Hydration and Characteristics of Metakaolin Pozzolanic Cement Pastes

HBRC Journal (2016) xxx, xxx–xxx

Housing and Building National Research Center HBRC Journal

http://ees.elsevier.com/hbrcj

Hydration and characteristics of metakaolin pozzolanic cement pastes

Hamdy El-Diadamony a, Ahmed A. Amer a,*, Tarek M. Sokkary b, Samir El-Hoseny a a Faculty of Science, Zagazig University, Zagazig, Egypt b Housing and Building National Research Center, Cairo, Egypt

Received 20 April 2015; revised 6 May 2015; accepted 25 May 2015

KEYWORDS Abstract The industrial area produces lots of solid waste materials with CO2 emission. One of the Supplementary cementing most effective ways to solve these problems is the utilization of these waste materials. The produc- materials; tion process of cements from its raw materials produces a lot of CO2. The most effective way to Cement hydration; decrease CO2 emission of cement industry is the substitution of a proportion of cement with sup- OPC; plementary cementing materials. Cement blended with metakaolin (MK) is also required as a coun-

Metakaolin termeasure to reduce the amount of CO2 generation. Metakaolin (MK), Al2Si2O7, is a highly amorphous dehydration product of kaolinite, Al2(OH)4Si2O5. The aim of our research was to investigate the effect of up to 20 wt% substitutions of OPC by MK on the hydration characteristics of MK-blended cement pastes. The physico-chemical properties of the hardened cement pastes were studied up to 90 days of hydration. The hydration products of some selected samples were investi- gated using XRD, DTA and DTG techniques. The results indicated that substitution of up to 20 wt% OPC by MK as pozzolanic materials resulted in an increase in the standard water of consistency, acceleration of the initial setting times, high compressive strength values at earlier ages and improvement of the mechanical and durability properties. ª 2015 Production and hosting by Elsevier B.V. on behalf of Housing and Building National Research Center. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/ by-nc-nd/4.0/).

