- Email: [email protected]
Effect of supplementary cementitious materials on autogenous shrinkage of ultra-high performance concrete
Construction and Building Materials 127 (2016) 43–48
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Effect of supplementary cementitious materials on autogenous shrinkage of ultra-high performance concrete Ehsan Ghafari a,⇑, Seyed Ali Ghahari a, Hugo Costa b, Eduardo Júlio c, Antonio Portugal d, Luisa Durães d a
Dep. of Civil Engineering, Purdue University, USA CERIS & Polytechnic Institute of Coimbra – ISEC, Coimbra, Portugal c CERIS, Instituto Superior Técnico, Universidade de Lisboa, Portugal d CIEPQPF & Dep. of Chemical Engineering, University of Coimbra, Portugal b
h i g h l i g h t s The amount of fine pores in UHPC can highly affect the autogenous shrinkage. A strong correlation between the autogenous shrinkage and the porosity of UHPC was established. Reducing the fine pores in FA/GGBS samples leads to a reduction of the autogenous shrinkage.
a r t i c l e
i n f o
Article history: Received 6 July 2016 Received in revised form 22 August 2016 Accepted 28 September 2016
Keywords: UHPC Autogenous shrinkage Porosity Cementitious materials
a b s t r a c t Ultra-high performance concrete (UHPC) not only presents ultra-high compressive strength but also exhibits ultra-high durability, due to its extremely dense structure and consequently highly reduced porosity. However, high dosages of silica fume (SF), typically adopted in UHPC, also lead to high autogenous shrinkage. This phenomenon, occurring at early ages, induces high internal stresses that, in turn, cause microcracking and increase permeability and, therefore, reduce the durability of concrete structures. The experimental study was conducted aiming to replace SF by another fine supplementary cementitious materials (SCMs), such as fly ash (FA) or ground granulated blast furnace slag (GGBS), in order to reduce the amount of autogenous shrinkage. The adopted approach involved partial or total replacement of SF by SCMs. Results indicate that the amount of fine pores in UHPC is a predominant factor that can highly affect the autogenous shrinkage. A strong correlation between the natural logarithm of autogenous shrinkage and the total porosity of UHPC mixtures was established. It was found that reducing the amount of fine pores in specimens containing FA or GGBS leads to a reduction of the autogenous shrinkage. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Ultra-high performance concrete (UHPC) concept was developed in the last decade. This concrete type presents not only ultra-high compressive strength but also ultra-high durability [1], because of its extremely dense structure and, thus, highly reduced porosity. To produce UHPC, it is mandatory to minimize the aggregate size and simultaneously to increase the paste/aggregate ratio [2]. Several types of fine size powders have been used as microfiller in this scope to meet the need for high degree of compactness and high compressive strength [3–5]. SF is known as a main constituent of a typical UHPC mixture and reactive powder concrete. ⇑ Corresponding author. E-mail address: [email protected] (E. Ghafari). http://dx.doi.org/10.1016/j.conbuildmat.2016.09.123 0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
It plays a significant role in improving both rheological and mechanical properties of UHPC. These prominent effects are divided into three main functions: i) filling effect, which improves the particle packing density; ii) lubricating effect, which results in an enhancement of rheological properties due to sphericity of particle shape; and iii) pozzolanic reaction, which leads to the production of additional C-S-H gel [2,6]. A wide-range of dosage from 10 up to 30% of SF in UHPC mixture has been reported in different investigations [7,8]. However, the optimum dosage has been recommended to be 25% in cement weight [2]. Besides of these advantages, SF presents some disadvantages, such as cost and scarcity if large quantities are requested. In addition, SF also presents some limitation in terms of aesthetic application due to its dark gray color. Although SF improves the rheology of concrete, the high specific surface of its particles results in an increased water
44
E. Ghafari et al. / Construction and Building Materials 127 (2016) 43–48
micro-fibers with 60 mm length and 0.15 mm diameter was also used. The mixing procedure includes the following steps: (1) First, in order to prevent agglomeration, and also to promote uniform distribution of the very fine particles, all powder and silica sand were mixed in dry state for 5 min at low speed; (2) Afterwards, water and the superplasticizer were added gradually in two steps; after 5 min, the mixtures became fluid; (3) Subsequently, fibers were added and additional mixing was applied for about 2 min at high speed; (4) After mixing, concrete was poured in a mold; and (5) 24 h later, specimens were removed from the mold. In order to investigate the effect of curing, half of the specimens were cured in water at 20 °C, and the remainders were cured at 90 °C and 95% relative humidity (RH) for 48 h, including 1-h ramp-up and ramp-down. Table 1 shows five different types of mixtures, where ‘SFS’ sample is a reference mixture that includes 24% of SF by weight of cement. This amount of SF was totally replaced by GGBS (GBS) and FA (FAS) in the mixtures and also partially replaced in binary blends of GGBS-SF (GBSF) and FA-SF (FASF). SF has a lower density than FA and GGBS. As a consequence, replacement of SF leads to a reduction of the paste volume. Hence, the amount of SF was replaced volumetrically, keeping the paste/aggregate ratio constant. Moreover, since both autogenous shrinkage and porosity are highly affected by water to binder (W/B) ratio, this ratio was also kept constant in volume. Keeping the aforementioned parameters constant enabled measuring the autogenous shrinkage, as well as porosity, in the same condition.
demand [9] and it affects the fluidity of the mixtures depending on carbon content [10]. In general, higher percentages of SF lead to higher dosages of superplasticizer and therefore, the mixtures become sticky [10,11]. For a low water to cement ratio (W/C) concrete, particularly in UHPC, SF presents the additional disadvantage of affecting durability. It is now well understood that autogenous shrinkage, caused by self-desiccation at early ages, induces high internal stresses that, in turn, provoke microcracking [12]. The latter increases permeability and, thus, reduces durability. It is also known that autogenous shrinkage, due to self-desiccation, is mostly related to fine pore structures [13]. Thus, mineral additions containing more fine pores are more susceptible to self-desiccation and consequently to autogenous shrinkage. Up to now, several researchers have reported high autogenous shrinkage of concrete specimens containing SF. Igarashi et al. [14] found that the presence of larger amounts of fine capillary pores in concrete containing SF is responsible for higher autogenous shrinkage at early ages. Mazloom et al. [15] performed an experimental study on the autogenous shrinkage of high strength concrete. The results indicated that, as the proportion of SF increased, the autogenous shrinkage also increased. Zhang et al. [16] reported two factors, SF content and W/C ratio, as having a significant effect on the autogenous shrinkage of concrete. As mentioned above, SF has been used in high percentage in UHPC typical mixtures. Therefore, autogenous shrinkage can be more critical in the case of UHPC. The study herein described focused on evaluating the potential use of fine supplementary cementitious materials (SCMs), such as FA and GGBS, as a replacement of SF. It is well known that the inclusion of SCMs in concrete mixtures enhances durability, decreases the heat of hydration, and generally improves concrete properties [17–19]. SCMs enhance concrete properties by two primary means. The first is by reaction with cement hydration products and the other by increasing particle packing efficiency. The effect of different SCMs as an alternative powder in the composition of reactive powder concrete and ultra high performance fiber reinforced concrete (UHPFRC) has already been studied. Rougeau et al. [10] investigated the mechanical properties as well as the durability of very high performance concrete and UHPC with some ultra-fine particles instead of SF, such as limestone microfiller, pulverized FA, and metakaolin, and the results pointed out that these ultrafine particles are potentially promising to produce UHPC. Tafraoui et al. [20] replaced SF by metakaolin and obtained an UHPC with almost equivalent mechanical performance. Yazici [21] reported that cement and SF content can be replaced by FA and/or GGBS keeping satisfactory mechanical properties. Based on the state of the art, the following two main objectives were defined for the present study: (1) to replace SF by other fine supplementary cementitious materials (SCMs), in order to reduce the amount of autogenous shrinkage, but without experiencing a significant reduction in UHPC mechanical properties, and (2) to establish the correlation between the pore structures and the autogenous shrinkage of UHPC, adopting the nitrogen gas adsorption technique to study the fine pore distribution of the specimens.
