%28asce%29mt%2e1943-5533%2e0000421.pdf

Concrete Containing Natural Pozzolans: New Challenges for Internal Curing

Downloaded from ascelibrary.org by Chonbuk National University on 05/20/16. Copyright ASCE. For personal use only; all rights reserved.

Gaston Espinoza-Hijazin1; Álvaro Paul2; and Mauricio Lopez, Ph.D.3 Abstract: Natural pozzolans (NP) have proven to be an effective supplementary cementitious material; however, the replacement of ordinary portland cement (OPC) with NP might increase the autogenous and drying shrinkage of concrete. Internal curing (IC) might be of great help when using NP because it can promote the pozzolanic reactions and reduce shrinkage. The aim of this research is to assess the effect of IC in concrete containing NP. Results indicate that a 39% replacement of OPC with NP decreased compressive strength by 15%, decreased chloride ion permeability by 66%, and increased autogenous shrinkage by 40%. IC with prewetted lightweight aggregate showed no significant effect in compressive strength or permeability, but it decreased autogenous shrinkage by up to 58%. NP used in this investigation presented higher chemical shrinkage than OPC, making IC less effective as levels of NP increased. The important decrease in permeability attained through the use of NP and the higher chemical shrinkage of NP makes IC a critical technology to consider in concrete mixtures with NP. DOI: 10.1061/ (ASCE)MT.1943-5533.0000421. © 2012 American Society of Civil Engineers. CE Database subject headings: Concrete; Cement; Compressive strength; Shrinkage; Permeability; Curing. Author keywords: Concrete; Cement; Curing; Compressive strength; Autogenous shrinkage; Permeability; Internal curing; SCM.

Introduction The curing of concrete is an important process during the first hours after casting to maintain optimal conditions for cement hydration and for assuring the required durability and strength of the hardened concrete, thus enabling a high performance of the structure during its service life (Mehta and Monteiro 2006). Therefore, to maximize the degree of hydration of cement and possibly that of the supplementary cementitious materials (SCM) and to reduce early shrinkage cracking (autogenous shrinkage, plastic shrinkage, and drying shrinkage), it is important to apply effective curing techniques during an extended period of time (Rilem TC 199-ICC 2007). Common curing techniques consider an external water supply to maintain a high internal relative humidity (RH) by water spraying or fogging, watering, use of wet coverings, or ponding of the concrete element (ACI Committee 308; Kovler and Jensen 2005). Although these external curing techniques may be useful for conventional concrete, they might be difficult to apply for the long periods of time needed for hydration because as they can interrupt or delay other critical activities during construction. Insufficient external curing can produce an important drop in the internal 1

Civil Engineer, School of Engineering, Pontificia Universidad Catolica de Chile, Santiago, Chile; and Adjunct Professor, School of Civil Engineering, Universidad Diego Portales, Ejército 441, 8370191, Santiago, Chile. E-mail: [email protected] 2 Civil Engineer, Assistant Professor, School of Civil Engineering, Universidad Diego Portales, Ejército 441, 8370191, Santiago, Chile. E-mail: [email protected] 3 Civil Engineer, Assistant Professor, School of Engineering, Pontificia Universidad Catolica de Chile, Casilla 306, Correo 22, 6904411, Santiago, Chile (corresponding author). E-mail: [email protected] Note. This manuscript was submitted on June 6, 2011; approved on November 7, 2011; published online on November 10, 2011. Discussion period open until January 1, 2013; separate discussions must be submitted for individual papers. This paper is part of the Journal of Materials in Civil Engineering, Vol. 24, No. 8, August 1, 2012. ©ASCE, ISSN 0899-1561/ 2012/8-981–988/$25.00.

relative humidity of the concrete and even interrupt the hydration process. Curing is especially important when the concrete includes SCMs such as pozzolans, silica fume or fly ash because secondary hydration reactions require water for a longer period of time. In concretes with a water-cement ratio (w∕c) below approximately 0.42, there would be insufficient water to promote complete hydration of the portland cement under sealed conditions, producing a quick consumption of water and an accelerated drop of the internal humidity of the mixture. Thus, there might not be enough water available for the SCMs to react with calcium hydroxide to form calcium silicate hydrates (C-S-H). A failure to reach high degrees of hydration makes unhydrated cementitious materials act as inert filler, reducing both the depercolation of the pore structure and the potential durability of the concrete. To overcome this problem, it is necessary to supply additional water during curing because avoiding water loss is not as sufficient as other conventional external curing methods (ACI 308 2001). Additionally, it is well established that concrete with a w∕c below 0.4 tends to possess low permeability; hence, even with good external curing practices, water entry will be limited (Bentz 2007, Lopez et al. 2005). A couple of decades ago, the internal curing method (IC) was suggested to complement external curing (Philleo 1991; Jensen and Hansen 2001) to overcome the problems described in the preceding paragraphs. The method consists of storing water within concrete, which is not available during mixing so it does not influence the initial w∕c ratio but is later released to promote further hydration reactions of cementitious materials. IC has been successfully applied and understood in concrete mixtures containing ordinary portland cement (OPC) and OPC with SCMs such as fly ash and silica fume. However, the effect of IC on mixtures containing SCMs such as natural pozzolans (NP) has not been extensively investigated even though such SCMs can be a sustainable alternative for the next few decades. To understand the implications of using IC in concrete with NP, it is necessary to introduce some concepts related to hydration.