Introduction

Cement industry produces the 7% of the global CO2 emission [1]. Researchers investigate the opportunities of how to * Corresponding author. decrease this level. The application of different supplementary Peer review under responsibility of Housing and Building National materials can be the proper solution for this problem. With Research Center. the development of industry, more and more by-products or wastes have been generated, causing serious environmental pol- lution problems. To solve this problem, a way must be found to Production and hosting by Elsevier consume or decrease such wastes. It has been discovered that http://dx.doi.org/10.1016/j.hbrcj.2015.05.005 1687-4048 ª 2015 Production and hosting by Elsevier B.V. on behalf of Housing and Building National Research Center. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Please cite this article in press as: H. El-Diadamony et al., Hydration and characteristics of metakaolin pozzolanic cement pastes, HBRC Journal (2016), http:// dx.doi.org/10.1016/j.hbrcj.2015.05.005 2 H. El-Diadamony et al. many industrial wastes can be recycled as a substitute for a constant W/B ratio 0.3. MK mixture with cement (replacement) for cement or aggregate in concrete [2]. replacement of 5, 10 and 15 wt% was prepared. The results Supplementary cementitious materials (SCMs) are now showed that 10 wt% replacement level was the optimum level commonly used to reduce the clinker factor of cement. These of MK content. materials can improve concrete properties such as compressive The objectives of this study were to investigate the substitu- strength, durability and impermeability through hydraulic or tion of OPC by MK up to 20 wt% on the hydration character- pozzolanic activity. The main component of SCMs additive istics of MK-blended cement pastes. The physico-chemical is usually an active amorphous SiO2. The availability of com- properties of cement pastes were determined up to 90 days. mon used industrial by-products such as fly ash, blast furnace The hydration products of some selected samples were investi- slag, rice husk ash, silica fume, and metakaolin. Kaolinitic gated by using XRD, DTA and DTG techniques. clays are widely available in the earth’s crust, and a heat treat- ment between 600 and 800 C of such clays leads to the dehy- Materials and methods of investigation droxylation of the crystalline structure of kaolinite to give metakaolin [3–5]. The materials used in this study are ordinary Portland cement Metakaolin (MK), Al2Si2O7, is a largely amorphous dehy- (OPC) and kaolinite clay. OPC was provided by Suez Cement dration product of kaolinite, Al2(OH)4Si2O5, which exhibits Company, Suez plant, El-Ain El-Sokhna, and the kaolinite strong pozzolanic activity [6–9]. MK is processed from kaolin clay was derived from Ras Abu Zneima Zone, South of Sinai, clay by calcination at moderate temperature (650–800 C). At Egypt. higher temperatures (>900 C), the metakaolin undergoes Metakaolin (MK) is a product from dehydroxylation of a further reactions to form crystalline compounds, the clay mineral, kaolinite, which is very fine powder prepared end-products being free silica and mullite. It contains silica by firing in a muffle furnace from room temperature up to and alumina in an active form which will react with CH. The 800 C for 2 h. The ground MK passed through 90 lm B.S. principal reasons for the use of clay-based pozzolans in mortar sieve. The chemical and physical properties of starting materi- and concrete have been materials availability and durability als are shown in Table 1. enhancement. In addition depending on the calcining tempera- The mineralogical composition of MK is seen from XRD ture and clay type, it is also possible to obtain enhancement in pattern in Fig. 1. It shows the presence of quartz as the main strength, particularly during the early stages of curing. The very mineral and amorphous alumino-silicate phase, and the amor- early strength enhancement is due to a combination of the filler phous phase is formed as a result of reactions between SiO effect and accelerated cement hydration [10]. Subsequently, 2 and Al2O3 as well as fluxing oxides impurities oxides at high these effects are enhanced by the pozzolanic reaction between temperature. The mix composition of MK-pozzolanic cement MK and the CH produced by the hydration of the cement. is shown in Table 2. The reactivity of MK has been linked to its content of penta- Mixes were prepared by substitution of OPC with 5, 10, 15 coordinated aluminum ions that are formed during the dehy- and 20 wt% of MK. The dry constituents of each mix were droxylation process [9,11]. The pozzolanic activation of MK mechanically mixed for one hour in a porcelain ball mill using by various activators (calcium hydroxide (CH), sulfates as well four balls to attain complete homogeneity, then kept in airtight as alkali hydroxide) and the properties of these binders have containers for further investigation. The standard water of con- been previously reported [2]. The principal reaction between sistency as well as initial and final setting times was determined MK and CH was derived from cement hydration, in the presence of water. This reaction forms additional cementitious CSH gel, together with crystalline products, which include calcium aluminate hydrate and alumino-silicate hydrates Table 1 Characteristics of OPC and metakaolin. (C2ASH8,C4AH13 and C3AH6). The crystalline products depend principally on the MK/CH ratio and reaction tempera- OPC MK ture [4,13]. This reaction, which is even slower than the hydra- (a) Oxide composition, % tion of plain Portland cement improves the binding properties CaO 62.72 0.28 of blended cements [11]. SiO2 20.68 55.10 Production of concrete with the incorporation of industrial Al2O3 4.90 34.10 Fe O 3.35 5.24 waste not only provides an effective way to protect the envi- 2 3 MgO 2.64 0.25 ronment, but also leads to better performance annually, and SO3 2.65 0.01 it is possible to completely consume most of the industrial Na2O 0.11 0.10 waste in the world, provided that suitable techniques for indi- K2O 0.14 0.02 vidual waste incorporation are available [14]. Many research- TiO2 0.12 2.00 ers have shown a lot of interest in MK as it has been found P2O5 0.10 1.00 to possess both pozzolanic and microfiller characteristics [6]. L.O.I 2.73 1.50 The replacement with 30 wt% of MK leads to substantial (b) Blaine surface area (cm2/g) 3400 11,000 improvement in strength and transport properties of blended concrete when compared to that of unblended concrete [15]. (c) Residue on sieve (%): 90 lm 0.00 1.00 Inclusion of MK as partial replacement of cement enhanced 45 lm 12.00 12.80 the compressive strength of concrete, but the optimum replace- 3 ment level of OPC by MK was about 20 wt% [14]. Dinakar (d) Specific gravity (g/cm ) 3.15 2.65 et al. [16] studied the effect of incorporating MK on the (e) Insoluble residue (%) HCl/Na2CO3 0.45 63.45 mechanical and durability properties of high strength concrete

1 50 1= Quartz

Intensity 1 20

1 10 1

10 20 30 40 50 2-Theta Scale

Fig. 1 XRD pattern of MK.