Experimental testing included three main parts: i) porosity assessment of the samples; ii) measurement of autogenous shrinkage; and iii) evaluation of UHPC mechanical properties. Measurements of cumulative pores’ volume and pores’ size distribution of UHPC specimens were performed using the nitrogen gas adsorption method, considered as the most appropriate test for evaluating fine pore structures [22]. Autogenous shrinkage can be calculated by two methods, namely linear and volumetric measurements. It has been proven that autogenous shrinkage values obtained with linear methods are considerably lower than those measured with volumetric methods [23]. In the scope of the study herein described, a vertical test setup was developed and used in order to measure the composite cement paste autogenous shrinkage. The fresh concrete was cast in the molds, sealed with plastic, and placed between two support heads. Moreover, a plastic sheet was embedded in the mold, in order to reduce the friction between concrete and the mold. The whole setup was then placed in a moisture room for 24 h. Therefore, the autogenous shrinkage was measured for all types of mixtures under isothermal conditions. The measurements started at early ages and shrinkage was recorded based on the dial gauge values. The temperature was kept constant in order to eliminate the effect of thermodynamic heat
2. Experimental
Table 1 Mixture proportions, in kg/dm3.
2.1. Materials and mixture proportions The UHPC mixtures were prepared with the following main constituents: Portland cement type 1 (52.5 R); SF; a new type of quartz flour (P600) used as a micro filler (particle size less than 10 lm); silica sand with maximum aggregate size of 0.6 mm; and polycarboxylate ether-based superplasticizers. In addition, FA class C and GGBS obtained by grinding method were used as binders to replace SF in the mixture of UHPC. A new type of steel
2.2. Experimental tests
Mixture
SFS
GBS
GBSF
FAS
FASF
Cement Sand SF FA GGBS Quartz Water SP Fibers
692 899.6 166.1 0 0 200.1 190.5 36 194
692 899.6 0 0 206.7 200.1 190.5 36 194
692 899.6 99.7 0 82.7 200.1 190.5 36 194
692 899.6 0 196 0 200.1 190.5 36 194
692 899.6 99.7 78.4 0 200.1 190.5 36 194
45
E. Ghafari et al. / Construction and Building Materials 127 (2016) 43–48
conduction on the autogenous shrinkage. The assessment of axial compressive strength was performed at 28 days of age, using cubic samples (40 40 40 mm3), in accordance with the standard BS EN 196–1 [24]. The tests were performed on four specimens and the average values were considered. 3. Results and discussion 3.1. Pore size distribution of the powders Barrett-Joyner-Halenda (BJH) analysis was performed to measure the pore structure of the SCM powders [25]. Fig. 1 represents the pore size distribution of the SCMs powders. In addition, the surface area of the powders was measured using the BET (Brunauer Emmett Teller) method [26]. Table 2 shows the pore volume and the specific surface area of the materials. As shown, SF has the highest pore volume and also exhibits a significant larger pore size when compared with other SCMs. The highest pore volume intensity was obtained approximately at 70 Å. This is in agreement with its large surface area that causes high water demand [27]. According to the results, FA powder showed the lowest amount of porosity. Furthermore, the highest pore volume value occurs for a pore size of 30 Å and his material mostly has micropores (size < 20 Å). Previous studies show that the porosity of the powders highly influences the water demand as well as superplasticizers adsorption [28]. 3.2. Pore size distribution of the UHPC samples In order to measure the pore size distribution of UHPC samples, the gas adsorption technique was used as in Section 3.1 [29]. The critical pore size was also determined for further discussion about the pore size distribution. The critical pore size is the width associated with the highest change in the rate of the N2 adsorption. Above this point, the interior pore structure of the specimen is occupied by the gas. It is the diameter that mostly occurs in the
Fig. 1. Pore size distribution of the SCMs.
Fig. 2. Volume adsorbed vs. relative pressure for SFS, GBS, and FAS.
interconnected pores and the maximum percolation of ions and chemical particles through the cement paste happens through such pores [30]. Adsorption isotherm, as shown in Fig. 2, is obtained by plotting the volume of the gas that is adsorbed (cc/g) over a wide range of relative pressures, i.e. specimen pressure (P) over the vapor saturation pressure (P0). The relative pressure is important to be considered to decrease the sensitivity of the technique to changes in pressure due to temperature small variations. The adsorption isotherm for SFS, FAS and GBS samples are presented in Fig. 2. The results indicate that all adsorption isotherm curves follow the type III adsorption isotherm, in which the adsorbed gas increases without limit when the relative saturation point gets closer to unity. This indicates the formation of multilayers in all specimens and, since no plateau is observable, no monolayer has been formed. However, SFS showed a fast increase in the gas adsorption when compared to GBS and FAS. This indicates that the pore size of SFS is relatively smaller than that of GBS and FAS. According to Fig. 2, the specimen containing SF showed a significant higher porosity compared with the specimen containing GGBS and FA. The pore parameters of all mixtures were calculated and are presented in Table 3. The sample containing SF exhibits the highest amount of porosity in which the high specific surface area confirms the formation of the small pores in the sample. It can be seen that the incorporation of FA and GGBS at the highest level of SF replacement results in a significant decrease in porosity. The lowest pore surface area was obtained for FAS indicating the formation of larger pore structure. The pore size distribution curves, obtained for SFS, GBS, and FAS with the nitrogen gas adsorption test, are shown in Fig. 3. The obtained curves cover the pore size distribution range from 10 Å to 1000 Å. The pore size distributions show the existence of much finer pores distribution in the range of 10–100 Å for UHPC specimens containing SF. From the cumulative pore volume, the
Table 3 Pore parameters of all UHPC samples.
Table 2 Specific surface area and pore volume of SCMs. Materials
FA
GGBS
SF
Specific surface area (m2/g) Pore volume (cc/g)
1.42 0.009
2.31 0.012
9.66 0.128
Sample
Specific surface (m2/g)
Mean diameter (nm)
Porosity (%)
SFS GBS FAS GBSF FASF
9.5 6.7 3.4 7.1 3.9
33.2 39.8 49.3 38.1 46.1
17.01 12.8 11.12 16.95 13.23
46
E. Ghafari et al. / Construction and Building Materials 127 (2016) 43–48
Fig. 3. Pore size distribution for SFS, GBS, and FAS.