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012 / 981

J. Mater. Civ. Eng., 2012, 24(8): 981-988

Downloaded from ascelibrary.org by Chonbuk National University on 05/20/16. Copyright ASCE. For personal use only; all rights reserved.

Hydration of Cementitious Materials Two conditions are needed to obtain the theoretical complete hydration of the cementitious materials: (1) enough space for the hydration products and (2) enough water for hydration reactions. It has been established that concrete with a water-to-cementitious materials ratio (W∕CM) below 0.36 might not provide enough space for hydration, whereas concrete mixtures with W∕CM between 0.36 and 0.42 might not provide enough water, and some additional curing water is needed (Powers 1960; Mindess et al. 2003). Hydration occurs when cementitious materials come in contact with water, thus beginning the formation of hydration products such as C-S-H, which explain most of the concrete’s strength, and calcium hydroxide, ettringite, and monosulphate. Part of the water that does not chemically react with cementitious materials is absorbed by the surface of the hydration products, another fraction is retained within the interlayer voids of C-S-H, and the rest remains as a solution inside the capillary pores formed during chemical shrinkage (Mehta and Monteiro 2006). Chemical shrinkage can be understood as the reduction of the volume of the hydrated cement paste as a result of hydration products occupying less volume than the original products that form the cement paste. Cementitious materials obtain the necessary water to promote their hydration from the capillary pores, which are emptied as hydration proceeds. This drainage of capillary pores generates capillary stresses that produce volumetric reductions in a closed isothermal system not subjected to external forces, known as autogenous shrinkage (Jensen and Hansen 2001). Secondary Chemical Reactions in Cements with Natural Pozzolans There are two main classifications of pozzolans depending on the source: Artificial and natural pozzolans. Artificial pozzolans are those originated as by-products of some industrial processes, such as power plants (i.e., fly ash), steel production (i.e., ground granulated blast furnace slag), and ferrosilicon production (i.e., silica fume). Natural pozzolans are naturally occurring pozzolans present in the earth. They are usually abundant in silica and have high aluminum and iron contents. They can be classified in two categories: Natural rocks that only require a grinding process to be used (i.e., volcanic glass and pumicite) and soils that require a thermal process to activate (i.e., calcined clays and metakaolin). The chemical composition and fineness of each category of natural pozzolans will determine their reactivity. ACI 116R (2000) defines natural pozzolans (NP) as “siliceous materials that possesses little or no cementitious value but that will, in finely divided form and in the presence of moisture, chemically react with calcium hydroxide at ordinary temperatures to form compounds having cementitious properties.” This reaction is known as a pozzolanic reaction. During the pozzolanic reaction in concrete, two phenomena occur: First, there is a slow decrease of free calcium hydroxides over time, and second, there is an increase in the formation of silicates and aluminous-silicate hydrates, similar to the hydration products of OPC. Studies show that calcium hydroxide content decreases from 2.2 to 0.7% in concrete made with portland cement with 40% pozzolan additions compared with concrete prepared with pure portland cement (Lea 1971), which indicates a greater content of C-S-H and other hydration products with low porosity (Mehta 1987). Characteristics of Concrete with Pozzolan Additions Physical properties of NP, such as shape, fineness, particle-size distribution, density, and composition of NP, have a great influence in

fresh concrete properties and in the development of hardened concrete strength (ACI 232 2002). Partial replacements of portland cement by high SiO2 ∕ðAl2 O3 þ Fe2 O3 Þ ratio pozzolans increase concrete resistance to both sulfate attack and substances contained in sea waters (Mehta 1987). In some proportion, this is ascribed to the decrease in free calcium hydroxide formed during portland cement hydration. The result is that the hardened concrete contains less calcium hydroxide and greater amounts of C-S-H and other low-porosity products. On the other hand, the decrease in calcium hydroxides in concrete reduces the initial pH, making it more vulnerable to carbonation, steel depassivation, and probably to corrosion (ACI 232 2002). Concrete with NP additions has proven to have a smaller pore structure as a result of the pozzolanic reaction, which becomes an important characteristic in the increase of concrete durability and mechanical strength (Mehta 1987). Moreover, a finer porosity of the C-S-H obtained from the pozzolanic reaction has been previously reported (Baroghel-Bouny 1996; Bentz et al. 2000). Use of Natural Pozzolans in Concrete NP is recommended as a partial replacement or addition to portland cement for concretes that need a low heat of hydration; for example, dams or other structures built with mass concrete and in structures that require sulfate resistance concrete, such as bridges or piers (Mielenz et al. 1950). Nowadays, the use of NP is suggested for cost reasons, to reduce energy consumption by reducing the CO2 emissions from portland cement manufacturing, and to reach some technical benefits, such as a lower permeability of the hydrated paste (Mehta 1987). Previous research has demonstrated that the use of pozzolans as cement volume replacements in 20 to 30% produces less heat of hydration, increases paste cohesion, and reduces expansion caused by the alkali–silica reaction (Andriolo 1975). Effects of Natural Pozzolans in Concrete Properties The specific gravities of NP are generally lower than that of OPC, so when OPC is replaced with NP by volume, a concrete of lower density is obtained; when OPC is replaced with NP by mass, a larger volume of fresh concrete is obtained. Some NPs produce an important increase in water requirements and others have an insignificant effect (Mather 1958). Additionally, because the pozzolanic reaction takes longer to occur in comparison with OPC, concrete mixtures containing NP will require moist curing for a longer periods of time. The effect of the pozzolanic reaction will be highly dependent on the NP used, namely, its fineness and reactivity. For example, metakaolin (calcined clay) presents an early age reaction similar to that of silica fume, and grinded volcanic glass reacts at a lower rate than that of a Class F fly ash. The use of NPs, in general, produces a more cohesive mixture, maintaining plastic consistency and improving workability; however, this effect is highly dependent on the fineness of the NP (Turanli et al. 2004). The NP can adsorb water from the mixture and keep it within the system for improving concrete finishing (ACI 232 2002). When using aggregates that are deficient in particles smaller than 75 μm, the use of very fine pozzolans can reduce bleeding and segregation and can increase concrete strength by improving particle packing. The addition of pozzolans such as NP produced concrete mixtures with lower permeability, lower heat of hydration, lower expansion as a result of the alkali–silica reaction, greater strength at later ages, and greater resistance to sulfate attack when compared with pure portland cement samples (Mather 1958; Lopez and Castro 2010).