31.5 Table 2 Mix composition of different OPC–MK pozzolanic Is w/c Fs 220 cement blends, wt%. 31.0 Fs Mix. no. OPC M1 M2 M3 M4 200 OPC 100 95 90 85 80 MK 00 05 10 15 20 30.5 w/c 180

30.0 Is 160 according to ASTM methods using Vicat apparatus [17]. The Water consistency, % mixing of OPC and blended cement dry mixtures was carried 29.5 140 out with the required water of consistency and the fresh pastes Intial and Final setting time, min were first cured with their molds at 100% relative humidity for 24 h as described in a previous work [18]. After 24 h, the spec- 29.0 120 imens were demolded and cured under water until the desired 0 5 10 15 20 Metakaolin content, wt% curing times of 3, 7, 28 and 90 days. The hydration reaction of the pastes was stopped using an acetone–methanol mixture Fig. 2 Water of consistency, initial and final setting times of [19]. The powdered sample was dried at 70 C for 2 h, and kept OPC and OPC–MK pozzolanic cement pastes. in airtight containers for further investigation. The rate of hydration was followed by the determination of free lime [20]. The combined water content was determined from the ignition the initial and final setting times of cement pastes due to coat- loss of the dried paste at 900 C for 20 min. The bulk density, ing effect of MK particles on the cement grains as well as the compressive strength and the total porosity were determined formation of ettringite and the dilution of OPC. As MK as described elsewhere [21]. The hydration products of some increases (15–20 wt%) the initial and final setting times are selected pastes were characterized using XRD, DTA and accelerated due to the slight decrease of water of consistency DTG techniques. of cement pastes and filling effect of MK. Generally, setting of binary MK–OPC pastes does not show consistent changes Results and discussion with increase of MK content whereas 15–20 wt% of MK leads to shorten the setting times. So, the increase of MK levels was Water of consistency and setting times not proportional to replacement level [23].

The water of consistency, and initial and final setting times of Chemically combined water content the neat OPC and OPC–MK blended cement pastes are represented in Fig. 2. Water of consistency of the net OPC cement The combined water contents of OPC and OPC–MK poz- paste was 30%, then increased with increasing MK content up zolanic cement pastes are graphically plotted as a function of 10 wt% and decreased at 20 wt% MK. The substitution of 5 curing time in Fig. 3. The combined water contents increase and 10 wt% of OPC by MK increases the water of consistency gradually with the increase of curing time for all cement pastes of MK-pozzolanic cement pastes up to 30% and 50% respec- as a result of the progressive hydration of anhydrous cement tively. This is due to the high reactivity of MK, very high speci- phases as well as pozzolanic reaction and formation of fic surface area and its amorphous structure. Therefore, it increasing amounts of hydration products. It is clear that the needs more water of consistency [22]. The increase of MK con- chemically combined water content increases with increasing tent up to 20 wt% decreases the water of consistency from MK content up to 10%. This is mainly due the increase of 30.50% to 29.50%. This is due to the dilution effect of OPC water of consistency of cement pastes and very high specific by larger amounts of MK; the excessive amounts of MK act surface area of MK [24] which reacts with CH liberated from as a filler. OPC replacement by MK up to 20 wt% elongates hydration of OPC with the formation of additional hydration

18 XRD analysis 8.5 17 8.0 Combined water Fig. 4 shows the XRD-patterns of hydrated OPC paste after 16 ------Free lime 7.5 curing for 3, 28 and 90 days. The patterns show that the

7.0 remaining parts of clinker minerals, namely b-C2S and C3S, 15 are still existing up to 90 days of hydration. This is indicated 6.5 by the decrease of the relative intensities of their peaks with 14 6.0 increasing age of the hydration up to 90 days as a result of 13 hydration of OPC phases. Also, the peaks of portlandite OPC 5.5 Free lime contents, % M1 (CH) phase increase with the increase of curing time due to

