MODELING OF HYDRATION OF PORTLAND CEMENTS INCORPORATING SUPPLEMENTARY CEMENTING MATERIALS

Transcription

1 MODELING OF HYDRATION OF PORTLAND CEMENTS INCORPORATING SUPPLEMENTARY CEMENTING MATERIALS Ki-Bong Park (), Han-Seung Lee (2) and Xiao-Yong Wang (2) () Division of Architecture, College of Engineering, Kangwon National University, Korea (2) School of Architecture & Architectural Engineering, Hanyang University, Korea Abstract Supplementary cementing materials (SCM), such as silica fume (SF) and low-calcium fly ash (FL), have been widely used as mineral admixtures in high strength and high performance concrete. Due to the pozzolanic reaction between calcium hydroxide and SCM, compared with ordinary Portland cement, hydration process of cement incorporating SCM is much more complex. This paper presents a numerical hydration model which is based on multicomponent concept and can simulate hydration of cement incorporating SCM. The proposed model starts with mixture proportion of concrete and considers both Portland cement hydration reaction and pozzolanic activity. Using this proposed model, this paper predicts the following properties of hydrating cement-scm blends as function of hydration time: reaction ratio of SCM, calcium hydroxide content, heat evolution, porosity, chemically bound water and the development of the compressive strength of concrete. The prediction results agree well with experiment results.. INTRODUCTION The Supplementary cementing materials (SCM), such as silica fume and low-calcium fly ash, are by-products from metallurgical industries and thermal power plants. They have been widely used as mineral admixtures in high strength and high performance concrete. In addition, the benefits from economic and ecology aspects, such as energy-saving and resource-conserving, can be achieved by using SCM []. The experimental and practical investigation of the influence of SCM on durability and mechanical property of concrete is very abundant. Based on systematic experimental investigation and quantitative description of pozzolanic reaction between calcium hydroxide and pozzolanic materials, Papadakis [2-7] estimated the ultimate value, such as porosity, chemically bound water and calcium hydroxide amount during hydration process of blended cement. In reference[8], a multiphastic hydration model was built which can predict coupling hydration and the thermal and hydric states of cement-based materials (not only clinker but also mineral additions like fly ash or silica fume). Based on 3-D microstructural hydration 49

2 model, Bentz[9,] predicted heat evolution process of cement paste blended with silica fume and the influence of silica fume on characteristics of the interfacial zone. In reference [], a general hydration model which can predict heat evolution process of cementitious materials was presented. From reference [8-], we can see that based on hydration model, the evolution of properties of hydrating cementitious materials can be described accurately. Due to the pozzolanic reaction between calcium hydroxide and SCM, compared with ordinary Portland cement, hydration of cement-scm blends is much more complex. It is difficult to build chemical-based kinetic equation to quantitatively describe the evolution of cement mineral component during hydration [2] and only the ultimate value is available, such as the model proposed in reference [2-7]. However, only ultimate value is not enough to predict evolution of properties of concrete in construction site. In this paper, based on multi-component concept, a numerical hydration model which can simulate hydration of cement-scm blends is built. The proposed model starts with mixture proportion of concrete and considers both Portland cement hydration and pozzolanic activity. Using this proposed model, this paper predicts the properties of hydrating cement-scm blends as function of hydration time. The prediction results agree well with experiment results. 2. HYDRATION MODEL OF ORDINARY PORTLAND CEMENT 2. Assumption of hydration model In this hydration model, the influence of water to cement ratio, cement particle size distribution, cement mineral components and curing temperature on hydration reaction of Portland cement is considered. The assumptions of this model are shown as following:. Cement particles are cast randomly in the representative unit cell space, as shown in Figure. Cement hydration will start since cement particles contact with water. 2. The degree of hydration of cement paste can be regarded as a weighted sum of all cement particles and four maor mineral components, as shown in figure The liquid phase, which is assumed to be water, diffuses through the hydrate layer, reaches Fig. cement particles distribute randomly in cell space Fig. 2 schematic of multicomponent hydration model the surface of the cement particle and chemically reacts with cement. 4. Particle size distribution of cement can be approximated by R-Rammler function[3]. 2.2 Hydration Mechanism On the basis of Park s model [3], some improvements are done to consider cement particle size distribution and cement mineral component. The basic hydration equation of each mineral composition in cement particles can be described as following equation (): 42