critical pore size (dcr) can be obtained. According to Fig. 3, all the mixtures had a maximum peak of pore volume in the range of 20–50 Å, although SFS showed a slight decrease in the value of dcr indicating the finer pore structure. In this critical interval, SFS has higher cumulative volume of the fine pores, whereas FAS had the lowest amount of pore volume in the mentioned range. The critical pore size of FAS is found to be around 40 Å while this value for GBS is about 30 Å. Thus, these results indicate that the FA and GBS contain larger pores when compared to SFS. 3.3. Autogenous shrinkage The measured autogenous shrinkage is shown in Fig. 4 for all mixtures. Since the W/B ratio is quite reduced in UHPFRC, this concrete is more susceptible to self-desiccation and consequently to autogenous shrinkage. In fact, the capillary pores contain air with a very small amount of water. The water/air menisci in capillary pores develop a pressure which results in a significant volume contraction. Therefore, as shown in Fig. 4, all UHPC specimens exhibit very fast shrinkage. As expected, results show very high autogenous shrinkage for mixtures containing only SF. It has been reported that the induced pressure by water/air menisci is inver-
sely proportional to the size of the capillary pores [31]. Hence, the amount of autogenous shrinkage induced by self-desiccation is more pronounced in the sample containing SF due to the finer pore structure. The results of this study corroborate the results of previous studies [13,32]. Also, the sample containing SF exhibits a significant amount of autogenous shrinkage at early ages. In fact, the addition of SF can accelerate the hydration process of tricalcium silicate (C3S) clinker phase due to the large and highly reactive surface of particles leading to the formation of more calcium silicate hydrate (C-S-H) at early ages [27]. The internal porosity of C-S-H gel, the so-called ‘gel pores’, ranges from 5 Å to 100 Å. Hence, in theory, the SFS is expected to have more fine pores. Consequently, the fine pores lead to a larger capillary pressure so that the autogenous shrinkage induced by self-desiccation will be accelerated. The results regarding the amount of fine pores obtained by nitrogen gas adsorption method are in good agreement with this theory. As shown in Fig. 4, the autogenous shrinkage was reduced when 50% of SF was replaced by FA and GGBS. However, when the replacement ratio reached 100%, the autogenous shrinkage became remarkably small. The obtained values yielded the lowest and approximately equal shrinkage for FAS and GBS. Due to intimate relation between the amount of fine pores and autogenous shrinkage, it can be expected that reducing the amount of fine pores in FAS and GBS specimens leads to a reduction in the autogenous shrinkage. The higher amount of large pores in FAS and GBS results in a smaller capillary pressure in the paste when water is consumed by the hydration of cement [33,34]. In addition, the autogenous shrinkage was drastically reduced at early ages for FAS specimens. This clearly indicates that the onset of the shrinkage is highly affected by the material type. The addition of FA can delay the initial hydration phase so that less C-S-H will be formed and, as a consequence, the amount of fine pores will be reduced. As it can be seen, the addition of GGBS was also found to be effective in reducing the autogenous shrinkage. This might be due to the critical range of the fine pores which contributes to the autogenous shrinkage. It has been reported that 50–500 Å is a critical pore size which affects the autogenous shrinkage [35]. As presented in Fig. 3, GBS exhibits more fine pores in the range of 50–500 Å (compared to SFS) that resulted in a lower autogenous shrinkage. 3.3.1. Prediction model for UHPC Most of the developed models for predicting the autogenous shrinkage of current concrete are not applicable for UHPC due to the fact that these take into account the effect of coarse aggregates. In addition, the effect of supplementary cementitious materials has not been well addressed in these models. In this scope, a prediction model was developed to establish the correlation between the pore structure of UHPC and the autogenous shrinkage, based on the method proposed by Li et al. [35] for the prediction of autogenous shrinkage of cement paste. In this method, a polynomial equation was adopted to predict the autogenous shrinkage, which can be expressed as (Eq. (1)):
ea ðtÞ ¼ at3 þ bt2 þ ct þ d
Fig. 4. Autogenous shrinkage of specimens.
ð1Þ
where ea (t) is the autogenous shrinkage of cement paste at the age t (in days), and a, b, c and d are the constants of the prediction model equation. The autogenous shrinkage regression equations along with the correlation coefficients of each equation are presented in Table 4. As presented, the high value of correlation coefficient corroborated the adequacy of the model used to predict the autogenous shrinkage of UHPC specimens containing different SCMs. Once the prediction equation was obtained, all the attempts were focused on establishing the relationship between the obtained fine pores and
47
E. Ghafari et al. / Construction and Building Materials 127 (2016) 43–48 Table 4 The equations for predicting the autogenous shrinkage of UHPC samples. Mixture
a (lm/(m day3))
b (lm/(m day2))
c (lm/(m day))
d (lm/m)
Correlation coefficient
GBS FAS GBSF FASF SFS
0.0335 0.0366 0.0311 0.0302 0.0343
2.0704 2.1433 1.8015 1.5464 2.117
47.365 47.796 41.183 35.663 47.967
35.189 17.311 24.809 6.0488 30.981
0.99 0.99 0.99 0.98 0.97
Fig. 5. The relationship between the porosity and the autogenous shrinkage at 28 days.
the autogenous shrinkage of UHPC. In this scope, a numerical regression method based on the previous literature was adopted which is expressed as follows [35]:
P ¼ A þ lnðea Þ
ð2Þ
where, ea is the autogenous shrinkage at the age of 28 days, p is the porosity (%), and A is constant to improve the prediction accuracy. The total porosity of the UHPC mixtures along with the autogenous shrinkage dates were used as an input to Eq. (2). Fig. 5 shows the relationship between the autogenous shrinkage value and the total porosity of the UHPC mixtures. A correlation coefficient R2 = 0.97 was found, revealing an excellent linear correlation between the natural logarithm of autogenous shrinkage and the total porosity of the UHPC samples. The results indicate that the amount of fine pores in UHPC is a predominant factor that can highly affect the autogenous shrinkage. Recalling the discussion regarding the fine pore distribution and pore volume in Fig. 3, it can be concluded that the combination of these two parameters accelerated the self-desiccation process, resulting in a high autogenous shrinkage. 3.4. Compressive strength Results for compressive strength in both steam-cured and water-cured conditions are shown in Fig. 6. Regardless of the mixture proportion, all the specimens that were submitted to the former curing condition exhibited higher compressive strength, compared to those maintained in water. In the latter case, no significant difference was observed in the compressive strength of all mixtures. However, the compressive strength of SF specimen exposed to steam was approximately 20% higher than that of the other specimens that were cured with steam. This implies that
Fig. 6. Compressive strength of UHPC samples containing different SCMs in different curing conditions at 28 days.
steam curing could accelerate the reaction between the mineral admixtures and cement particles. There is a synergistic effect between the water adsorption of the mineral admixtures and the pozzolanic reactivity. SF in the UHPC mixture showed the best results in terms of compressive strength which would be due to the availability of higher amount of portlandite that accelerates the pozzolanic reaction of the particle [36,37]. FAS reached a compressive strength of 150 MPa when cured with steam, which was the lowest value of compressive strength among the specimens. Constantinides et al. [38] have performed extensive research on the low density (LD) C-S-H and high density (HD) C-S-H. Their research has shown that LD C-S-H covers the surface of the cement and FA, and that HD C-S-H penetrates into the hydration particles. The low amount of compressive strength obtained by FAS specimen denotes that FA could not react well within the UHPC mixture, although a significant pozzolanic reaction was expected. However, further research is required for proving the fact that FA would lead to significant lower mechanical properties, since the cement hydration is considerably faster than the FA pozzolanic reaction [39,40]. Compared with the specimens cured in water, SFS showed 21% increase in the compressive strength when cured with steam, recorded as the highest increase among the compressive strength values. This could be due to the high pozzolanic reactivity of SF that has led to a higher formation of C-S-H gel and crystalline CH [41]. The compressive strength of GBS cured in water was 135 MPa, the lowest compressive strength among the specimens cured in water. This can be attributed to that fact that UHPC contains an amount of water that is consumed by the powders in the mixture. If a mineral admixture adsorbs a considerable amount of water, some amount of cement powder would remain intact and less C-S-H gel would be produced accordingly. GGBS is known by its ability in reducing the pore size and the cumulative pore volume of the paste [42]. Therefore, more water might have been
48
E. Ghafari et al. / Construction and Building Materials 127 (2016) 43–48
adsorbed by GGBS compared with other admixtures leading to lower compressive strength values. It is worth mentioning that the incorporation of both FA and GGBS increased the fluidity of the mixtures with the same W/B ratio. As mentioned before, the adopted strategy aimed at measuring the autogenous shrinkage of all mixtures in the same conditions. Therefore, it was expected that the compressive strength of the specimens with GGBS and FA would have increased with the reduction in the water content by designing the mixtures to have the same consistency instead of the same water content. Finally, the FASF and GBSF specimens showed insignificant difference in the mechanical properties when cured in water and cured with steam. 4. Conclusions In this study, the ability of SCMs, such as FA and ground granulated blast furnace slag, to be used as replacement of SF in UHPC was evaluated. The addition of these was found to be effective in reducing the autogenous shrinkage of UHPC. Results also indicate that the amount of fine pores in UHPC is a predominant factor that can highly affect the autogenous shrinkage. A strong linear correlation between the natural logarithm of the latter and the total porosity of the UHPC mixtures was established. It was found that reducing the amount of fine pores in FAS and GBS leads to a reduction in the autogenous shrinkage. However, it was also concluded that the incorporation of FA and GGBS gives similar performance in compressive strength when compared to the reference mixture (containing only SF). In particular, it was observed that although incorporating FA into the mixtures results in a slight decrease of the mechanical properties, much lower autogenous shrinkage can be achieved. Acknowledgements This work has been supported by the Portuguese Foundation for Science and Technology (FCT) under Grants SFRH/BD/91402/2012 and the Project PTDC/ECM/098497/2008 entitled ‘‘Intelligent Super Skin (ISS)”. The authors wish to thanks SIKA company for their support and partnerships in the execution of this project. References [1] E. Ghafari, M. Arezoumandi, H. Costa, E. Júlio, Influence of nano-silica addition on durability of UHPC, Constr. Build. Mater. 94 (2015) 181–188. [2] P. Richard, M. Cheyrezy, Composition of reactive powder concrete, Cem. Concr. Res. 25 (7) (1995) 1501–1511. [3] M. Heikal, S. Abd El-Aleem, W.M. Morsi, Characteristics of blended cements containing nano-silica, HBRC J. 9 (3) (2013) 243–255. [4] S. Abd El-Aleem, M. Heikal, W.M. Morsi, Hydration characteristic, thermal expansion and microstructure of cement containing nano-silica, Constr. Build. Mater. 59 (2014) 151–160. [5] A. Abdelaziz, Saleh Abd El-Aleem, Wagih M. Menshawy, Effect of fine materials in local quarry dusts of limestone and basalt on the properties of Portland cement pastes and mortars, Int. J. Eng. Res. Technol. (IJERT) 3 (6) (2014) 1038– 1056. [6] S. Ghahari, A.M. Ramezanianpour, A.A. Ramezanianpour, M. Esmaeili, An accelerated test method of simultaneous carbonation and chloride ion ingress: durability of silica fume concrete in severe environments, Adv. Mater. Sci. Eng. 2016 (2016). [7] First International Symposium on Ultra High Performance Concrete September 2004 Kassel, Germany. [8] Second International Symposium on Ultra High Performance Concrete March 2008 Kassel, Germany. [9] T.C. Holland, Silica Fume User’s Manual. 2005, Federal Highway Administration (FHWA). [10] P. Rougeau, B. Borys, Ultra High Performance Concrete with ultrafine particles other than silica fume in International Symposium on Ultra High Performance Concrete 2004 Kassel, Germany. [11] M.I. Khan, Properties of High Performance Concrete, King Saud University, 2006. [12] P. Lura, Ole Mejlhede Jensen, Jason Weiss, Cracking in cement paste induced by autogenous shrinkage, Mater. Struct. 42 (8) (2009) 1089–1099.
