Service life of metakaolin-based concrete exposed to carbonation

Cement and Concrete Research 99 (2017) 18–29

Contents lists available at ScienceDirect

Cement and Concrete Research journal homepage: www.elsevier.com/locate/cemconres

Service life of metakaolin-based concrete exposed to carbonation Comparison with blended cement containing fly ash, blast furnace slag and limestone filler

MARK

Raphaël Bucher, Paco Diederich, Gilles Escadeillas, Martin Cyr⁎ Laboratoire Matériaux et Durabilité des Constructions (LMDC), Université de Toulouse, INSA/UPS Génie Civil, 135 Avenue de Rangueil, 31077 Toulouse Cedex 04, France

A R T I C L E I N F O

A B S T R A C T

Keywords: Metakaolin Carbonation CO2 ingress Model Supplementary cementing materials

Carbonation of cement-based materials can lead to corrosion of the steel bars in reinforced concrete, and supplementary cementing materials (SCMs) often increase the carbonation kinetics compared to reference concretes made of cements composed of clinker only. The aim of this work was to assess the consequences, in terms of service life of structures, of a possible increase in the carbonation depth of metakaolin concretes. Experimental testing using accelerated (4% CO2) and natural carbonation conditions showed that increasing the metakaolin content tended to increase the carbonation depth of concretes, due to the consumption of portlandite by the pozzolanic reaction. However, most of the time, the carbonation was within the range of carbonation depths found in commercially available blended cements (including fly ash, GGBS or limestone filler) that had already proved their worth on the market. The combination of MK and limestone filler (in CEM II/A-LL with 15% MK) behaved very well with respect to carbonation, the carbonation depth being almost equivalent to that of CEM I samples. The modelling of CO2 ingress into the concretes showed that, although metakaolin increased the carbonation (except when associated with limestone filler), the carbonation depth did not exceed 30 mm after 50 years, far from the value of 50 mm generally used as concrete cover to protect the steel bars. This means that the formulations including metakaolin would not have deteriorated by the end of the building's service life.

1. Introduction Carbonation of cement-based materials is an important pathology that can cause severe degradation of concrete, since it reduces the pH of the pore solution and thus leads to the corrosion of steel rebars through the destruction of their passivation layer [1]. This decrease of the pH is due to the consumption of portlandite (produced by the hydration of the cement) by atmospheric CO2. The pozzolanic reaction of metakaolin (MK)—and more generally of all other supplementary cementing materials (SCM)—leads to a decrease of the portlandite reserve in the cement paste, so several authors have concluded [2–6] that the replacement of Portland cement by metakaolin usually causes an increase of the carbonation kinetics of concrete. Carbonation results obtained by San Nicolas [2] after concrete exposure of 28 days in a carbonation chamber (50% CO2, 20 °C and 65% relative humidity) showed that the carbonation depth increased from 2 to 9 mm (+ 450%) when 25% of cement was replaced by metakaolin. These results were in correlation with the work of Kim et al. [3], who observed increases of 40, 70, 100 and 370% for the replacement of 5, 10, 15 and 25% of cement, respectively, by ⁎

metakaolin (56 days at 5% CO2, 30 °C and 60% relative humidity). McPollin et al. [4] confirmed these results with 10% of metakaolin (5% CO2, 20 °C and 65% relative humidity). Mejia de Gutiérrez et al. [5] studied the effect of the duration of the wet cure of metakaolin-based mixtures on the carbonation kinetics (exposure for 3 and 6 weeks in a carbonation chamber: 2.25% CO2, 30 °C and 70% of relative humidity), as it is well known that the pozzolanic reaction is slower than the cement hydration reaction. Their results were in agreement with those in the literature [2–4], with a greater carbonation depth in metakaolin concrete than in a reference sample when the curing time was limited to 28 days. However, for longer curing times (wet cure of 240 days), the carbonation kinetics was slower in the case of concrete made with metakaolin than in the control concrete with cement only. The use of other SCMs also causes an increase in the carbonation kinetics when compared to a reference sample composed of cement only. In the case of ground granulated blast furnace slag (GGBS), high doses could lead to a significant increase in the carbonation depth [6,7]. This effect can be reduced by longer wet curing [8,9] or by the use of finer GGBS [9]. Increases in the carbonation kinetics are also seen when high doses of silica fume [10] or fly ash (> 30%) are used

Corresponding author. E-mail address: [email protected] (M. Cyr).

http://dx.doi.org/10.1016/j.cemconres.2017.04.013 Received 26 November 2016; Received in revised form 24 April 2017; Accepted 26 April 2017 0008-8846/ © 2017 Elsevier Ltd. All rights reserved.