Combined water contents, % 5.0 12 M2 M3 the hydration of the silicate phases of OPC clinker with the M4 4.5 formation of CSH and liberation of CH. The carboaluminate 11 1 10 100 hydrates can be formed and increase with time. Calcium Curing time, days carboaluminate is mainly formed by the reaction of atmo- spheric CO2 with the CAH as a result of hydration of Fig. 3 Chemically combined water as well as free lime contents MK-pozzolanic cement [28]. versus age of hydration of OPC and OPC–MK pozzolanic cement XRD patterns of hydrated cement pastes of M1 (5 wt% pastes. MK) cured at 3, 28 and 90 days are shown in Fig. 5. The patterns show that anhydrous phases of clinker minerals b-C2S and C3S are still existing up to 90 days; their peak intensities decrease with curing time up to 90 days. Also, the intensities products, mainly as calcium aluminosilicate hydrates (gehlen- of the peaks of portlandite increase with increase of curing ite hydrate C2ASH8), calcium aluminate hydrate (C4AH13) time due to the hydration of these silicate minerals with the and calcium silicate hydrate (CSH) in the pozzolanic cement formation of CSH and liberation of portlandite. pastes which have higher water contents than those of CSH Fig. 6 illustrates the XRD-patterns of OPC and OPC–MK in the neat OPC paste [25]. pastes containing 5, 15 and 20 wt% of MK (M1, M3 and M4) As the amount of MK increases to 15–20 wt% the com- up to 28 days of hydration. The patterns show that the inten- bined water content decreases. This is mainly attributed to sities of portlandite peaks of OPC paste are higher than those the dilution of OPC which has higher rate of hydration than of set of Pozzolanic cement pastes due to the pozzolanic reac- MK as well as low water of consistency that decreases the rate tion of MK with CH. The intensity of calcite (CaCO3) peak of hydration. 10–15 wt% MK substitution acts as a nucleating decreases with increasing MK contents; this is an indication agent which accelerates the rate of hydration of cement paste. for the consumption of CH by MK. The intensity of C3S and b-C2S minerals decreases with the increase of curing age Free lime content due to the progress of hydration process. The peak characteristic for CaCO3 is mainly overlapped with that of CSH gel. The The free lime contents of OPC and OPC–MK pozzolanic quartz peaks are still present in OPC–MK blended cement cement pastes are graphically represented as a function of pastes (M1, M3 and M4) due to its unhydraulic properties. hydration age in Fig. 3. The free lime content of OPC paste The peak intensity of CaCO3 decreases with curing time due increases with the increase of curing time up to 90 days; this to the effect of CO2 in the presence of H2O to form Ca(HCO3)2. is due to the continuous hydration of C3S and b-C2S which lib- erate free lime. On the other hand, the free lime contents of all MK-pozzolanic cement pastes increase with increasing hydra- Thermal analysis tion time up to 7 days, then decrease up to 90 days. The initial increase of free lime of pozzolanic cement pastes is mainly due The most widely used thermal method of analysis is differen- to its liberation during the hydration of OPC and then the tial thermal analysis (DTA). Fig. 7 illustrates the DTA/DTG decrease after 7 days is due to its consumption by the MK. thermograms of the hydrated OPC paste after curing up to There are two different processes, the ﬁrst tending to increase 90 days. The DTA thermograms show the occurrence of four Ca(OH)2 content and the other tending to decrease it. At early endothermic peaks at 100, 150, 420–450, 700–750 C. The ages of hydration the rate of liberation of CH by OPC cement endothermic peaks located below 200 C are mainly due to pastes exceeds the rate of consumption by aluminosilicate the dehydration of interlayer water of CSH and CAH or materials of MK; therefore, the CH content increases up to CASH, whereas the endothermic peak appeared at 7 days. After 7 days up to 90 days the rate of liberation of 420–450 C is due to the decomposition of Ca(OH)2 [26]. CH by OPC cement pastes is less than the rate of CH con- The last endothermic peak located at 700–750 C is due to sumption by MK and thereby, the CH content decreases. Also, the decomposition of CaCO3 [27]. It is clear that the intensities the CH contents of all OPC–MK cement pastes are less than of the endothermic peaks characteristic for Ca(OH)2 and those of OPC at all hydration ages due to the decrease of the CaCO3 increase with the increase of curing time up to 90 days amounts of OPC minerals which are the source of CH. It is due to progress of hydration of OPC pastes. Also, the clear that, the residual CH content of OPC–MK pozzolanic endothermic peaks of CSH and CASH increase with curing cement pastes is mainly due to the pozzolanic activity of MK. time due to progress of hydration.

1= Portlandite 2= C S 1 1 3 OPC - 3 days 3= B-C2S 4= Calcite OPC - 28 days 5= Cc OPC - 90 days

4 2+3 1 90 days 1 5 2+3

4 2+3 28 days 1 1 1

Intensity 2+3 1 5

2+3 1 3 days 1 4 2+3 1 1

10 20 30 40 50 2-Theta - Scale

Fig. 4 XRD-patterns of hardened OPC paste cured at 3, 28 and 90 days.

3 1 M1- 3 days 1= Portlandite 2= C3S 1 2 M1-28 days 3= B-C S 4= Calcite 2 M1-90 days 5= Cc 4 2 2 90 days 3 5 3 1 3 2 4 1 1 28 days 2 5 2 1 3 3 3 Intensity 4 1 2 1 3 days

2 1 2 3

10 20 30 40 50 2-Theta-Scale

Fig. 5 XRD patterns of OPC–MK pozzolanic cement pastes of M1 cured at 3, 28 and 90 days.