[13] M.S. Meddah, Tagnit-Hamou Arezki, Pore structure of concrete with mineral admixtures and its effect on self-desiccation shrinkage, ACI Mater. J. (Jun 2009). [14] A.W. Shin-ichi Igarashi, Mitsunori Kawamura, Evaluation of capillary pore size characteristics in high-strength concrete at early ages, Cem. Concr. Res. 35 (2005) 513–519. [15] M. Mazloom, A.A. Ramezanianpour, J.J. Brooks, Effect of silica fume on mechanical properties of high-strength concrete, Cement Concr. Compos. 26 (4) (2004) 347–357. [16] M.H. Zhang, C.T. Tam, M.P. Leow, Effect of water-to-cementitious materials ratio and silica fume on the autogenous shrinkage of concrete, Cem. Concr. Res. 33 (10) (2003) 1687–1694. [17] F. Robert, M.W.M. Wescott, Martha G. VanGeem, John Gajda, Prospects for Expanding the Use of Supplementary Cementitious Materials in California, February 2010. [18] M.C.G. Juenger, Siddique Rafat, Recent advances in understanding the role of supplementary cementitious materials in concrete, Cem. Concr. Res. 78 (2015) 71–80. [19] Z. Li, Drying shrinkage prediction of paste containing meta-kaolin and ultrafine fly ash for developing ultra-high performance concrete, Mater. Today Commun. 6 (2016) 74–80. [20] A. Tafraoui, G. Escadeillas, S. Lebaili, T. Vidal, Metakaolin in the formulation of UHPC, Constr. Build. Mater. 23 (2009) 669–674. [21] H. Yazıcı, H. Yig˘iter, A.S ß. Karabulut, B. Baradan, Utilization of fly ash and ground granulated blast furnace slag as an alternative silica source in reactive powder concrete, Fuel 87 (12) (2008) 2401–2407. [22] G. Herold, H.S. Müller, Measurement of porosity of Ultra High Strength Fibre Reinforced Concrete in International Symposium on Ultra High Performance Concrete September 13–15, 2004, Kassel, Germany. [23] S.N. Lim, T.H. Wee, Autogenous shrinkage of ground-granulated blast-furnace slag concrete, Mater. J. 97 (5) (2000) 587–593. [24] EN196-1, Methods of testing cement-Part 1: Determination of strength, 2005. [25] E.P. Barrett, L.G. Joyner, P.P. Halenda, The determination of pore volume and area distributions in porous substances. I. Computations from nitrogen isotherms, J. Am. Chem. Soc. 73 (1) (1951) 373–380. [26] G. Fagerlund, Determination of specific surface by the BET method, Matériaux Constr. 6 (3) (1973) 239–245. [27] H. Madani, A. Bagheri, T. Parhizkar, The pozzolanic reactivity of monodispersed nanosilica hydrosols and their influence on the hydration characteristics of Portland cement, Cem. Concr. Res. 42 (12) (2012) 1563–1570. [28] E. Ghafari, A. Baig, K. Nicoletta, D. Feys, R.D. Ferron, K.H. Khayat, Admixture Compatibility of Alternative Supplementary Cementitious Materials for Pavement and Structural Concrete. 2014, Center for Transportation Infrastructure and Safety/NUTC program: Published. In: The National Academies of Sciences Engineering and Medicine. [29] C.G. Schull, The determination of pore size distribution from gas adsorption data, J. Am. Chem. Soc. 70 (4) (1948) 1405–1410. [30] D.N. Winslow, S. Diamond, A mercury porosimetry study of the porosity in Portland cement, J. Mater. (1970). [31] D. Saje, Reduction of the early autogenous shrinkage of high strength concrete, Adv. Mater. Sci. Eng. 2015 (2015). [32] J.D. Jianxin Ma, Ultra High Performance Self Compacting Concrete LACER, 2002, p. 7. [33] S. Tangtermsirikul, Effect of chemical composition and particle size of fly ash on autogenous shrinkage of paste, in: Proceedings of an International Workshop on the Autogenous Shrinkage of Concrete, E & FN Spon, London, 1999. [34] P. Termkhajornkit, T. Nawa, M. Nakai, T. Saito, Effect of fly ash on autogenous shrinkage, Cem. Concr. Res. 35 (3) (2005) 473–482. [35] Y. Li, J. Bao, Y. Guo, The relationship between autogenous shrinkage and pore structure of cement paste with mineral admixtures, Constr. Build. Mater. 24 (10) (2010) 1855–1860. [36] E. Ghafari, H. Costa, E. Júlio, A. Portugal, L. Durães, The effect of nanosilica addition on flowability, strength and transport properties of ultra high performance concrete, Mater. Des. 59 (2014) 1–9. [37] E. Ghafari, H. Costa, E. Júlio, Critical review on eco-efficient ultra high performance concrete enhanced with nano-materials, Constr. Build. Mater. 101 (2015) 201–208. [38] G. Constantinides, F.-J. Ulm, The nanogranular nature of C–S–H, J. Mech. Phys. Solids 55 (1) (2007) 64–90. [39] V.G. Papadakis, Effect of fly ash on Portland cement systems: Part I. Low-calcium fly ash, Cem. Concr. Res. 29 (11) (1999) 1727–1736. [40] S. Hanehara et al., Effects of water/powder ratio, mixing ratio of fly ash, and curing temperature on pozzolanic reaction of fly ash in cement paste, Cem. Concr. Res. 31 (1) (2001) 31–39. [41] A.M. Ramezanianpour, Kh. Esmaeili, S.A. Ghahari, A.A. Ramezanianpour, Influence of initial steam curing and different types of mineral additives on mechanical and durability properties of self-compacting concrete, Constr. Build. Mater. 73 (2014) 187–194. [42] P.A.M. Basheer, P.R.V. Gilleece, A.E. Long, W.J. Mc Carter, Monitoring electrical resistance of concretes containing alternative cementitious materials to assess their resistance to chloride penetration, Cement Concr. Compos. 24 (5) (2002) 437–449.