982 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012

J. Mater. Civ. Eng., 2012, 24(8): 981-988

Downloaded from ascelibrary.org by Chonbuk National University on 05/20/16. Copyright ASCE. For personal use only; all rights reserved.

Concrete with a W∕CM of 0.4 and a 15% replacement of OPC with NP by mass shows an increase from 36 to 48 MPa in compressive strength (Zhang and Malhotra 1996). A research study on the alkali–silica reaction showed a reduction in the expansion, at one year of age, from 0.28 to 0.02% with a 30% replacement of OPC with NP by mass (Saad et al. 1982). Although the use of low-alkali cements is satisfactory to avoid expansion in a large number of the most reactive aggregates, some aggregates require an additional control, which is provided through the use of NP or other SCMs. Bentz (2007) carried out an extensive study on the autogenous and chemical shrinkage of different blended cement with different SCMs and observed that autogenous shrinkage was greater in blended cements than in OPC; blended cements with NP were not included in this study, so their effect needs to be investigated. The drying shrinkage of concrete prepared with blended cements containing NP depends on the hydrations of the cementitious materials and the water requirements of the mixture. A research study (Mehta 1981) showed that drying shrinkage in concretes with NP (10, 20, and 30% replacement of OPC with NP) did not show significant differences in drying shrinkage compared with the control sample, which had only OPC; all specimens showed drying shrinkage values between 500 and 600 με. Conversely, Zhang and Malhotra (1995) found an increase from 400 to 600 με in drying shrinkage when replacing OPC with 10% of NP (metakaolin). Similarly, Videla et al. (2004) found that all drying shrinkage models greatly underestimated the drying shrinkage of blended cements containing NP.

Research Significance The use of NP in concrete has proven beneficial for several concrete properties; however, curing is even more important than for conventional concrete because of the longer hydration times involved and higher shrinkage. To take full advantage of the use of NP, appropriate curing methods should be ensured. The objective of this research is to assess the potential beneficial effects of applying IC to concrete mixtures with NP. Particularly, the research seeks to quantify and explain the effects of IC on compressive strength, chloride ion permeability, and autogenous shrinkage for concretes containing NP.

Experimental Procedure The experimental procedure was designed to quantify the effect of IC on concrete performance as the replacement of OPC with NP increases. Three relevant properties of concrete were analyzed: The compressive strength to symbolize the structural behavior; The autogenous shrinkage, to represent the cracking tendency; and the chloride ion permeability, to represent the durability of concrete.