2 1 OPC - 28 days 1= Portlandite 2= C3S 3 M1- 28 days 3= B-C S 4= Calcite 1 2 M3- 28 days 5= Cc 6= Quartz M4- 28 days 6 4 6 2+3 1 M4

4 1 6 2+3 M3 1 1 6

4 1 1 1 2+3 Intensity M1 5

3 2

4 1 1 1 2+3 OPC 5

10 20 30 40 50 2-Theta-Scale

Fig. 6 XRD-patterns of OPC and hardened OPC–MK pozzolanic cement pastes of M1, M3 and M4 cured at 28 days.

100 OPC 3 days 98 90 d (DTG) OPC 28 days OPC 90 days 96 28 d 94

Exo 3d 92 90 88

T 86 3 d 84 82 Weight loss, %

Δ 28 80 78 Endo (DTA) 90 d 76 74

0 200 400 600 800 1000 Temperature,oC

Fig. 7 DTA/DTG thermograms of the hydrated OPC pastes cured to 90 days.

The hydration progress can be studied from thermal gravi- of the amorphous part of calcium silicate hydrates (CSH). The metric analysis (TGA). The weight loss of the hydrated phases second endothermic peak observed at about 150–160 C repre- without Ca(OH)2 and CaCO3 increases with curing time. The sents the decomposition of crystalline part of CSH, CAH, weight loss increases with slow rate up to 3 days, followed by CASH as well as the presence of carboaluminate and gehlenite higher rate up to 90 days. The TG losses characteristic for the hydrates [28]. The third endotherm located at about 450 C hydration products at low temperatures (up to 200 C) are represents the dehydration of calcium hydroxide (CH) [5,26]. 7.796%, 9.668% and 11.78% after 3, 28 and 90 days, The last endothermic peak located at 710–715 C is due to respectively. This means the degree hydration increases with decomposition of CaCO3 [26,27]. It is clear that the second increasing time of curing. On the other hand, the TG losses endothermic peak located at about 155 C increases with cur- characteristic for CH are 4.451%, 5.718% and 6.804% after ing time due to the formation of excessive amounts of CAH 3, 28 and 90 days, respectively. The increase of TG loss of and CASH on the expense of CSH. These phases are formed portlandite with curing time is due to the continuous liberation as a result of the pozzolanic reaction of MK with CH. The of CH as a result of hydration of OPC. The TG losses due to main features of the thermograms are characterized by the rel- the decomposition of CaCO3 are 4.994%, 6.883% and 5.678% ative increase of the peak areas characteristic for CH, CAH, after 3, 7 and 28 days, respectively. Also, the TG losses due to CASH and CSH phases as the hydration time increases up CaCO3 decrease with curing time; this is attributed to the reac- to 90 days; this is due to that the small amount of MK acts tion of CO2 and the moisture with CaCO3 with the formation as nucleation agent which enhances the hydration process. of Ca(HCO3)2. The endothermic peak located at 709–713 C is mainly due The DTA/DTG thermograms of hydrated OPC–MK poz- to the decomposition of CaCO3. zolanic cement paste made of mix M1 and cured up to 90 days The weight losses of hydrated phases up to 200 C are are shown in Fig. 8. Evidently, there are four endothermic 8.248%, 11.650%, 10.620% for OPC–MK blend of mix M1 peaks. The ﬁrst endothermic peak located below 100 Cis hydrated for 3, 28 and 90 days, respectively. The TG losses of mainly due to the removal of free water and the decomposition hydrated cement paste made of OPC–MK blend of composition

M1 3 days M1 28 days 100 M1 90 days 98 96 94 (DTG) 92 3 d 28 d 90 T Exo Δ 90 d 88 3 d 86 84

28 d 82 Weight loss, %

Endo (DTA) 80 90 d 78 76

0 200 400 600 800 1000 Temperature, oC

Fig. 8 DTA/DTG thermograms of hydrated metakaolin pozzolanic cement pastes (M1) cured up to 90 days.