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Effect of supplementary cementitious materials on autogenous shrinkage of ultra-high performance concrete Ehsan Ghafari a,⇑, Seyed Ali Ghahari a, Hugo Costa b, Eduardo Júlio c, Antonio Portugal d, Luisa Durães d a
Dep. of Civil Engineering, Purdue University, USA CERIS & Polytechnic Institute of Coimbra – ISEC, Coimbra, Portugal c CERIS, Instituto Superior Técnico, Universidade de Lisboa, Portugal d CIEPQPF & Dep. of Chemical Engineering, University of Coimbra, Portugal b
h i g h l i g h t s The amount of fine pores in UHPC can highly affect the autogenous shrinkage. A strong correlation between the autogenous shrinkage and the porosity of UHPC was established. Reducing the fine pores in FA/GGBS samples leads to a reduction of the autogenous shrinkage.
a r t i c l e
i n f o
Article history: Received 6 July 2016 Received in revised form 22 August 2016 Accepted 28 September 2016
Keywords: UHPC Autogenous shrinkage Porosity Cementitious materials
a b s t r a c t Ultra-high performance concrete (UHPC) not only presents ultra-high compressive strength but also exhibits ultra-high durability, due to its extremely dense structure and consequently highly reduced porosity. However, high dosages of silica fume (SF), typically adopted in UHPC, also lead to high autogenous shrinkage. This phenomenon, occurring at early ages, induces high internal stresses that, in turn, cause microcracking and increase permeability and, therefore, reduce the durability of concrete structures. The experimental study was conducted aiming to replace SF by another fine supplementary cementitious materials (SCMs), such as fly ash (FA) or ground granulated blast furnace slag (GGBS), in order to reduce the amount of autogenous shrinkage. The adopted approach involved partial or total replacement of SF by SCMs. Results indicate that the amount of fine pores in UHPC is a predominant factor that can highly affect the autogenous shrinkage. A strong correlation between the natural logarithm of autogenous shrinkage and the total porosity of UHPC mixtures was established. It was found that reducing the amount of fine pores in specimens containing FA or GGBS leads to a reduction of the autogenous shrinkage. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Ultra-high performance concrete (UHPC) concept was developed in the last decade. This concrete type presents not only ultra-high compressive strength but also ultra-high durability [1], because of its extremely dense structure and, thus, highly reduced porosity. To produce UHPC, it is mandatory to minimize the aggregate size and simultaneously to increase the paste/aggregate ratio [2]. Several types of fine size powders have been used as microfiller in this scope to meet the need for high degree of compactness and high compressive strength [3–5]. SF is known as a main constituent of a typical UHPC mixture and reactive powder concrete. ⇑ Corresponding author. E-mail address: [email protected] (E. Ghafari). http://dx.doi.org/10.1016/j.conbuildmat.2016.09.123 0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
It plays a significant role in improving both rheological and mechanical properties of UHPC. These prominent effects are divided into three main functions: i) filling effect, which improves the particle packing density; ii) lubricating effect, which results in an enhancement of rheological properties due to sphericity of particle shape; and iii) pozzolanic reaction, which leads to the production of additional C-S-H gel [2,6]. A wide-range of dosage from 10 up to 30% of SF in UHPC mixture has been reported in different investigations [7,8]. However, the optimum dosage has been recommended to be 25% in cement weight [2]. Besides of these advantages, SF presents some disadvantages, such as cost and scarcity if large quantities are requested. In addition, SF also presents some limitation in terms of aesthetic application due to its dark gray color. Although SF improves the rheology of concrete, the high specific surface of its particles results in an increased water
44
E. Ghafari et al. / Construction and Building Materials 127 (2016) 43–48
micro-fibers with 60 mm length and 0.15 mm diameter was also used. The mixing procedure includes the following steps: (1) First, in order to prevent agglomeration, and also to promote uniform distribution of the very fine particles, all powder and silica sand were mixed in dry state for 5 min at low speed; (2) Afterwards, water and the superplasticizer were added gradually in two steps; after 5 min, the mixtures became fluid; (3) Subsequently, fibers were added and additional mixing was applied for about 2 min at high speed; (4) After mixing, concrete was poured in a mold; and (5) 24 h later, specimens were removed from the mold. In order to investigate the effect of curing, half of the specimens were cured in water at 20 °C, and the remainders were cured at 90 °C and 95% relative humidity (RH) for 48 h, including 1-h ramp-up and ramp-down. Table 1 shows five different types of mixtures, where ‘SFS’ sample is a reference mixture that includes 24% of SF by weight of cement. This amount of SF was totally replaced by GGBS (GBS) and FA (FAS) in the mixtures and also partially replaced in binary blends of GGBS-SF (GBSF) and FA-SF (FASF). SF has a lower density than FA and GGBS. As a consequence, replacement of SF leads to a reduction of the paste volume. Hence, the amount of SF was replaced volumetrically, keeping the paste/aggregate ratio constant. Moreover, since both autogenous shrinkage and porosity are highly affected by water to binder (W/B) ratio, this ratio was also kept constant in volume. Keeping the aforementioned parameters constant enabled measuring the autogenous shrinkage, as well as porosity, in the same condition.
demand [9] and it affects the fluidity of the mixtures depending on carbon content [10]. In general, higher percentages of SF lead to higher dosages of superplasticizer and therefore, the mixtures become sticky [10,11]. For a low water to cement ratio (W/C) concrete, particularly in UHPC, SF presents the additional disadvantage of affecting durability. It is now well understood that autogenous shrinkage, caused by self-desiccation at early ages, induces high internal stresses that, in turn, provoke microcracking [12]. The latter increases permeability and, thus, reduces durability. It is also known that autogenous shrinkage, due to self-desiccation, is mostly related to fine pore structures [13]. Thus, mineral additions containing more fine pores are more susceptible to self-desiccation and consequently to autogenous shrinkage. Up to now, several researchers have reported high autogenous shrinkage of concrete specimens containing SF. Igarashi et al. [14] found that the presence of larger amounts of fine capillary pores in concrete containing SF is responsible for higher autogenous shrinkage at early ages. Mazloom et al. [15] performed an experimental study on the autogenous shrinkage of high strength concrete. The results indicated that, as the proportion of SF increased, the autogenous shrinkage also increased. Zhang et al. [16] reported two factors, SF content and W/C ratio, as having a significant effect on the autogenous shrinkage of concrete. As mentioned above, SF has been used in high percentage in UHPC typical mixtures. Therefore, autogenous shrinkage can be more critical in the case of UHPC. The study herein described focused on evaluating the potential use of fine supplementary cementitious materials (SCMs), such as FA and GGBS, as a replacement of SF. It is well known that the inclusion of SCMs in concrete mixtures enhances durability, decreases the heat of hydration, and generally improves concrete properties [17–19]. SCMs enhance concrete properties by two primary means. The first is by reaction with cement hydration products and the other by increasing particle packing efficiency. The effect of different SCMs as an alternative powder in the composition of reactive powder concrete and ultra high performance fiber reinforced concrete (UHPFRC) has already been studied. Rougeau et al. [10] investigated the mechanical properties as well as the durability of very high performance concrete and UHPC with some ultra-fine particles instead of SF, such as limestone microfiller, pulverized FA, and metakaolin, and the results pointed out that these ultrafine particles are potentially promising to produce UHPC. Tafraoui et al. [20] replaced SF by metakaolin and obtained an UHPC with almost equivalent mechanical performance. Yazici [21] reported that cement and SF content can be replaced by FA and/or GGBS keeping satisfactory mechanical properties. Based on the state of the art, the following two main objectives were defined for the present study: (1) to replace SF by other fine supplementary cementitious materials (SCMs), in order to reduce the amount of autogenous shrinkage, but without experiencing a significant reduction in UHPC mechanical properties, and (2) to establish the correlation between the pore structures and the autogenous shrinkage of UHPC, adopting the nitrogen gas adsorption technique to study the fine pore distribution of the specimens.