Eight mixtures were prepared at a W∕CM of 0.3, with different amounts of replacement of OPC with NP, with or without an IC agent. Testing was performed at the age of 90 days to ensure a high degree of hydration. The strength and permeability specimens were cast in 100 mm ðdiameterÞ × 200 mm ðheightÞ cylinders, which were demolded at 24 h of age and kept in a fog room with an approximately relative humidity of 100% and temperature of 23°C. Autogenous shrinkage specimens were prepared in flexible corrugated molds, 350 mm ðlengthÞ × 22 mm ðdiameterÞ, as detailed in subsequent paragraphs, and the measurements started after the final set. With the purpose of controlling the hydration process of cementitious materials and to avoid variability, the paste and aggregates volume was kept constant in all trial mixtures. With regard to the cementitious materials, concrete with four different combinations of OPC and NP were prepared. The replacement of OPC with NP was made at 0, 13, 26, and 39% levels expressed by mass. Two mixtures were prepared for each combination, a control mixture without IC and another one with IC through the use of saturated lightweight fine aggregate. The NP used in this investigation had a volcanic origin and could be classified as a Type-N pozzolan according to ASTM C618 (2008). It corresponded to a volcanic glass known as rhyolite pumicite, and it was obtained from the area known as Polpaico. The chemical compositions of the NP and OPC used in this investigation are shown in Table 1. Additionally, a scanning electron microscope (SEM) image of the NP can be seen in Fig. 1, in which a void and fluid structure and pyroplastic texture are appreciable. The chemical composition of the NP used in this investigation suggests an almost negligible hydraulic activity (2.7% of CaO content) and a relatively high pozzolanic activity because of its 69.2% of SiO2 . Therefore, the effect of using NP on the properties of the concrete will be more noticeable at later ages (Uzal and Turanli 2003). Although there are several potentially useful curing agents to perform internal curing, this research considered the use of expanded clay fine aggregate (particle size of under 5 mm) because it has been successfully applied in several studies for mixtures of low W∕CM (e.g., Paul and Lopez 2011; Jensen and Lura 2006). This curing agent was chosen for its absorption and desorption capacity and its particle size, which allows a better distribution within the concrete mixture. This optimizes the efficiency of internal curing by minimizing the water transport distance (Bentz 2007). The expanded clay’s properties used are shown in Table 2. The mixtures, prepared at a W∕CM of 0.3, were designed to maintain the paste volume constant at 35%. The lightweight aggregate (LWA), used as the internal curing agent, was replaced with normal fine aggregate in equal volume, and the amount was calculated on the basis of the method developed by Bentz and Snyder (Bentz et al. 2005). According to this method, the curing water amount is directly proportional to the cement amount, the degree of maximum hydration expected (83.33%), and the chemical shrinkage (CS) to complete hydration.

Table 1. Chemical Composition of NP (Rafael Cepeda, Personal Communication, May 11, 2011) and OPC, by Courtesy of Rafael Cepeda Natural pozzolan (rhyolite pumicite–volcanic glass) (%) SiO2 69.2

Al2 O3 13.2

Fe2 O3 1.7

CaO 2.7

MgO 0.8

K2 O 3.0

SO3 0.1

Na2 O 3.9

TiO2 0.2

P2 O5 0.1

LOI 4.36

Ordinary portland cement (Type I) (%) SiO2 20.8

Al2 O3 6.0

Fe2 O3 2.9

CaO 63.2

SO3 1.8

Blaine fineness (m2 ∕kg) 360

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012 / 983

J. Mater. Civ. Eng., 2012, 24(8): 981-988

Downloaded from ascelibrary.org by Chonbuk National University on 05/20/16. Copyright ASCE. For personal use only; all rights reserved.

Tests Performed

Fig. 1. SEM image of the natural pozzolan surface, 1,200 X (Rafael Cepeda, personal communication, May 11, 2011); by courtesy of Rafael Cepeda

Table 2. Properties of the Internal Curing Agent, Expanded Clay Particle size range (mm)

0.08–5

1-day absorption (%) 3-day absorption (%) 32-day absorption (%) 90-day absorption (%) Oven dry density (kg∕m3 ) SSD density (kg∕m3 ) Porosity (%) Water release rate (grams of water∕grams of dry aggregate∕day × 105 ) desorption

21.2 28.5 30.8 31.1 1,653 1,920 41 10.1

The compressive strength tests were carried out at 90 days, according to ASTM C39 (2001). At least three cylinders of 200 mm in height and 100 mm in diameter were tested for each one of the eight mixtures prepared. The chloride ion permeability test was performed in accordance with ASTM C1202 (2005); test specimens of 50
2 mm in height and 100 mm in diameter were cut from the 100 × 200 mm cylinders, and they were sealed with rubber cement on their lateral section. They were then vacuum saturated before testing. Specimens were placed on two cells containing different solutions: 3% NaCl solution (which will be connected to the negative pole) and 0.3 N NaOH solution (which will be connected to the positive pole). The two poles are then connected a 60-V power supply, and the resulting current intensity through the specimen is measured every 5 min for 6 h. The chloride ion permeability is expressed as the charged passed during the 6 h through the specimen. The autogenous shrinkage was measured by applying the method developed by Jensen and Hansen (1995) that recently became ASTM Standard C1698 (2009). The concrete mixture was placed in a corrugated, flexible, and airtight cylinder mold, which does not restrain shrinkage and prevents moisture loss. Specimens were kept at a constant temperature to avoid thermal strains. Therefore, measured strains correspond to autogenous shrinkage (shrinkage by self-desiccation: Autogenous deformation in a material's system) (Jensen and Hansen 2001). The setting time was measured by the Vicat test (ASTM C191 2004). The length of the specimen after the final set was considered as the initial length of the specimen, and the change in length was measured for a period of 50 days.

Test Result Analysis Compressive Strength

This study considered a CS of 0.055 (typical for OPC; Kovler 2007), regardless the replacement of OPC with NP. Therefore, the mixtures with NP considered an amount of IC agent that might be insufficient if the NP has a chemical shrinkage greater than the OPC, as suggested by previous studies (Bentz 2007). Table 3 presents the mixture designs used in each of the eight mixtures analyzed.