M1 are 4.846%, 5.884% and 7.165% for 3, 28 and 90 days, graphically represented as a function of curing time in respectively; these are due to the dehydration of portlandite. Fig. 10. The bulk density of OPC–MK pozzolanic cement This indicates that 5 wt% MK acts as nucleating agent which pastes depends on the value of degree of hydration. As the accelerates the rate of hydration of OPC. The losses due to the hydration progresses, the hydration products fill a part of decomposition of CaCO3 are 4.465%, 5.188% and 5.553% for the pore volume because the volume of the hydration products cement pastes hydrated for 3, 28 and 90 days, respectively. is more twice than that of the anhydrous cement; this increases The TG losses of CaCO3 increase with curing time, and this is the bulk density of the hardened pastes [28]. The bulk density due to the continuous carbonation of portlandite. of all cement pastes increases with curing time. As the amount The DTA/DTG thermograms obtained for OPC and of MK increases the bulk density decreases. This is mainly due OPC–MK mixes having MK contents of 5, 15, 20 wt% to the increase of the amount of water of consistency with MK (M1, M2 and M3) pozzolanic cement pastes cured for 28 days content in addition to the decrease of specific gravity of MK are seen in Fig. 9. This figure shows the same four endothermic (2.65 g/cm3) in comparison with 3.15 g/cm3 of OPC (Table 1); peaks appeared in the thermograms obtained for the the initial water/cement ratio plays an important role in the OPC–MK made of M1. The peak areas appeared in the ther- values of bulk density. This is also due to that the rate of mograms obtained for OPC–MK pozzolanic cement pastes hydration of MK is lower than that of OPC especially at the having MK contents of 5–15%, by weight, are nearly identical early ages of hydration. The values of bulk density of poz- to those of the neat OPC pastes and then decrease at 20 wt% zolanic cement pastes are lower than those of OPC pastes, at MK substitution due to the dilution of OPC as well as the all curing times. In addition, the CSH formed from the poz- pozzolanic reaction of MK with portlandite. zolana reaction has low density in comparison with that The TG losses of the CSH, CAH and CASH are 9.668%, formed from the hydration of OPC. Therefore, the pozzolana 11.65%, 11.65% and 10.95%, respectively; this means that cement pastes show lower values of bulk density [24]. substitution of 5–15 wt% MK acts as microfiller and nucleating agent which accelerates the rate of hydration and thereby Total porosity the combined water increases. This is in agreement with the results of DTA. The weight loss of M1 cement paste due to The total porosity of OPC and OPC–MK pozzolanic cement the dehydration of CH is higher than that of OPC, M3 and pastes is also graphically represented as a function of curing M4. This indicates that 5 wt% MK acts as a nucleating agent. time in Fig. 10. The porosity of the hardened pastes depends On the other hand, the weight loss decreases with increasing on many factors and typically increases with increasing W/C MK content due to the pozzolanic reaction of MK with CH. ratio and decreases with curing time; in addition, the type of The last step in the TG thermograms is the decomposition of cement plays a certain role [29]. The paste has the highest CaCO3; the TG losses are 4.49%, 5.044%, 5.183% and porosity immediately after mixing with water. As the hydra- 6.883% for OPC and OPC–MK pozzolanic cement pastes tion proceeds, the hydration products fill a part of the pore made with 5, 15 and 20 wt% MK. Therefore, the CaCO3 con- volume, and accordingly the porosity decreases [30]. The total tents increase with increasing MK content in the OPC–MK porosity of OPC and OPC–MK pozzolanic cement pastes pozzolanic cement pastes. decreases with curing time due to the continuous precipitation of more hydration products within the pore system which leads Bulk density to decrease the porosity. As the MK content of the OPC–MK mixture increases up to 20 wt% the total porosity values are The bulk density of OPC and pozzolanic cement pastes relatively high due to the increase of water of consistency of made of OPC substituted with 5, 10, 15 and 20 wt% MK are OPC–MK blended cement pastes in addition to the relatively

100 OPC 28 days 98 M1 28 days M3 28 days 96 M4 28 days 94 Exo. M4 92 90 88 86

Τ 84 Δ 82

M1 WEIGHT, % M3 80 OPC 78 76 DTG 74 ------DTA Endo. 72 70 0 200 400 600 800 1000 Temperature,oC

Fig. 9 DTA/DTG thermograms of OPC and OPC–MK pozzolanic cement pastes (M1, M3 and M4) cured up to 28 days.