Experimental testing included three main parts: i) porosity assessment of the samples; ii) measurement of autogenous shrinkage; and iii) evaluation of UHPC mechanical properties. Measurements of cumulative pores’ volume and pores’ size distribution of UHPC specimens were performed using the nitrogen gas adsorption method, considered as the most appropriate test for evaluating fine pore structures [22]. Autogenous shrinkage can be calculated by two methods, namely linear and volumetric measurements. It has been proven that autogenous shrinkage values obtained with linear methods are considerably lower than those measured with volumetric methods [23]. In the scope of the study herein described, a vertical test setup was developed and used in order to measure the composite cement paste autogenous shrinkage. The fresh concrete was cast in the molds, sealed with plastic, and placed between two support heads. Moreover, a plastic sheet was embedded in the mold, in order to reduce the friction between concrete and the mold. The whole setup was then placed in a moisture room for 24 h. Therefore, the autogenous shrinkage was measured for all types of mixtures under isothermal conditions. The measurements started at early ages and shrinkage was recorded based on the dial gauge values. The temperature was kept constant in order to eliminate the effect of thermodynamic heat
2. Experimental
Table 1 Mixture proportions, in kg/dm3.
2.1. Materials and mixture proportions The UHPC mixtures were prepared with the following main constituents: Portland cement type 1 (52.5 R); SF; a new type of quartz flour (P600) used as a micro filler (particle size less than 10 lm); silica sand with maximum aggregate size of 0.6 mm; and polycarboxylate ether-based superplasticizers. In addition, FA class C and GGBS obtained by grinding method were used as binders to replace SF in the mixture of UHPC. A new type of steel
2.2. Experimental tests
Mixture
SFS
GBS
GBSF
FAS
FASF
Cement Sand SF FA GGBS Quartz Water SP Fibers
692 899.6 166.1 0 0 200.1 190.5 36 194
692 899.6 0 0 206.7 200.1 190.5 36 194
692 899.6 99.7 0 82.7 200.1 190.5 36 194
692 899.6 0 196 0 200.1 190.5 36 194
692 899.6 99.7 78.4 0 200.1 190.5 36 194
45
E. Ghafari et al. / Construction and Building Materials 127 (2016) 43–48
conduction on the autogenous shrinkage. The assessment of axial compressive strength was performed at 28 days of age, using cubic samples (40 40 40 mm3), in accordance with the standard BS EN 196–1 [24]. The tests were performed on four specimens and the average values were considered. 3. Results and discussion 3.1. Pore size distribution of the powders Barrett-Joyner-Halenda (BJH) analysis was performed to measure the pore structure of the SCM powders [25]. Fig. 1 represents the pore size distribution of the SCMs powders. In addition, the surface area of the powders was measured using the BET (Brunauer Emmett Teller) method [26]. Table 2 shows the pore volume and the specific surface area of the materials. As shown, SF has the highest pore volume and also exhibits a significant larger pore size when compared with other SCMs. The highest pore volume intensity was obtained approximately at 70 Å. This is in agreement with its large surface area that causes high water demand [27]. According to the results, FA powder showed the lowest amount of porosity. Furthermore, the highest pore volume value occurs for a pore size of 30 Å and his material mostly has micropores (size < 20 Å). Previous studies show that the porosity of the powders highly influences the water demand as well as superplasticizers adsorption [28]. 3.2. Pore size distribution of the UHPC samples In order to measure the pore size distribution of UHPC samples, the gas adsorption technique was used as in Section 3.1 [29]. The critical pore size was also determined for further discussion about the pore size distribution. The critical pore size is the width associated with the highest change in the rate of the N2 adsorption. Above this point, the interior pore structure of the specimen is occupied by the gas. It is the diameter that mostly occurs in the
Fig. 1. Pore size distribution of the SCMs.
Fig. 2. Volume adsorbed vs. relative pressure for SFS, GBS, and FAS.
interconnected pores and the maximum percolation of ions and chemical particles through the cement paste happens through such pores [30]. Adsorption isotherm, as shown in Fig. 2, is obtained by plotting the volume of the gas that is adsorbed (cc/g) over a wide range of relative pressures, i.e. specimen pressure (P) over the vapor saturation pressure (P0). The relative pressure is important to be considered to decrease the sensitivity of the technique to changes in pressure due to temperature small variations. The adsorption isotherm for SFS, FAS and GBS samples are presented in Fig. 2. The results indicate that all adsorption isotherm curves follow the type III adsorption isotherm, in which the adsorbed gas increases without limit when the relative saturation point gets closer to unity. This indicates the formation of multilayers in all specimens and, since no plateau is observable, no monolayer has been formed. However, SFS showed a fast increase in the gas adsorption when compared to GBS and FAS. This indicates that the pore size of SFS is relatively smaller than that of GBS and FAS. According to Fig. 2, the specimen containing SF showed a significant higher porosity compared with the specimen containing GGBS and FA. The pore parameters of all mixtures were calculated and are presented in Table 3. The sample containing SF exhibits the highest amount of porosity in which the high specific surface area confirms the formation of the small pores in the sample. It can be seen that the incorporation of FA and GGBS at the highest level of SF replacement results in a significant decrease in porosity. The lowest pore surface area was obtained for FAS indicating the formation of larger pore structure. The pore size distribution curves, obtained for SFS, GBS, and FAS with the nitrogen gas adsorption test, are shown in Fig. 3. The obtained curves cover the pore size distribution range from 10 Å to 1000 Å. The pore size distributions show the existence of much finer pores distribution in the range of 10–100 Å for UHPC specimens containing SF. From the cumulative pore volume, the
Table 3 Pore parameters of all UHPC samples.
Table 2 Specific surface area and pore volume of SCMs. Materials
FA
GGBS
SF
Specific surface area (m2/g) Pore volume (cc/g)
1.42 0.009
2.31 0.012
9.66 0.128
Sample
Specific surface (m2/g)
Mean diameter (nm)
Porosity (%)
SFS GBS FAS GBSF FASF
9.5 6.7 3.4 7.1 3.9
33.2 39.8 49.3 38.1 46.1
17.01 12.8 11.12 16.95 13.23
46
E. Ghafari et al. / Construction and Building Materials 127 (2016) 43–48
Fig. 3. Pore size distribution for SFS, GBS, and FAS.