Cement and Concrete Research 99 (2017) 18–29

R. Bucher et al.

2. Materials and methods

[11]. However, Bai et al. [12] showed that the combination of metakaolin and fly ash led to a reduction of the carbonation kinetics with respect to a concrete with fly ash only (4% of CO2). The aim of the present work was to assess the consequences of the possible increase in the carbonation depth of metakaolin concretes, in terms of service life of structures. For this purpose, three points were treated and are reported in this paper:

2.1. Materials Table 1 gives the characteristics of the binders used in this study. All the cements were commercial ones and complied with the European standard EN 197-1 [15]. Two plain cements containing a minimum of 95% of clinker were used as references: normal (CEM I) and low-sulfate (CEM I PM-ES) cement, the latter containing less C3A. Five blended cements containing one or a combination of additions were also used as references:

1- Most of the carbonation data found in the literature concerns relatively pure metakaolin and a rotary kiln. The work presented here used flash-calcined metakaolin having a particular composition [13]. The reported data could thus improve our knowledge on the effect of different types of metakaolin on the concrete carbonation process. 2- The most typical studies found in regard to carbonation of SCM matrices compare a pure Portland cement and an SCM concrete containing less cement, so less portlandite. The carbonation results are rarely in favor of the SCM, whatever its nature. Nevertheless, the use of blended cements remains the usual practice when the aim is to improve other durability aspects of concrete. These blended cements are accepted for carbonation exposure classes (governed by the XC class in European standard EN 206-1 [14]), even if they are used in concretes subjected to carbonation, as long as they are included in a standard such as EN 197-1 [15]. For this reason, several commercial blended cements (based on limestone filler, blast furnace slag and fly ash) were used as references and to provide a basis of comparison for accelerated and natural carbonation studies of metakaolin concretes. 3- Even in the case of an increase of the carbonation of concrete with SCMs, it remains pertinent to evaluate the possible consequences of such behavior in terms of service life of the concrete. A model [16] predicting the carbonation depth of reinforced concrete structures, adjusted with the experimental results of natural carbonation, was thus used to compare the long-term carbonation behavior of pure Portland cement, blended cements and metakaolin concretes.

– – – –

Limestone cement: a CEM II/A-LL (16% limestone filler); Fly ash cement: a CEM II/A-V with 15% of fly ash; Blast furnace slag cement: a CEM III/A with 62% of GGBS; Combination of GGBS and fly ash cement: a CEM V/A with 22% of each addition.

The metakaolin used was a flash calcined one [17]. Flash calcination refers to the combustion process where the particles of kaolinite are transformed into metakaolin by passing near a flame (temperature around 700 °C) for a few tenths of a second [18]. This process is faster and consumes less energy than traditional methods (e.g. rotary kiln) [13]. Due to the lower purity of the deposit, this metakaolin has an impurity rate of about 50%, mainly composed of quartz. This leads to a marked decrease in the surface area and therefore in water demand compared to a pure metakaolin. The siliceous aggregates used were divided into six size classes (0–0.315 mm, 0.315–1 mm, 1–4 mm, 4–8 mm, 8–12 mm, 12–20 mm). The particle size distribution of the aggregates in the concretes was established by means of the Dreux method in order to optimize the compactness of the grading curve [18]. The High Range Water Reducing Agent (HRWRA) was a polycarboxylate type available as a commercial solution (density = 1.05 kg/m3; active solid content by weight = 30.5%). 2.2. Concrete design Table 2 summarizes the 11 concrete designs evaluated in this study. Several concretes without metakaolin were cast to be used as references of commercial mixtures accepted in the ready-mixed concrete industry:

Table 1 Composition and physical properties of the commercial cements and metakaolin. Cement type

CEM I

CEM I PM-ES

CEM III/A

CEM II/A-LL

CEM II/A-V

CEM V/A

Metakaolin

Additions Strength class

– 52.5

– 52.5

Blast furnace slag 52.5

Limestone filler 42.5

Fly ash 42.5

Blast furnace slag and Fly ash 42.5

– –

Chemical properties (%) SiO2 Al2O3 CaO MgO Fe2O3 K2O Na2O TiO2 SO3 Cl− Loss on ignition