2.19 OPC M1 M2 M3 M4 2.18 ------Total porosity

2.17 Bulk density 22

2.16 3

2.15

2.14

2.13 20 Total porosity, % Bulk density, g/cm 2.12

2.11

2.10

2.09 18 1 10 100 Curing time, days

Fig. 10 Bulk density as well as total porosity of OPC and OPC–MK pozzolanic cement pastes up to 90 days. low hydraulic character of MK in comparison with OPC. [32]. It is clear that the compressive strength of all cement pastes Therefore, the change in total porosity of OPC–MK blended increases with increasing time of hydration up to 90 days. This pastes depends not only on curing age, but also on the MK mainly due to the increase of the amount of hydration products content. The total porosity is sharply decreased up to 7 days such as CSH, CAH and CASH and their later accumulation and slightly decreased up to 90 days. This is due to the nucle- within the available pores giving high strength [33]. It is clear ating agent of MK which accelerates the hydration and then also that pozzolanic cement paste having 5% MK gives the the amount of hydration products increases and ﬁlls excessive highest compressive strength values than those of OPC or the parts of the open pores at early ages of hydration [30,31]. other pozzolanic cement pastes. This is also due to that It is clear that the OPC–MK paste containing 10 wt% MK 5 wt% MK acts as nucleating agent which accelerates the gives apparent porosity values similar to those of OPC pastes. hydration of OPC. On the other hand, there is no real difference of compressive strength of OPC and the other blended cement Compressive strength pastes containing 10–20 wt% MK. As the MK increases up to 20 wt% the compressive strength is slightly decreased. This is The compressive strength values of OPC and OPC–MK due to the dilution of OPC, which decreases the compressive Pozzolanic cement pastes are graphically represented as a func- strength of pozzolanic cement pastes [34]. tion of curing time in Fig. 11. The effect of pozzolana (MK) on The increase of strength would be due to higher surface the strength of OPC–MK pozzolanic cement pastes depends on area of MK resulting in more nucleation sites to improve reac- a number of factors such as the pozzolana content of the tivity and improve packing as well as their pozzolanic reactiv- cement, the type and surface area of pozzolana as well as the ity [35]. individual characteristics of the blending Portland cement Conclusion

140 From the above finding it may be concluded that: OPC M1 M2 (1) Metakaolin fired at 850 C for two hours is mainly com- 120 M3 posed of quartz as a crystalline mineral in addition to M4 the amorphous aluminosilicate glassy phase. (2) Substitution of 10 wt% OPC by MK increases the water 100 of consistency and then decreases up to 20 wt%. On the other hand, the initial and final setting times are elon- gated up to 10 wt% then shortened. 80 (3) The free lime contents of OPC–MK pozzolanic cement pastes increase up to 7 days then decrease up to 90 days Compressive strength, MPa due to the pozzolanic reaction of MK. 60 (4) The results of free lime of OPC–MK pozzolanic cement decrease with increasing time of hydration. (5) The results of XRD are in agreement with those of 110 100 Curing time, days DTA. (6) The compressive strength values of OPC–MK Fig. 11 Compressive strength of OPC and MK pozzolanic pozzolanic cement pastes containing up to 20 wt% have cement pastes up to 90 days. small differences up to 90 days.