3 dαi 3C = dt ( v w ) r r r w 2 i + ag ρi 3 3 ( ) + ( αi ) + ( αi ) kd Dei Dei kri i Where α denotes hydration degree of mineral component in a given cement particle. Subscript i, denotes mineral component, such as C 3 S, C 2 S, C 3 A and C 4 AF. Superscript denotes cement particle number. v i : stoichiometric ratio by mass of water to mineral component, as proposed in reference. w ag : physically bound water, equals to 3% of the weight of reacted cement. ρ i : density of anhydrate cement mineral component. k d : reaction coefficient in dormant period. r : radius of anhydrate cement particle. () D ei : effective diffusion k : coefficient of coefficient of water in the hydration product of each mineral component. ri reaction rate of each mineral component. The influence of temperature on reaction rate can be considered by Arrhenius law. The global degree of hydration of cement is a weighted average of mineral component and cement particles. 2.3 Water Withdrawl Mechanism During hydration period, given a certain time after initial setting time, due to the increasing interconnect among cement particles, the contact area between cement particle and surrounding water will be decreased. As a result, the slower hydration rate will be achieved. On the other hand the available space for packing of cement hydration product will be decreased because hydration products volume is 2.2 times bigger as reacted solid cement volume. As proposed in reference [3], with the increasing of gel-space ratio, the relative hydration rate will be decreased. Considering these two points, the modifications of equation () can be expressed as equation (2): ' d α i d α i freesurface w.38* Ce* = *( ) * α (2) dt dt totalsurface w freesurface In equation (2), the item ( ) is the ratio between free surface area (the area totalsurface which contacts with water) and total surface area. w is water content and Ce is cement w.38* Ce * α content in mixing proportion. The item considers the decreasing of w available space for packing of cement hydration product. 3. HYDRATION MODEL FOR CEMENT-SCM BLENDS 3. the amount of CH and CSH during hydration period As proposed in Papadakis paper [2-7], during hydration period, the chemical reactions of mineral component of Portland cement can be expressed as following equation (3): 2CS 3 + 6H CSH CH (3-) 2CS+ 4H CSH + CH (3-2)