19.75 5.27 63.97 1.93 2.39 0.48 0.17 – 2.95 0.03 1.77

21.70 3.70 65.70 1.00 4.60 0.29 0.16 0.23 2.59 0.02 0.90

26.20 7.44 55.00 4.00 1.70 0.36 0.57 – 1.82 0.33 2.23

18.00 3.80 63.00 1.20 2.20 0.70 0.09 0.20 2.60 0.03 7.70

25.55 8.08 55.54 1.02 3.44 1.09 0.20 0.40 2.55 0.01 1.40

28.70 10.00 46.80 2.60 3.40 1.27 0.24 0.60 2.80 0.01 2.10

67.10 26.80 1.12 0.11 2.56 0.12 0.01 1.37 – – 0.84

Physical properties Specific density (kg/m3) Blaine/BET specific surface area (cm2/g) Compressive strength MPa (28 days)

3140 3801 65.3

3190 3550 60.8

3010 4300 64.7

3070 4085 54.1

3000 3613 49.4

2910 4900 48.1

2510 165 000 (BET) –

Bogue composition (clinker) (%) C3S C2S C3A C4AF

58.1 10.0 10.7 8.0

64.0 11.0 2.5 14.1

58.1 10.0 10.7 10.0

65.0 13.0 8.0 8.0

63.0 12.9 7.9 9.7

66.0 11.0 7.0 11.0

– – – –

19

Cement and Concrete Research 99 (2017) 18–29

R. Bucher et al.

Table 2 Mixture proportions and fresh properties of the concretes. Materials (kg/m3)

CEM CEM CEM CEM CEM CEM CEM CEM CEM CEM CEM

I I PMES II/A-LL II/A-V III/A V/A I 15 I 20 I 25 II/A-LL 15 II/A-V 15

Cement

280 280 280 280 280 280 238 224 210 238 238

MK

0 0 0 0 0 0 42 56 70 42 42

Aggregates 0–0.3

0.3–1

1–4

4–8

8–12

12–20

246 247 252 252 245 253 245 245 244 251 251

104 104 106 106 103 107 103 103 103 106 106

388 380 397 396 385 399 385 385 385 396 395

71 71 73 73 72 74 71 71 71 73 72

310 311 318 317 309 319 308 308 308 317 316

768 769 787 785 764 789 762 763 762 784 782

– Two plain cement concretes based on normal (CEM I) and lowsulfate cement (CEM I PM-ES); – Five blended cement concretes containing the additions alone or combined.

– Replacement of 15, 20 and 25% of plain CEM I; – Replacement of 15% of limestone cement (CEM II/A-LL) and fly ash cement (CEM II/A-V). The choice of using MK to replace a fraction of blended cement was made to evaluate the interaction of MK with other SCMs. Moreover, since 2015, the French annex to the European standard EN 206-1 [14] (concerning to the specification, performance, production and conformity of concrete) allows the use of SCMs such as MK to replace a fraction of blended cement (CEM II/A, which initially contained a minimum of 80% of clinker). In the case of MK, the standard accepts 10% of replacement, so it was decided to increase the MK content to 15% to evaluate the consequences of such a higher value. The concretes were designed with the objective of obtaining an equivalent compressive strength of 50 MPa for all mixtures on a cube at 28 days. In order to reach the objective 50 MPa at 28 days, the water/ binder (W/B) ratios were set at different values for each concrete mixture to take account of the different strength classes of the cements used. A slump of 17 ± 2 cm was targeted for the mixtures by adjusting the HRWRA quantity (expressed in mass percentage of powder). Each batch comprised around 70 l of concrete. The concrete was cast, with vibration, in several molds for mechanical and durability tests. 2.3. Methods The compressive strengths were tested on 10 cm cubes after a wet cure of 7, 28, 90 or 365 days. The tests were carried out at a loading speed of 0.5 MPa/s on three replicate samples. The concrete porosity was evaluated by means of a water accessible porosity test according to the French standard (NF P18-459) [19]. The aim of this test is to measure the connected open porosity. The sample was saturated with water under vacuum in order to fill all the open porosity. The accessible water porosity (Eq. (1)) was calculated from the mass of the dry sample, the mass of the wet sample and the wet sample mass when submerged in water.