[16] ASTM Designation: C191, Standard Method for Normal Conflict of interest Consistency and Setting of Hydraulic Cement, ASTM Annual Book of ASTM Standards, 04,01, 2008. [17] H. El-Didamony, A.A. Amer, H. Abdel-Aziz, Properties and There is no conflict of interest. durability of alkali-activated slag pastes immersed in sea water, Ceram. Int. 38 (2012) 3773–3780. References [18] H. El-Didamony, A.M. Sharara, I.M. Helmy, S. Abdel-Aleem, Hydration characteristics of b-C2S in the presence of some accelerators, Cem. Concr. Res. 26 (8) (1996) 1179–1187. [1] R. Siddique, M.I. Kahn, Supplementary Cementing Material, [19] Egyptian Specification, The Standard Methods for the Chemical Springer-Verlag, Berlin Heidelberg, 2011. Analysis, 2010. [2] R. Siddique, J. Klaus, Influence of metakaolin on the properties [20] ASTM Designation: C150, Standard Test Methods for of mortar and concrete: a review, Appl. Clay Sci. 43 (2009) 392– Compressive Strength of Hydraulic Cements, 2007, pp. 710–773. 400. [21] H. El-Didamony, M. Heikal, M. Shoaib, Homra pozzolanic [3] C. He, B. Osbaeck, E. Makovicky, Pozzolanic reactions of six cement silicates industrials, Ceram. Sci. Technol. 65 (3–4) (2000) principal clay minerals; activation, reactivity assessments and 39–43. technological effects, Cem. Concr. Res. 25 (8) (1995) 1691–1702. [22] J.J. Brooks, M.A. Johari, M. Mazloom, Effect of admixtures on [4] M. Murat, C. Comel, Hydration reaction and hardening of setting times of high-strength concrete, Cem. Concr. Compos. 22 calcined clays and related minerals III, influence of calcination (2000) 293–301. process of kaolinite on mechanical strengths of hardened [23] E. Badogiannis, G. Kakali, G. Dimopoulou, E. Chaniotakis, S. metakaolinite, Cem. Concr. Res. 13 (5) (1983) 631–637. Tsivilis, Metakaolin as a main cement constituent Exploitation [5] L. Mina´rik, K. Kopecsko´, Impact of metakaolin – a new of poor Greek kaolins, Cem. Concr. Compos. 27 (2) (2005) 197– supplementary material – on the hydration mechanism of 203. cement, Acta Tech. Napocensis: Civ. Eng. Arch. 56 (2) (2013) [24] M.F. Rojas, M.I. Sanchez de Rojas, Chemical assessment of the 100–110. electric arc furnace slag as construction material: expansive [6] M. Frıás, M.I. Sańchez de Rojas, J. Cabrera, The effect that the compounds, Cem. Concr. Res. 34 (10) (2004) 1881–1888. pozzolanic reaction of metakaolin has on the heat evolution in [25] L.P. Esteves, On the hydration of water-entrained cement-silica metakaolin-cement mortars, Cem. Concr. Res. 30 (2) (2000) system combined SEM, XRD and thermal analysis in cement 209–216. pastes, Thermochim. Acta 518 (2011) 27–35. [7] Z.J. ZLi, Z. Ding, Property improvement of Portland cement by [26] R. Vedalakshmi, Quantification of hydrated cement products of incorporating with metakaolin and slag, Cem. Concr. Res. 33 (4) blended cements in low and medium strength using TG and (2003) 579–584. DTA technique, Thermochim. Acta 407 (2003) 49–60. [8] C.A. Love, I.G. Richardson, A.R. Brough, Composition and [27] M. Antoni, J. Rossen, F. Martirena, K. Scrivener, Cement structure of C–S–H in white Portland cement-20% metakaolin substitution by a combination of metakaolin and limestone, pastes hydrated at 25 C, Cem. Concr. Res. 37 (2) (2007) 109– Cem. Concr. Res. 42 (12) (2012) 1579–1589. 117. [28] M.I. Khan, C.J. Lynsdale, P. Waldron, Porosity and strength [9] R. Fernandez, F. Martirena, K.I. Scrivener, The origin of the PFA/SF/OPC ternary blended pastes, Cem. Concr. Res. 30 (8) pozzolanic activity of calcined clay minerals: a comparison (2000) 1225–1229. between kaolinite, illite and montmorillonite, Cem. Concr. Res. [29] V.S. Ramachandran, J.J. Beaudoin, Handbook of Analytical 41 (1) (2011) 113–122. Techniques Concrete Science and Technology Principles [10] S. Wild, J.M. Khatib, A. Jones, Relative strength, pozzolanic Techniques and Application, William Andrew Publishing, New activity and cement hydration in superplasticized metakaolin York USA, 2001. concrete, Cem. Concr. Res. 26 (10) (1996) 1537–1544. [30] M. Frıás, J. Cabrera, Pore size distribution and degree of [11] N.J. Coleman, W.R. McWhinnie, The solid state chemistry of hydration of MK-cement pastes, Cem. Concr. Res. 30 (2000) metakaolin-blended Ordinary Portland cement, J. Mater. Sci. 35 561–569. (11) (2000) 2701–2710. [31] P.K. Mehta, P.C. Aitcin, Principles underlying production of [12] J.M. Kinuthia, S. Wild, B.B. Sabir, J. Bai, Self-compensating high-performance concrete, Cem. Concr. Aggr. 12 (2) (1990) 70– autogenous shrinkage in Portland cement-metakaolin-fly ash 78. pastes, Adv. Cem. Res. 12 (1) (2000) 35–43. [32] P.C. Hewlett, Lea ’s Chemistry of Cement and Concrete, [13] Z. Li, Advanced Concrete Technology, Wiley, Hoboken, NJ, Elsevier, Butterworth-Heinmann, 2004. USA, 2011, 14, 34, ISBN: 978-0-470- 43743-8. [33] A. Echart, H.M. Ludwig, J. Stark, Hydration of the four main [14] V. Kannan, K. Ganesan, Mechanical and transport properties in Portland cement clinker phases, Zem-Kalk-Gip 28 (8) (1995) ternary blended self compacting concrete with metakaolin and 443–452. fly ash, J. Mech. Civil. Eng. 2 (4) (2012) 22–31. [34] H. EL-Didamony, S. Abo-EL-Enein, A.H. Ali, T.M. EL- [15] P. Dinakar, P.K. Sahoo, G. Sriram, Effect of metakaolin Sokkary, Effect of silica fume on the slag cement containing content on the properties of high strength concrete, Int. J. wet cement dust, Ind. J. Eng. Mater. Sci. 6 (1999) 274–278. Concr. Str. Mater. 7 (3) (2013) 215–223.

Please cite this article in press as: H. El-Diadamony et al., Hydration and characteristics of metakaolin pozzolanic cement pastes, HBRC Journal (2016), http:// dx.doi.org/10.1016/j.hbrcj.2015.05.005