4 CA 3 + CSH2 + H CASH 4 2 (3-3) CAF 4 + 2CH+ H CAFH 6 2 (3-4) By aforementioned multi-component hydration model and equation (3), the content of hydration product in unit volume of Portland cement, such as calcium hydroxide and CSH content, can be obtained from equation (4) as following: CH = (.5g α +.5g α 2 g α )*74.* Ce (4-) c3s c3s c2s c2s c4af c4af CSH = (.5g α +.5 g α )*342.4* Ce (4-2) c3s c3s c2s c2s In Papadakis paper [2-7], based on the experiment results of chemically bound water amount, calcium hydroxide amount and porosity during hydration period, the reaction stoichiometry of cement incorporating SCM is obtained. Silica fume is composed primarily of SiO 2 (S, 9.9%), which is almost completely vitreous (96%) and thus reactive. It is assumed that silica fume is practically active S and the rest is inert. The pozzolanic reaction of cement blended with silica fume is shown as equation (5-). In case of low-calcium fly ash, there is aluminosilicate(a-s) glass with a high content of silicon in reaction form. The hydration product of an aluminosilicate glass/hydrated lime mixture should be a CSH gel incorporating significant amounts of A. The silicon which presents as quartz or in crystalline aluminosilicate phases is inert. In Papadakis paper [6], the pozzolanic reaction in cement-low-calcium fly ash blends is written as following equations: 2S+ 3CH C S H (5-) A+ CSH2 + 3CH + 7H C4ASH2 (5-2) A+ 4CH + 9H C4AH3 (5-3) Based on hydration model and pozzolanic reaction shown in equation (5), the content of calcium hydroxide, chemical bound water and CSH can be rewritten as following equation (6): CH = (.5gc3sαc3s +.5gc2sαc2s 2 gc4afαc4af)*74.* Ce (.85* γs* fs, P+ 2.97* γ A* f A, P)* αreacted* P (6-) H = Hce + mh * * 2 O γ A fa, P * αreacted * P (6-2) CSH = (.5gc3sαc3s +.5 gc2sαc2s )*342.* Ce * fs, P * γ S * αreacted * P (6-3) In equation (6-), (6-2) and (6-3), f SP, and f AP, is weight fraction of S and A in SCM; γ S and γ A is their active (glass) part; α reacted is reacted degree of active(glass) part of SCM; P is SCM content in mixing proportion; H ce is chemically bound water of Portland cement; m H 2 O is chemically bound water of unit weight of reacted active(glass) part of SCM. In fact, as shown in equation (6), it is assumed that the active (glass) S and A, hydrate at equal rate. 3.2 simulation of pozzolanic reaction in cement-scm blends The hydration rate of pozzolanic material depends on the content of calcium hydroxide and degree of hydration. Compared to normal Portland cement, particle size distributions of silica fume are two orders of magnitude finer. In this paper it is assumed that pozzolanic reaction in cement-silica fume blends includes two processes: reaction process and diffusion process. So that the hydration equation of active(glass) part in silica fume can be written as following: 422

5 dαreacted mch 3 = dt P v r ρ r r 2 SF SF SF SF 3 SF 3 ( αreacted ) + ( ) + ( αreacted ) DeSF DeSF krsf (7-) DeSF = DeSF *ln( ) (7-2) αreacted Due to higher particle size, hydration rate of low-calcium fly ash is much lower than silica fume. In early duration period of -2 days no traces of reaction among FL particles can be detected [6]. In simulation it is assumed that pozzolanic reaction in cement-fl blends includes dormant period, reaction process and diffusion process, which is shown as following : dαreacted mch 3 = dt P v r ρ r r B k C α 2 FL FL FL FL FL 3 3 ( ) + ( αreacted ) + ( αreacted ) kdfl DeFL DeFL krfl (7-3) FL 3 dfl = + *( ).5 FL reacted (7-4) ( αreacted ) DeFL = DeFL *ln( ) (7-5) αreacted Where m CH is calcium hydroxide content in unit volume of hydrating blends; P is SCM content in mixture; v SF and v FL is stoichiometry ratio by mass of CH to SF and FL; rsf and r FL is radius of SF and FL particle; ρ SF and ρ FL is density of SF and FL; k dfl is reaction rate coefficients in dormant period; D esf and D efl are initial diffusion coefficients and k rsf and k rfl are reaction rate coefficients. Because water will be entrapped in pozzolanic CSH gel, compared with hydration of plain cement paste, hydration rate of cement in cement-scm blends will be decreased. By water reducing coefficient, this mechanism is considered. The reacted ratio of total fly ash (both active part and inert part), can be obtained based on equation (7) and weight fractions of constitutes in fly ash. The reacted ratio of total fly ash can be calculated as following equation (8): α flyash = ( γs * fs, P + γa * fa, P )* αreacted = (.82* *.24)* αreacted =.66* αreacted (8) Similarly, the reacted ratio of total silica fume, can be obtained as following: αsilica = γ s * fs, P * αreacted =.99*.96* αreacted =.87* αreacted (9) The heat evolution of hydrating blends consists of contribution from Portland cement hydration and SCM hydration reaction. The total heat evolution is expressed as equation (): dq * dαi * * dαreacted = Ce g H i i P H SCM dt + () dt dt Where H i and H SCM are specific heat generation contents of individual component. The porosity of hydrating blends can be determined as equation (): w ε = εh εp () ρ w Where ε h is porosity reduction due to hydration of Portland cement. And reduction due to pozzolanic activity. ε p is porosity 423