Ms − Md × 100 Ms − Mw

W/B

HRWRA (%)

Air content (%)

Slump (cm)

167 167 147 147 167 147 167 167 167 147 147

0.60 0.60 0.53 0.53 0.60 0.53 0.60 0.60 0.60 0.53 0.53

0.6 0.5 1.5 1.5 0.5 1.5 1.0 1.3 1.3 2.5 2.5

2.3 1.8 1.3 1.6 1.7 1.6 1.9 2.0 1.9 1.7 1.6

17.5 18.5 16.0 18.0 17.0 18.0 17.0 18.0 17.5 15.0 14.5

saturated sample. These tests were carried out after 28 and 365 days of curing on three replicate samples of height 5 cm (three discs cut from a cylinder (Ø = 11 cm, h = 22 cm)). Gas permeability was tested using the French standard XP P18-463 [20], at relative pressures of 1.0 bar to assess apparent permeability, and between 1.5 and 4.0 bars to assess permeability by the Klinkenberg approach. The aim of this test was to measure the quantity of pressurized air crossing a dry sample. The greater the air volume passing through the sample, the more permeable the sample. This higher permeability could reduce the durability of the concrete by making it easier for aggressive agents to penetrate it. This permeability gives an empirical evaluation of the tortuosity of the porous network. The tests were performed after 28 and 365 days of curing on three replicate dried samples (three discs coming from a cylinder (Ø = 11 cm, h = 22 cm)). All the accelerated carbonation tests were performed on three replicate prismatic samples (H = 28 cm, L = 7 cm, l = 7 cm), after curing times of 28 and 365 days. Between the curing time and the beginning of the carbonation test, the samples were stored in an atmosphere with 50 ± 5% relative humidity at 20 °C for 14 days. The samples for accelerated carbonation were kept in a chamber at 20 °C, 55% relative humidity and a 4 ± 0.5% CO2 content. The measurements of carbonation depth were carried out after 56, 63 and 70 days on split samples, on twelve points per sample observable after spraying phenolphthalein solution. Natural carbonation was allowed to occur on the same type of samples after 28 days of curing. In this case the specimens were stored in-situ outside the laboratory (mean temperature in Toulouse from 8 to 18 °C between winter and summer, with an average relative humidity of 75%). The carbonation depth was measured after 1 and 2 years of natural exposure. The quantity of portlandite in concrete after 28 days was assessed by thermogravimetric analysis using a Netzch STA 44F3 device. The heating rate of the test was 10 °C/min and the CH quantity was determined by the tangent method.

Metakaolin was used to replace a fraction of the cement for three types of mixtures:

p=

Water

3. Results 3.1. Concrete properties 3.1.1. Compressive strength Table 3 summarizes the compressive strength results. The intention was to work on concrete for which the compressive strength at 28 days was the same for all samples (50 ± 5 MPa). Only one sample deviated from this value, i.e. the concrete made of CEM II/A-LL 15 (64.7 ± 1.1 MPa). In the short term (7 days), a decrease was noticed in the compressive strength of concretes with some SCMs compared to CEM I. Concretes with GGBS (CEM III/A), metakaolin (CEM I 25), fly ash

(1)

where p is the open porosity, Mw is the mass of the wet sample dipped in water, Md is the mass of the dry sample and Ms is the mass of the 20

Cement and Concrete Research 99 (2017) 18–29

R. Bucher et al.

Table 3 Compressive strength of concrete mixtures.

CEM I CEM I PMES CEM II/A-LL CEM II/A-V CEM III/A CEM V/A CEM I 15 CEM I 20 CEM I 25 CEM II/A-LL 15 CEM II/A-V 15

Cement class

W/B

Compressive strength at 7 days (MPa)

Compressive strength at 28 days (MPa)

Compressive strength at 90 days (MPa)

Compressive strength at 365 days (MPa)