Compressive strength was tested 90 days after mixing (Fig. 2). As expected, the replacement of the OPC with NP increased the compressive strength for low levels of replacement. However, when replacement levels were higher (26 and 39% by mass), a reduction in the 90-day compressive strength was observed. Recent studies

Table 3. Mixture Designs of Concretes (kg∕m3 ) Mixture IDa

OPC

NP

0% 0% IC 13% 13% IC 26% 26% IC 39% 39% IC

518 518 442 442 370 370 300 300

0 0 66 66 130 130 192 192

Coarse Fine Water NWAb NWAc 155 155 153 153 150 150 148 148

790 790 790 790 790 790 790 790

965 799 966 802 965 804 965 807

LWA for IC

HRWRd

0 119 0 117 0 115 0 113

13 13 13 13 13 13 13 13

a

All mixtures were designed on a 2% entrained air basis. NWA: Normal weight aggregate, siliceous aggregate (maximum size aggregate ¼ 10 mm). c NWA: Normal weight aggregate, siliceous aggregate (maximum size aggregate ¼ 5 mm). d HRWR: High range water reducing admixture. b

Fig. 2. Effect of IC on compressive strength at 90 days: 0, 13, 26, and 39% replacement of OPC with NP by mass (W∕CM ¼ 0:3); samples internally cured present higher variability; error bars show the range of results for each mixture

984 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012

J. Mater. Civ. Eng., 2012, 24(8): 981-988

Downloaded from ascelibrary.org by Chonbuk National University on 05/20/16. Copyright ASCE. For personal use only; all rights reserved.

have suggested that an approximately 20% replacement of OPC with NP of a different chemical composition would maximize the compressive strength of concrete (Pekmezci and Akyüz 2004; Mouli and Khelafi 2008). Fig. 2 shows that there would be no significant differences between the compressive strength of mixtures with or without IC (for an equal replacement proportion with natural pozzolan) and that the variability of the tested specimens was relatively low. Because all of the mixtures had the same amount of IC agent and the intrinsic strength of LWA is lower than that of the normal aggregate being replaced (Videla and Lopez 2002; Paul and Lopez 2011), a reduction in the compressive strength of the mixtures with internal curing is expected if no other effect is considered. Nevertheless, Fig. 2 does not show strength reduction when using an IC agent (i.e., there was a strength increase in the cement paste as a result of the internal curing). In other words, the hydration increase produced by IC (Espinoza-Hijazin and Lopez 2011) has produced a strength increase in cement paste that is capable of counteracting the strength loss caused by using LWA. When analyzing the levels of NP replacement, a general trend of compressive strength reduction is observed as the NP content increases. For instance, the mixtures with a 39% replacement of OPC with NP presented a

Fig. 3. Effect of IC on chloride ion permeability at 90 days: 0, 13, 26, and 39% replacement of OPC with NP by mass (W∕CM ¼ 0:3); dashed lines indicate the frontier between each classification of chloride ion penetrability, according to ASTM C1202(2005)

compressive strength of approximately 86.7% of those with only OPC. This suggests that at 90 days of age, some NP has already hydrated, contributing to compressive strength; otherwise, the drop in strength would have been greater. These results agree with those obtained previously using NP with similar SiO2 contents (Uzal and Turanli 2003), in which a reduction in compressive strength at 90 days was observed. Nevertheless, results are not in agreement with those obtained by Papadakis and Tsimas (2002) using NP with less SiO2 and more CaO. Even though IC seems to either maintain or increase the 90-day compressive strength, it does not show statistically relevant benefits regarding a compressive strength gain, as concluded by a t-student statistical test by comparison of means. Therefore, the increase in compressive strength as a result of enhanced hydration of the specimens with IC was somehow counteracted by the increase in porosity when including LWA in the mixture as an IC agent. Chloride Ion Penetration A chloride ion penetration test was performed at 90 days of age. Fig. 3 shows the results with the average value of at least two samples. Results show that there is no significant difference in the chloride ion permeability between the mixtures with and without IC. Thus, it may be concluded that internal curing would not offer additional benefits to those of an efficient external curing. On the other hand, the results show a clear permeability reduction as the level of NP replacement increases. This would indicate that the NP reacts with calcium hydroxide in the presence of moisture, so the hardened concrete contains less calcium hydroxide and more C-S-H and other low-porosity products, which entails lower permeability (Mehta 1987). A reduction in compressive strength and a great reduction in permeability because of the use of NP has been observed previously (Uzal and Turanli 2003) and might be a result of the fact that the contribution of NP is more efficient in helping the depercolation of capillary pores than in reducing the overall porosity; this, however, needs to be further investigated. Even though IC seems to decrease chloride ion permeability in some cases, it does not show statistically relevant benefits in the reduction of permeability, as concluded from the t-student statistical test. This might be a result of the fact that control mixtures

Fig. 4. Effect of IC on autogenous shrinkage: 0, 13, 26, and 39% replacement of OPC with NP by mass (W∕CM ¼ 0:3); zero value corresponds to time of final set; samples internally cured show lower autogenous shrinkages until Day 56 JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012 / 985

J. Mater. Civ. Eng., 2012, 24(8): 981-988

Downloaded from ascelibrary.org by Chonbuk National University on 05/20/16. Copyright ASCE. For personal use only; all rights reserved.