6 4. RESULT AND DISCUSSION 4. hydration degree of mineral components of ordinary Portland cement Experiment results of hydration of ordinary Portland cement in reference [4] is used. The comparison between experiment results and prediction resultsis shown in figure 3. In figure 4 and 5, the content of CH and CSH is shown as function of degree of hydration. It is obvious CH and CSH amounts are almost linearly dependent on hydration degree. 4.2 hydration of cement-scm blends In this part, the experiment results in [2,6,5] are adopted to verify proposed hydration model. The experiment results including compressive strength, bound water, calcium hydroxide content, heat evolution and porosity are used. The mixing proportion of specimens can be found in references[2,6]. Fig. 3 the comparison between experiment and simulation result of hydration degree of cement mineral components Fig. 4 calculated CH amount as function of degree of hydration Fig.5 calculated CSH amount as a function of degree of hydration Fig.6 the comparison of CH evolution : SF-cementpaste Fig.7 the comparison of CH evolution : FL-cementpaste Fig.8 calculated reacted ratio of total fly ash versus time Fig.9(a) calculated reacted ratio of total silica fume versus time :SFA5 424

7 Fig.9(b) calculated reacted ratio of total silica fume versus time :SFA5 Fig. comparison of chemically bound water evolution Fig. comparison of Porosity evolution evoltuion of compressive strength of concrete(mpa) days days 8 9 days days R 2 = days calculated CSH amount(g/cm 3 ) evoltuion of compressive strength of concrete(mpa) days 82 days 9 days 49 days 28 days R 2 = calculated CSH amount(g/cm 3 ) Fig.2comparison of heat evolurion FIG.3the evolution of compressive strength of mortar as function of calculated CSH amount: SF blends Fig. 4 the evolution of compressive strength of mortar as function of calculated CSH amount: FL blends Based on the amount of calcium hydroxide, the reaction coefficients of pozzolanic reaction can be obtained. The evolution of CH amount is shown as function of hydration time in figure 6 and figure 7. As shown in figures, the simulation results agree well with experiment results. The reacted ratio of total fly ash and silica fume are shown in figure 8 and figure 9. As shown in figure 8, the reacted ratio of FLA almost equals to zero in initial 5 hours (3 weeks). This agrees with SEM images experiments result: in the early hydration period of -2 days no traces of reaction among FL particles can be detected [6]. The results of chemically bound water and porosity are shown in figure and figure. In case of silica fume-cement blends, because selective reaction of CH with SF instead of its reaction with C 3 A and C 4 AF phase of cement between 4 days and 6 months is not considered in proposed model, simulation result of bound water is a little bigger than experiment result. The results of heat evolution is shown in figure 2. As shown in figure 2, the prediction results, overall agree well with experiment results. The difference between experiments and simulation results may be in that the size of fly ash particle is simplified as monosize. The fly ash particles with small size maybe hydrate much quickly than particles with median size. The results of development of compressive strength of mortar specimens are shown in figure 3 and 4. As shown in figure 4, in case of different FL addition, by regression between calculated CSH and compressive strength of mortars, a direct proportion function is obtained (at early age 3 days and 4 days, to consider contribution of aluminoferrite phases on 425

