza MA, Scian AN, Irassar EF. Assessment of pozzolanic activity ofdifferent calcined clays. Cem Concr Comp 2013;37:319–27.

[18] Kabay N, Aköz F. Effect of prewetting methods on some fresh and hardenedproperties of concrete with pumice aggregate. Constr Build Mater 2012;34:503–7.

[19] European Standard, EN 196-3. Methods for testing cement-Part 3:determination of setting time and soundness; 2005.

[20] ASTM C1585. Standard test method for measurement of rate of absorption ofwater by hydraulic-cement concretes; 2004.

[21] ASTM C642. Standard test method for density, absorption, and voids inhardened concrete; 2006.

[22] European Standard, EN 12390-3. Testing hardened concrete-Part 3:compressive strength of test specimen; 2009.

[23] European Standard, EN 12390-6. Testing hardened concrete-Part 3: tensilesplitting strength of test specimen; 2009.

[24] European Standard, EN 12350-2. Testing fresh concrete-Part 2: slump test;2010.

[25] Sezer G_I. Compressive strength and sulfate resistance of limestone and/orsilica fume mortars. Constr Build Mater 2012;26:613–8.

[26] Aye T, Oguchi CT. Resistance of plain and blended cement mortars exposed tosevere sulfate attacks. Constr Build Mater 2011;25:2988–96.

[27] Sahmaran M, Erdem TK, Yaman IO. Sulfate resistance of plain and blendedcements exposed to wetting–drying and heating–cooling environments.Constr Build Mater 2007;21:1771–8.

[28] Bassuoni MT, Nehdi ML. Durability of self-consolidating concrete to sulfateattack under combined cyclic environments and flexural loading. Cem ConcrRes 2009;39:206–26.

[29] Siad H, Kamali-Bernard S, Mesbah HA, Escadeillas G, Mouli M, Khelafi H.Characterization of the degradation of self-compacting concretes in sodiumsulfate environment: influence of different mineral admixtures. Constr BuildMater 2013;47:1188–200.

[30] Al-Dulaijan SU. Sulfate resistance of plain and blended cements exposed tomagnesium sulfate solutions. Constr Build Mater 2007;21:1792–802.

[31] Skaropoulou A, Sotiriadis K, Kakali G, Tsivilis S. Use of mineral admixtures toimprove the resistance of limestone cement concrete against thaumasite formof sulfate attack. Cem Concr Comp 2013;37:267–75.

[32] Shannag MJ, Shaia HA. Sulfate resistance of high-performance concrete. CemConcr Comp 2003;25:363–9.

[33] Al-Amoudi OSB. Performance of 15 reinforced concrete mixtures inmagnesium-sodium sulphate environments. Constr Build Mater 1995;9:149–58.

[34] Behfarnia K, Farshadfar O. The effects of pozzolanic binders and polypropylenefibers on durability of SCC to magnesium sulfate attack. Constr Build Mater2013;38:64–71.

[35] Dehwah HAF. Effect of sulfate concentration and associated cation type onconcrete deterioration and morphological changes in cement hydrates. ConstrBuild Mater 2007;21:29–39.

[36] Binici H, Aksogan O. Sulfate resistance of plain and blended cement. Cem ConcrComp 2006;28:39–46.

[37] Lee ST, Moon HY, Hooton RD, Kim JP. Effect of solution concentrations andreplacement levels of metakaolin on the resistance of mortars exposed tomagnesium sulfate solutions. Cem Concr Res 2005;35:1314–23.

[38] Page CL, Page MM. Durability of concrete and cement composites. CambridgeEngland: CRC Press; 2007.

[39] Uysal M, Sumer M. Performance of self-compacting concrete containingdifferent mineral admixtures. Constr Build Mater 2011;25:4112–20.

[40] Sumer M. Compressive strength and sulfate resistance properties of concretescontaining Class F and Class C fly ashes. Constr Build Mater 2012;34:531–6.

[41] Al-Dulaijan SU, Maslehuddin M, Al-Zahrani MM, Sharif AM, Shameem M,Ibrahim M. Sulfate resistance of plain and blended cements exposed to varyingconcentrations of sodium sulfate. Cem Concr Comp 2003;25:429–37.