(without IC) already contained very low chloride ion permeability. The compressive strength and chloride ion penetrability tests showed a decrease in these properties as the NP content increases and did not prove any significant influence with the application of IC. One possible explanation of this nonrelevant effect of IC as NP increases might be found in the NP water demand. Previous research has concluded that NP water demand is greater than that of OPC (Ayres and Khan 1993; Uzal and Turanli 2003; Malhotra and Mehta 2009); in consequence, at greater amounts of NP, the lack of curing water increases. Moreover, it must be considered that the pozzolanic reaction produces a greater chemical shrinkage and is relatively unaffected by the relative humidity drop (Jensen and Hansen 2001). Additionally, a study on IC with different types of pozzolans concluded that the amount of IC water required is highly dependent on the type of pozzolan (Bentz 2007). Finally, all the specimens tested for compressive strength and chloride ion permeability were kept at 100% RH for the 90 days before testing. This ideal external curing could have helped to promote hydration, making the effect of IC less noticeable.

the sole exception of the mixture without IC and 13% NP, it may be stated that whenever the NP content increases, the autogenous shrinkage increases, as well. This suggests that NP presents a higher autogenous shrinkage and higher water consumption, which accelerates the mixture’s self-desiccation. Therefore, it may be concluded that efficient curing techniques, such as IC, become more necessary as the content of NP increases. It may also be concluded that the water content of IC was not enough to prevent the autogenous shrinkage occurrence of either the mixture with only OPC or of the mixtures with OPC and NP. Greater contents of IC water may be necessary to further decrease autogenous shrinkage. Results suggest that IC was able to reduce autogenous shrinkage between 26 and 53%. These results were compared with data obtained from IC applied to concrete mixtures with silica fume and fly ash, in which autogenous shrinkage was completely eliminated (Bentz 2007). This suggests that the NP used in this investigation might have a higher chemical shrinkage than other SCMs.

Autogenous Shrinkage

Conclusions

The autogenous shrinkage measured for at least 50 days in all investigated mixtures is shown in Fig. 4. After 50 days of measurement, mixtures without IC presented autogenous shrinkages between 963 and 1395 με, depending on the replacement level of OPC with NP. These values double or triple those reported previously for mixtures of blended cements (silica fume, fly ash, and slag) at the same W∕CM (Bentz 2007). This suggests that the NP used in this investigation has an elevated chemical shrinkage. Mixtures with IC presented autogenous shrinkages between 555 and 891 με, depending on the replacement level of OPC with NP (i.e., the higher the NP content, the higher the autogenous shrinkage). All of the samples with IC significantly reduced the autogenous shrinkage with respect to the mixtures containing the same amount of NP and without IC. Considering the replacement of the OPC with NP of 0, 13, 26, and 39% by mass, there was an autogenous shrinkage reduction of 621, 254, 641, and 382 με, respectively. Values obtained from the t-student statistical test indicate that there is a 99.9% probability of significant differences between the autogenous shrinkage of the samples with and without IC. The contribution of IC to the reduction of the autogenous shrinkage is mainly observed within the first 5 days, in which there is an autogenous expansion instead of an autogenous shrinkage. This has been previously observed by several authors (Bentz and Snyder 1999; Bentur et al. 2001; Jensen and Hansen 2002; Zhutovsky et al. 2002) and can be explained by the presence of water in the capillary pores, which reduces the capillary stresses, and by the presence of interlayer water on C-S-H, which produces a separation of layers and swelling. The reduction of autogenous shrinkage favors cracking reduction at early ages caused by stresses generated upon strain restraint. Autogenous shrinkage is delayed to later ages, when the concrete has gained higher strength and is more capable of resisting stresses, which reduces the cracking tendency; however, the elastic modulus of concrete has also increased, which might increase the cracking tendency. The experimental data and conclusions agree with the conclusions that IC minimizes the autogenous shrinkage without affecting the mechanical properties, such as compressive strength (Cusson and Hoogeveen 2008). Fig. 4 shows a clear relationship between the replacement of OPC with NP and the autogenous shrinkage in all mixtures. With

Concretes with a low W∕CM containing NP, such as those used in this research, require an efficient curing application because of high water demand. This is very difficult to achieve based only on traditional external curing because of the low permeability of these mixtures and the need for a long curing time to support the occurrence of pozzolanic reactions. Thus, IC seems a promising alternative to promote hydration in such complex circumstances. It has been observed that the main benefit of IC in these mixtures with a W∕CM ¼ 0:3 and NP is the autogenous shrinkage reduction. Additionally, the use of prewetted LWA as an IC agent neither reduces the concrete’s strength nor increases its chloride ion permeability, so it may be used to reduce autogenous shrinkage without the risk of negatively affecting other properties. The fact that IC is beneficial to the autogenous shrinkage but is not relevant to compressive strength or chloride ion permeability can be explained by the ideal external curing conditions. That is, the specimens used in strength and chloride ion permeability were moist cured for 90 days (100% relative humidity and 23
1°C), whereas the specimens for autogenous shrinkage were cured under sealed conditions. This implies that for compressive strength and chloride ion permeability specimens, the incremental benefit of IC may not have been significant compared with the external curing benefit; however, in the sealed specimens, in which gaining water from the outside was prevented, IC plays a critical role. It can also be concluded that the addition of NP produced differences in compressive strength as the replacement of OPC with NP increased. The difference between the maximum and minimum strength was 19.8%. The maximum strength was obtained with a 13% replacement of OPC with NP, whereas the lowest strength was obtained with a 39% replacement of OPC with NP. Additionally, results showed that the chloride ion permeability decreases as the NP content increases. Concrete mixtures having only OPC showed chloride ion permeability up to three times greater than that of concrete with 39% of NP. It was observed that autogenous shrinkage increases as NP content increases. Because the NP needs curing water at later ages because of the pozzolanic reaction, effective curing strategies, such as IC, become critical. Furthermore, NP reduces the permeability of concrete, so an external water supply (external curing) becomes more inefficient.