8 compressive strength, based on experimental results, the calculated CSH are multiplied by.85 and.92, respectively). But as shown in figure 3-a and 3-b, in case of different SF addition, the direct proportion function can not be found. As proposed in reference [6], according to a weak link theory for compressive strength, it is assumed that compressive strength of concrete should correlate with difference between the bulk and minimum loadbearing phase volume fractions, as this is a measure of the strength of the interfacial zone. When concrete incorporates 2% fly ash, the ratio of load-bearing phase volume fractions between minimum and bulk changes very small, from 47% to 44%[6]. Hence among different FL addition, as shown in figure 4, an unified linear relationship exists between compressive strengths and calculated CSH. Contrarily, when concrete incorporates % silica fume, this ratio changes very large, from 47% to 68% [6]. So among different SF addition, as shown in figure 3, the unified linear relationship does not exist.the realtion between compression strength and gel-space ratio now is under research. 5. CONCLUSION In this paper, based on multi-component concept, a numerical hydration model which can simulate hydration of cement-scm blends is built. The proposed model considers both Portland cement hydration and pozzolanic activity. Using this proposed model, this paper predicts the properties of hydrating cement-scm blends as function of hydration time. Furthermore, the durability aspect, such as carbonation and the early-age thermal crack of concrete structure in construction site can be a potential result. REFERENCES: [] P.Kumar Metha, Concrete-Microstructure, Properties and Materials, third edition(published by MaGraw-Hill Companies, 26) [2] Vagelis G. Papadakis, experimental investigation and theoretical modeling of silica fume activity in concrete, Cem. Concr. Res., 29(999)79-86 [3] Vagelis G. Papadakis, S.Tsimas, supplementary cement in materials in concrete, Part I: efficiency and design, Cem. Concr. Res., 32(22) [4] Vagelis G. Papadakis, S.Tsimas, supplementary cement in materials in concrete, Part II: a fundamental estimation of the efficiency factor, Cem. Concr. Res., 32(22) [5] Vagelis G. Papadakis, S.Tsimas, effect of supplementary cementing materials on concrete resistance against carbonation and chloride ingress, Cem. Concr. Res., 3(2) [6] Vagelis G. Papadakis, effect of fly ash on Portland cement systems, Part I: low-calcium fly ash, Cem. Concr. Res., 29(999) [7] Vagelis G. Papadakis, effect of fly ash on Portland cement systems, Part II: high calcium fly ash, Cem. Concr. Res.,3(2) [8] L.Buffo-Lacarriere, A.Sellier, G.Escadeillas, A.Turatsinze, multiphasic finite element modeling of concrete hydration, Cem. Concr. Res., 37(2)3-38 [9] D.P.Bentz,V.Waller and F.de Larrard, prediction of adiabatic temperature rise in conventional and high performance concretes using 3-D microstructural model, Cem. Concr. Res., 28(999) [] D.P.Bentz, Stutzman, P.E., and Garboczi, E.J., Experimental and Simulation Studies of the Interfacial Zone in Concrete, Cem. Concr. Res., 22(992) [] Anton K.Schindler and Kevin J.Folliard, heat of hydration models for cementitious materials, ACI mater. J., 2(25)

9 [2] Vagelis G. Papadakis, Michael N.Fardis, and Costas G. Vayenas, hydration and carbonation of pozzolanic cements, ACI Mater. J., 89(992) 9-3 [3] Ki-Bong Park, Takafumi Noguchib, Joel Plawsky, Modeling of hydration reaction using neural network to predict average properties of cement paste, Cem. Concr. Res., 35(25) [4] T.Matsushita, S.Hoshino,I.Maruyama,T.Noguchi,K.Yamada, Effect of curing temperature and water to cement ratio on hydration of cement compounds, Proc. 2th ICCC (27)TH2-7.3 [5] Ivindra Pane, Will Hansen, Investigation of blended cement hydration by isothermal calorimetry and thermal analysis, Cem. Concr. Res., 35(25)55-64 [6] D.P.Bentz and E.J.Garbociz, simulation studies of the effects of mineral admixtures on the cement paste-aggregate interfacial zone, ACI mater. J., 88(99)