critical pore size (dcr) can be obtained. According to Fig. 3, all the mixtures had a maximum peak of pore volume in the range of 20–50 Å, although SFS showed a slight decrease in the value of dcr indicating the finer pore structure. In this critical interval, SFS has higher cumulative volume of the fine pores, whereas FAS had the lowest amount of pore volume in the mentioned range. The critical pore size of FAS is found to be around 40 Å while this value for GBS is about 30 Å. Thus, these results indicate that the FA and GBS contain larger pores when compared to SFS. 3.3. Autogenous shrinkage The measured autogenous shrinkage is shown in Fig. 4 for all mixtures. Since the W/B ratio is quite reduced in UHPFRC, this concrete is more susceptible to self-desiccation and consequently to autogenous shrinkage. In fact, the capillary pores contain air with a very small amount of water. The water/air menisci in capillary pores develop a pressure which results in a significant volume contraction. Therefore, as shown in Fig. 4, all UHPC specimens exhibit very fast shrinkage. As expected, results show very high autogenous shrinkage for mixtures containing only SF. It has been reported that the induced pressure by water/air menisci is inver-
sely proportional to the size of the capillary pores [31]. Hence, the amount of autogenous shrinkage induced by self-desiccation is more pronounced in the sample containing SF due to the finer pore structure. The results of this study corroborate the results of previous studies [13,32]. Also, the sample containing SF exhibits a significant amount of autogenous shrinkage at early ages. In fact, the addition of SF can accelerate the hydration process of tricalcium silicate (C3S) clinker phase due to the large and highly reactive surface of particles leading to the formation of more calcium silicate hydrate (C-S-H) at early ages [27]. The internal porosity of C-S-H gel, the so-called ‘gel pores’, ranges from 5 Å to 100 Å. Hence, in theory, the SFS is expected to have more fine pores. Consequently, the fine pores lead to a larger capillary pressure so that the autogenous shrinkage induced by self-desiccation will be accelerated. The results regarding the amount of fine pores obtained by nitrogen gas adsorption method are in good agreement with this theory. As shown in Fig. 4, the autogenous shrinkage was reduced when 50% of SF was replaced by FA and GGBS. However, when the replacement ratio reached 100%, the autogenous shrinkage became remarkably small. The obtained values yielded the lowest and approximately equal shrinkage for FAS and GBS. Due to intimate relation between the amount of fine pores and autogenous shrinkage, it can be expected that reducing the amount of fine pores in FAS and GBS specimens leads to a reduction in the autogenous shrinkage. The higher amount of large pores in FAS and GBS results in a smaller capillary pressure in the paste when water is consumed by the hydration of cement [33,34]. In addition, the autogenous shrinkage was drastically reduced at early ages for FAS specimens. This clearly indicates that the onset of the shrinkage is highly affected by the material type. The addition of FA can delay the initial hydration phase so that less C-S-H will be formed and, as a consequence, the amount of fine pores will be reduced. As it can be seen, the addition of GGBS was also found to be effective in reducing the autogenous shrinkage. This might be due to the critical range of the fine pores which contributes to the autogenous shrinkage. It has been reported that 50–500 Å is a critical pore size which affects the autogenous shrinkage [35]. As presented in Fig. 3, GBS exhibits more fine pores in the range of 50–500 Å (compared to SFS) that resulted in a lower autogenous shrinkage. 3.3.1. Prediction model for UHPC Most of the developed models for predicting the autogenous shrinkage of current concrete are not applicable for UHPC due to the fact that these take into account the effect of coarse aggregates. In addition, the effect of supplementary cementitious materials has not been well addressed in these models. In this scope, a prediction model was developed to establish the correlation between the pore structure of UHPC and the autogenous shrinkage, based on the method proposed by Li et al. [35] for the prediction of autogenous shrinkage of cement paste. In this method, a polynomial equation was adopted to predict the autogenous shrinkage, which can be expressed as (Eq. (1)):
ea ðtÞ ¼ at3 þ bt2 þ ct þ d
Fig. 4. Autogenous shrinkage of specimens.
ð1Þ
where ea (t) is the autogenous shrinkage of cement paste at the age t (in days), and a, b, c and d are the constants of the prediction model equation. The autogenous shrinkage regression equations along with the correlation coefficients of each equation are presented in Table 4. As presented, the high value of correlation coefficient corroborated the adequacy of the model used to predict the autogenous shrinkage of UHPC specimens containing different SCMs. Once the prediction equation was obtained, all the attempts were focused on establishing the relationship between the obtained fine pores and
47
E. Ghafari et al. / Construction and Building Materials 127 (2016) 43–48 Table 4 The equations for predicting the autogenous shrinkage of UHPC samples. Mixture
a (lm/(m day3))
b (lm/(m day2))
c (lm/(m day))
d (lm/m)
Correlation coefficient
GBS FAS GBSF FASF SFS
0.0335 0.0366 0.0311 0.0302 0.0343
2.0704 2.1433 1.8015 1.5464 2.117
47.365 47.796 41.183 35.663 47.967
35.189 17.311 24.809 6.0488 30.981
0.99 0.99 0.99 0.98 0.97
Fig. 5. The relationship between the porosity and the autogenous shrinkage at 28 days.
the autogenous shrinkage of UHPC. In this scope, a numerical regression method based on the previous literature was adopted which is expressed as follows [35]:
P ¼ A þ lnðea Þ
ð2Þ
where, ea is the autogenous shrinkage at the age of 28 days, p is the porosity (%), and A is constant to improve the prediction accuracy. The total porosity of the UHPC mixtures along with the autogenous shrinkage dates were used as an input to Eq. (2). Fig. 5 shows the relationship between the autogenous shrinkage value and the total porosity of the UHPC mixtures. A correlation coefficient R2 = 0.97 was found, revealing an excellent linear correlation between the natural logarithm of autogenous shrinkage and the total porosity of the UHPC samples. The results indicate that the amount of fine pores in UHPC is a predominant factor that can highly affect the autogenous shrinkage. Recalling the discussion regarding the fine pore distribution and pore volume in Fig. 3, it can be concluded that the combination of these two parameters accelerated the self-desiccation process, resulting in a high autogenous shrinkage. 3.4. Compressive strength Results for compressive strength in both steam-cured and water-cured conditions are shown in Fig. 6. Regardless of the mixture proportion, all the specimens that were submitted to the former curing condition exhibited higher compressive strength, compared to those maintained in water. In the latter case, no significant difference was observed in the compressive strength of all mixtures. However, the compressive strength of SF specimen exposed to steam was approximately 20% higher than that of the other specimens that were cured with steam. This implies that
Fig. 6. Compressive strength of UHPC samples containing different SCMs in different curing conditions at 28 days.
steam curing could accelerate the reaction between the mineral admixtures and cement particles. There is a synergistic effect between the water adsorption of the mineral admixtures and the pozzolanic reactivity. SF in the UHPC mixture showed the best results in terms of compressive strength which would be due to the availability of higher amount of portlandite that accelerates the pozzolanic reaction of the particle [36,37]. FAS reached a compressive strength of 150 MPa when cured with steam, which was the lowest value of compressive strength among the specimens. Constantinides et al. [38] have performed extensive research on the low density (LD) C-S-H and high density (HD) C-S-H. Their research has shown that LD C-S-H covers the surface of the cement and FA, and that HD C-S-H penetrates into the hydration particles. The low amount of compressive strength obtained by FAS specimen denotes that FA could not react well within the UHPC mixture, although a significant pozzolanic reaction was expected. However, further research is required for proving the fact that FA would lead to significant lower mechanical properties, since the cement hydration is considerably faster than the FA pozzolanic reaction [39,40]. Compared with the specimens cured in water, SFS showed 21% increase in the compressive strength when cured with steam, recorded as the highest increase among the compressive strength values. This could be due to the high pozzolanic reactivity of SF that has led to a higher formation of C-S-H gel and crystalline CH [41]. The compressive strength of GBS cured in water was 135 MPa, the lowest compressive strength among the specimens cured in water. This can be attributed to that fact that UHPC contains an amount of water that is consumed by the powders in the mixture. If a mineral admixture adsorbs a considerable amount of water, some amount of cement powder would remain intact and less C-S-H gel would be produced accordingly. GGBS is known by its ability in reducing the pore size and the cumulative pore volume of the paste [42]. Therefore, more water might have been
48
E. Ghafari et al. / Construction and Building Materials 127 (2016) 43–48
adsorbed by GGBS compared with other admixtures leading to lower compressive strength values. It is worth mentioning that the incorporation of both FA and GGBS increased the fluidity of the mixtures with the same W/B ratio. As mentioned before, the adopted strategy aimed at measuring the autogenous shrinkage of all mixtures in the same conditions. Therefore, it was expected that the compressive strength of the specimens with GGBS and FA would have increased with the reduction in the water content by designing the mixtures to have the same consistency instead of the same water content. Finally, the FASF and GBSF specimens showed insignificant difference in the mechanical properties when cured in water and cured with steam. 4. Conclusions In this study, the ability of SCMs, such as FA and ground granulated blast furnace slag, to be used as replacement of SF in UHPC was evaluated. The addition of these was found to be effective in reducing the autogenous shrinkage of UHPC. Results also indicate that the amount of fine pores in UHPC is a predominant factor that can highly affect the autogenous shrinkage. A strong linear correlation between the natural logarithm of the latter and the total porosity of the UHPC mixtures was established. It was found that reducing the amount of fine pores in FAS and GBS leads to a reduction in the autogenous shrinkage. However, it was also concluded that the incorporation of FA and GGBS gives similar performance in compressive strength when compared to the reference mixture (containing only SF). In particular, it was observed that although incorporating FA into the mixtures results in a slight decrease of the mechanical properties, much lower autogenous shrinkage can be achieved. Acknowledgements This work has been supported by the Portuguese Foundation for Science and Technology (FCT) under Grants SFRH/BD/91402/2012 and the Project PTDC/ECM/098497/2008 entitled ‘‘Intelligent Super Skin (ISS)”. The authors wish to thanks SIKA company for their support and partnerships in the execution of this project. References [1] E. Ghafari, M. Arezoumandi, H. Costa, E. Júlio, Influence of nano-silica addition on durability of UHPC, Constr. Build. Mater. 94 (2015) 181–188. [2] P. Richard, M. Cheyrezy, Composition of reactive powder concrete, Cem. Concr. Res. 25 (7) (1995) 1501–1511. [3] M. Heikal, S. Abd El-Aleem, W.M. Morsi, Characteristics of blended cements containing nano-silica, HBRC J. 9 (3) (2013) 243–255. [4] S. Abd El-Aleem, M. Heikal, W.M. Morsi, Hydration characteristic, thermal expansion and microstructure of cement containing nano-silica, Constr. Build. Mater. 59 (2014) 151–160. [5] A. Abdelaziz, Saleh Abd El-Aleem, Wagih M. Menshawy, Effect of fine materials in local quarry dusts of limestone and basalt on the properties of Portland cement pastes and mortars, Int. J. Eng. Res. Technol. (IJERT) 3 (6) (2014) 1038– 1056. [6] S. Ghahari, A.M. Ramezanianpour, A.A. Ramezanianpour, M. Esmaeili, An accelerated test method of simultaneous carbonation and chloride ion ingress: durability of silica fume concrete in severe environments, Adv. Mater. Sci. Eng. 2016 (2016). [7] First International Symposium on Ultra High Performance Concrete September 2004 Kassel, Germany. [8] Second International Symposium on Ultra High Performance Concrete March 2008 Kassel, Germany. [9] T.C. Holland, Silica Fume User’s Manual. 2005, Federal Highway Administration (FHWA). [10] P. Rougeau, B. Borys, Ultra High Performance Concrete with ultrafine particles other than silica fume in International Symposium on Ultra High Performance Concrete 2004 Kassel, Germany. [11] M.I. Khan, Properties of High Performance Concrete, King Saud University, 2006. [12] P. Lura, Ole Mejlhede Jensen, Jason Weiss, Cracking in cement paste induced by autogenous shrinkage, Mater. Struct. 42 (8) (2009) 1089–1099.