986 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012

J. Mater. Civ. Eng., 2012, 24(8): 981-988

Acknowledgments The authors greatly acknowledge the support given by the Chilean Council for Science and Technology Research (Conicyt) through Fondecyt Project #11060341 and the support of Rafael Cepeda, from Cementos Polpaico, for the important information provided. Additionally, the authors want to thank Mauricio Guerra, Patricio Garcia, and several students for their assistance during the research.

Downloaded from ascelibrary.org by Chonbuk National University on 05/20/16. Copyright ASCE. For personal use only; all rights reserved.

References American Concrete Institute (ACI). (2000). “Cement and concrete terminology.” ACI manual of concrete practice, ACI Committee 116, American Concrete Institute, Farmington Hills, MI. American Concrete Institute (ACI). (2001). “Guide to curing concrete.” ACI manual of concrete practice, ACI Committee 308, American Concrete Institute, Farmington Hills, MI. American Concrete Institute (ACI). (2002). “Use of raw or processed natural pozzolans in concrete.” ACI manual of concrete practice, ACI Committee 232, American Concrete Institute, Farmington Hills, MI. Andriolo, F. R. (1975). A utilizacado da pozzolana na construcado do conjunto hidroelectrico de urubupunga, CESP Publication, Companhia Energetica de Sao Paolo, Brazil (in Portuguese). ASTM. (2001). “Standard test method for compressive strength of cylindrical concrete specimens.” C39/C39M, West Conshohocken, PA. ASTM. (2004). “Standard test method for time of setting of hydraulic cement by Vicat needle.” C191, West Conshohocken, PA. ASTM. (2005). “Standard test method for electrical indication of concrete’s ability to resist chloride ion penetration.” C1202, West Conshohocken, PA. ASTM. (2008). “Standard specification for coal fly ash and raw or calcined natural pozzolan for use in concrete.” C618, West Conshohocken, PA. ASTM. (2009). “Standard test method for autogenous strain of cement paste and mortar.” C1698, West Conshohocken, PA. Ayers, M. E., and Khan, M. S. (1993). “Overview of fly ash and silica fume concretes: The need for rational curing standards.” Proc., V. Mohan Malhotra Symp., American Concrete Institute, Farmington Hills, MI, 605–617. Baroghel-Bouny, V. (1996). “Texture and moisture properties of ordinary and high-performance cementitious materials.” Proc., Int. RILEM Conf. “Concrete: From Material to Structure”, RILEM Publications, Bagneux, France, 144–165. Bentur, A., Igarashi, S. I., and Kovler, K. (2001). “Prevention of autogenous shrinkage in high-strength concrete by internal curing using wet lightweight aggregates.” Cem. Concr. Res., 31(11), 1587–1591. Bentz, D. P. (2007). “Internal curing of high-performance blended cement mortars.” ACI Mater. J., 104(4), 408–414. Bentz, D. P., Jensen, O. M., Coats, A. M., and Glasser, F. P. (2000). “Influence of silica fume on diffusivity in cement-based materials: I. Experimental and computer modeling studies on cement pastes.” Cem. Concr. Res., 30(6), 953–962. Bentz, D. P., Lura, P., and Roberts, J. (2005). “Mixture proportioning for internal curing.” Concr. Int., 27(2), 35–40. Bentz, D. P., and Snyder, K. A. (1999). “Protected paste volume in concrete: Extension to internal curing using saturated lightweight fine aggregate.” Cem. Concr. Res., 29(11), 1863–1867. Cusson, D., and Hoogeveen, T. (2008). “Internal curing of highperformance concrete with pre-soaked fine lightweight aggregate for prevention of autogenous shrinkage cracking.” Cem. Concr. Res., 38(6), 757–765. Espinoza-Hijazim, G., and Lopez, M. (2011). “Extending internal curing to concrete mixtures with W∕C higher than 0.42.” Constr. Build. Mater., 25(3), 1236–1242. Jensen, O. M., and Hansen, P. F. (1995). “A dilatometer for measuring autogenous deformation in hardening cement paste.” Mater. Struct., 28(7), 406–409. Jensen, O. M., and Hansen, P. F. (2001). “Autogenous deformation and RH-change in perspective.” Cem. Concr. Res., 31(12), 1859–1865.