[13] M.S. Meddah, Tagnit-Hamou Arezki, Pore structure of concrete with mineral admixtures and its effect on self-desiccation shrinkage, ACI Mater. J. (Jun 2009). [14] A.W. Shin-ichi Igarashi, Mitsunori Kawamura, Evaluation of capillary pore size characteristics in high-strength concrete at early ages, Cem. Concr. Res. 35 (2005) 513–519. [15] M. Mazloom, A.A. Ramezanianpour, J.J. Brooks, Effect of silica fume on mechanical properties of high-strength concrete, Cement Concr. Compos. 26 (4) (2004) 347–357. [16] M.H. Zhang, C.T. Tam, M.P. Leow, Effect of water-to-cementitious materials ratio and silica fume on the autogenous shrinkage of concrete, Cem. Concr. Res. 33 (10) (2003) 1687–1694. [17] F. Robert, M.W.M. Wescott, Martha G. VanGeem, John Gajda, Prospects for Expanding the Use of Supplementary Cementitious Materials in California, February 2010. [18] M.C.G. Juenger, Siddique Rafat, Recent advances in understanding the role of supplementary cementitious materials in concrete, Cem. Concr. Res. 78 (2015) 71–80. [19] Z. Li, Drying shrinkage prediction of paste containing meta-kaolin and ultrafine fly ash for developing ultra-high performance concrete, Mater. Today Commun. 6 (2016) 74–80. [20] A. Tafraoui, G. Escadeillas, S. Lebaili, T. Vidal, Metakaolin in the formulation of UHPC, Constr. Build. Mater. 23 (2009) 669–674. [21] H. Yazıcı, H. Yig˘iter, A.S ß. Karabulut, B. Baradan, Utilization of fly ash and ground granulated blast furnace slag as an alternative silica source in reactive powder concrete, Fuel 87 (12) (2008) 2401–2407. [22] G. Herold, H.S. Müller, Measurement of porosity of Ultra High Strength Fibre Reinforced Concrete in International Symposium on Ultra High Performance Concrete September 13–15, 2004, Kassel, Germany. [23] S.N. Lim, T.H. Wee, Autogenous shrinkage of ground-granulated blast-furnace slag concrete, Mater. J. 97 (5) (2000) 587–593. [24] EN196-1, Methods of testing cement-Part 1: Determination of strength, 2005. [25] E.P. Barrett, L.G. Joyner, P.P. Halenda, The determination of pore volume and area distributions in porous substances. I. Computations from nitrogen isotherms, J. Am. Chem. Soc. 73 (1) (1951) 373–380. [26] G. Fagerlund, Determination of specific surface by the BET method, Matériaux Constr. 6 (3) (1973) 239–245. [27] H. Madani, A. Bagheri, T. Parhizkar, The pozzolanic reactivity of monodispersed nanosilica hydrosols and their influence on the hydration characteristics of Portland cement, Cem. Concr. Res. 42 (12) (2012) 1563–1570. [28] E. Ghafari, A. Baig, K. Nicoletta, D. Feys, R.D. Ferron, K.H. Khayat, Admixture Compatibility of Alternative Supplementary Cementitious Materials for Pavement and Structural Concrete. 2014, Center for Transportation Infrastructure and Safety/NUTC program: Published. In: The National Academies of Sciences Engineering and Medicine. [29] C.G. Schull, The determination of pore size distribution from gas adsorption data, J. Am. Chem. Soc. 70 (4) (1948) 1405–1410. [30] D.N. Winslow, S. Diamond, A mercury porosimetry study of the porosity in Portland cement, J. Mater. (1970). [31] D. Saje, Reduction of the early autogenous shrinkage of high strength concrete, Adv. Mater. Sci. Eng. 2015 (2015). [32] J.D. Jianxin Ma, Ultra High Performance Self Compacting Concrete LACER, 2002, p. 7. [33] S. Tangtermsirikul, Effect of chemical composition and particle size of fly ash on autogenous shrinkage of paste, in: Proceedings of an International Workshop on the Autogenous Shrinkage of Concrete, E & FN Spon, London, 1999. [34] P. Termkhajornkit, T. Nawa, M. Nakai, T. Saito, Effect of fly ash on autogenous shrinkage, Cem. Concr. Res. 35 (3) (2005) 473–482. [35] Y. Li, J. Bao, Y. Guo, The relationship between autogenous shrinkage and pore structure of cement paste with mineral admixtures, Constr. Build. Mater. 24 (10) (2010) 1855–1860. [36] E. Ghafari, H. Costa, E. Júlio, A. Portugal, L. Durães, The effect of nanosilica addition on flowability, strength and transport properties of ultra high performance concrete, Mater. Des. 59 (2014) 1–9. [37] E. Ghafari, H. Costa, E. Júlio, Critical review on eco-efficient ultra high performance concrete enhanced with nano-materials, Constr. Build. Mater. 101 (2015) 201–208. [38] G. Constantinides, F.-J. Ulm, The nanogranular nature of C–S–H, J. Mech. Phys. Solids 55 (1) (2007) 64–90. [39] V.G. Papadakis, Effect of fly ash on Portland cement systems: Part I. Low-calcium fly ash, Cem. Concr. Res. 29 (11) (1999) 1727–1736. [40] S. Hanehara et al., Effects of water/powder ratio, mixing ratio of fly ash, and curing temperature on pozzolanic reaction of fly ash in cement paste, Cem. Concr. Res. 31 (1) (2001) 31–39. [41] A.M. Ramezanianpour, Kh. Esmaeili, S.A. Ghahari, A.A. Ramezanianpour, Influence of initial steam curing and different types of mineral additives on mechanical and durability properties of self-compacting concrete, Constr. Build. Mater. 73 (2014) 187–194. [42] P.A.M. Basheer, P.R.V. Gilleece, A.E. Long, W.J. Mc Carter, Monitoring electrical resistance of concretes containing alternative cementitious materials to assess their resistance to chloride penetration, Cement Concr. Compos. 24 (5) (2002) 437–449.