Jensen, O. M., and Hansen, P. F. (2002). “Water-entrained cement-based materials: II. Experimental observations.” Cem. Concr. Res., 32(6), 973–978. Jensen, O. M., and Lura, P. (2006). “Techniques and materials for internal water curing of concrete.” Mater. Struct., 39(9), 817–825. Kovler, K. (2007). “Effect of internal curing on autogenous deformation.” Chapter 7, Internal curing of concrete, RILEM TC 196-ICC: State of the art report, K. Kovler and O. M. Jensen, eds., RILEM Publications, Bagneux, France. Kovler, K., and Jensen, O. M. (2005). “Novel techniques for concrete curing.” Concr. Int., 27(9), 39–42. Lea, F. M. (1971). The chemistry of cement and concrete, Chemical Publishing, Burlington, MA, 414–453. Lopez, M., Castro, J. T. (2010). “Effect of natural pozzolans on porosity and pore connectivity of concrete with time.” Revista Ingeniería de Construcción, 25(3), 419–431. Lopez, M., Kahn, L. F., and Kurtis, K. E. (2005). “Internal curing in high performance concretes—A new paradigm.” Revista Ingeniería de Construcción, 20(2), 117–126 (in Spanish). Malhotra, V. M., and Mehta, P. K. (2009). High-performance, high-volume fly ash concrete: Materials, mixture proportions, properties, construction practice, and case histories, Supplementary Cementing Materials for Sustainable Development Inc., Ottawa, ON. Mather, B. (1958). “The partial replacement of portland cement in concrete.” Cement and Concrete (STP-205), ASTM, West Conshohocken, PA, 37–73. Mehta, P. K. (1981). “Studies on blended portland cements containing santorin-earth.” Cem. Concr. Res., 11(4), 507–518. Mehta, P. K. (1987). “Natural pozzolans: Supplementary cementing materials.” Proc., Int. Symp. on Advances in Concrete Technology, V. M. Malhotra, ed., CANMET, Athens, Greece, 407–430. Mehta, P. K., and Monteiro, P. J. M. (2006). Concrete: Microstructure, properties and materials, McGraw-Hill, New York. Mielenz, R. C., Witte, L., and Glantz, O. (1950). Effect of calcination on natural pozzolans (STP-99), ASTM, West Conshohocken, PACI Committee 213.A, 43–92. Mindess, S., Young, J. F., and Darwing, D. (2003). Concrete, Prentice Hall, Upper Saddle River, NJ. Mouli, M., and Khelafi, H. (2008). “Performance characteristics of lightweight aggregate concrete containing natural pozzolan.” Build. Environ., 43(1), 31–36. Papadakis, V. G., and Tsimas, S. (2002). “Supplementary cementing materials in concrete, Part I: Efficiency and design.” Cem. Concr. Res., 32(10), 1525–1532. Paul, A., and Lopez, M. (2011). “Assessing lightweight aggregate efficiency for maximizing internal curing performance.” ACI Mater. J., 108(4), 385–393. Pekmezci, B. Y., and Akyüz, S. (2004). “Optimum usage of a natural pozzolan for the maximum compressive strength of concrete.” Cem. Concr. Res., 34(12), 2175–2179. Philleo, R. E. (1991). “Concrete science and reality.” Materials science of concrete II, J. Skalny and Mindess, S., eds., American Ceramic Society, Westerville, OH, 1–8. Powers, T. C. (1960). “Physical properties of cement paste.” Proc., Fourth Int. Symp. on the Chemistry of Cement, Portland Cement Association, Skokie, IL, 577–609. RILEM TC 199-ICC. (2007). State of the art report, K. Kovler and O. M. Jensen, eds., RILEM Publications, Bagneux, France. Saad, M., de Andrade, W. P., and Paulon, V. A. (1982). “Properties of mass concrete containing an active pozzolan made from clay.” Concr. Int., 4(7), 59–65. Turanli, L., Uzal, B., and Bektas, F. (2004). “Effect of material characteristics on the properties of blended cements containing high volumes of natural pozzolans.” Cem. Concr. Res., 34(12), 2277–2282. Uzal, B., and Turanli, L. (2003). “Studies on blended cements containing a high volume of natural pozzolans.” Cem. Concr. Res., 33(11), 1777–1781. Videla, C., Covarrubias, J. P., and Masana, C. (2004). “Updating concrete drying-shrinkage prediction models for local materials.” ACI Mater. J., 101(3), 187–198.

JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012 / 987

J. Mater. Civ. Eng., 2012, 24(8): 981-988

Zhang, M. H., and Malhotra, V. M. (1996). “High-performance concrete incorporating rice husk ash as a supplementary cementing material.” ACI Mater. J., 93(6), 629–636. Zhutovsky, S., Kovler, K., and Bentur, A. (2002). “Efficiency of lightweight aggregate for internal curing of high strength concrete to eliminate autogenous shrinkage.” Mater. Struct., 35(2), 97–101.

Downloaded from ascelibrary.org by Chonbuk National University on 05/20/16. Copyright ASCE. For personal use only; all rights reserved.

Videla, C., and Lopez, M. (2002). “Effect of lightweight aggregate intrinsic strength on lightweight concrete compressive strength and modulus of elasticity.” Mater. Constr., 52(265), 23–37. Zhang, M. H., and Malhotra, V. M. (1995). “Characteristics of a thermally activated alumino-silicate pozzolanic material and its use in concrete.” Cem. Concr. Res., 25(8), 1713–1725.

988 / JOURNAL OF MATERIALS IN CIVIL ENGINEERING © ASCE / AUGUST 2012

J. Mater. Civ. Eng., 2012, 24(8): 981-988