52.5 52.5 42.5 42.5 52.5 42.5 52.5 52.5 52.5 42.5 42.5

0.60 0.60 0.53 0.53 0.60 0.53 0.60 0.60 0.60 0.53 0.53

43.1 31.5 43.8 36.5 33.0 31.4 38.6 39.7 35.4 55.1 36.3

49.0 – 53.7 51.4 48.2 48.3 47.4 52.9 49.1 64.7 47.3

54.1 52.5 57.1 – 59.7 66.1 53.4 53.0 51.1 68.5 49.1

58.7 54.2 60.4 67.9 62.2 73.2 53.5 54.2 48.3 72.9 65.5

± ± ± ± ± ± ± ± ± ± ±

0.5 0.6 1.7 1.9 1.7 0.6 2.3 2.3 1.3 1.1 0.3

± 1.9 ± ± ± ± ± ± ± ± ±

0.8 1.5 2.3 0.3 1.0 1.2 0.3 1.1 4.0

(CEM II/A-V), and fly ash and metakaolin (CEM II/A-V 15) gave compressive strengths about 10 MPa lower than the CEM I. This decrease in compressive strength was smaller with 15% and 20% of MK (CEM I 15 and CEM I 20, respectively). It was noted that the combination of MK and limestone filler (CEM II/A-LL 15) led to a 7-d compressive strength > 10 MPa higher than the reference concrete made of CEM I. After 28 days, if strengths are considered equivalent when their values are within ± 10%, one concrete differed from specifications: CEM II/A-LL 15, which had a higher compressive strength (64.7 ± 1.1 MPa). The combination of MK and limestone filler seemed to lead to very good results, as will be discussed later. After a wet cure of 365 days, the compressive strength could almost be classified according to the W/B ratio. The compressive strengths were between 53 and 62 MPa (except for the CEM I 25) for concretes with a W/B ratio of 0.6, and between 60 and 73 MPa for the concretes with a W/B ratio of 0.53. The development of the compressive strength of concrete made with metakaolin varied according to the cement used. When each concrete containing metakaolin was compared to the reference concrete made with the same cement, it was observed that compressive strengths:

± 1.6 ± 1.5 ± 0.9 ± ± ± ± ± ± ±

0.8 2.8 0.5 3.1 1.9 1.6 3.0

± ± ± ± ± ± ± ± ± ± ±

1.3 1.0 1.1 2.2 2.6 0.9 2.0 2.2 2.6 1.4 0.7

of 1 to 1.5% after a wet cure of 28 days, but this tendency was less visible after a wet cure of 365 days. For the gas permeability test, the cement substitution by metakaolin was not significantly affected by a wet cure of 28 days or 365 days. For the formulations made with blended cements, the use of 15% of metakaolin (with CEM II/A-LL) did not have a significant effect on the water porosity at 28 or 365 days; the results remained in the same order of magnitude when the standard deviation was considered. However, the use of metakaolin led to a slight decrease in the gas permeability. The open porosity increased by 1% after wet cures of 28 and 365 days when 15% of metakaolin was used in replacement of CEM II/ A-V cement. 3.1.3. Portlandite content The portlandite contents are given in Table 4 for all concrete mixtures. As expected, CEM I cements contained high proportions of portlandite due to the large proportion of clinker they contained (> 95%), and to the absence of pozzolanic materials that could consume portlandite over time. The design based on cements with SCMs contained smaller quantities of portlandite than pure Portland cements (CEM I and CEM I PMES) did. On the one hand, this decrease could be related to the dilution effect of the clinker when a mineral powder was used to replace a fraction of the pure Portland cement. It concerned all the concretes made with other cements than CEM I here. For instance, the replacement of 25% of CEM I by MK would lead to a portlandite content of around 10% (75% of 13.6% for plain cement, but without considering that the MK could improve the cement hydration due to germination effect). On the other hand, portlandite was partially consumed by some chemically active pozzolanic SCMs such as fly ash (CEM II/A-V), GGBS (CEM III/A), a combination of fly ash and GGBS (CEM V/A) or metakaolin. In the case of MK, its use was clearly seen to lead to a decrease in the portlandite content: use of more MK implied less portlandite in the concrete (CEM I 15, CEM I 20, CEM I 25, for 15, 20 and 25% of MK, respectively).

– Decreased slightly for CEM I; – Were equal for CEM II/A-V (fly ash); – Increased significantly for CEM II/A-LL (limestone filler). 3.1.2. Water porosity and gas permeability The results of water porosity and gas permeability are given in Table 4. 3.1.2.1. W/B effect. The performance of the concretes depended strongly on the water content in the mixtures, as the concretes with a W/B = 0.6 had a water porosity between 12% and 14.5%, while the water porosity was < 12% for all mixtures with a W/B of 0.53. The tendency was the same for the permeability results: the coefficient of permeability was higher than 100 × 10− 12 m2/s for W/B of 0.6, while the mixtures with W/B = 0.53 led to a permeability lower than 100 × 10− 12 m2/s (except for the CEM II/A-V at 365 days).

3.2. Effect of metakaolin on the carbonation depth 3.2.1. Comparison with a reference: accelerated carbonation Fig. 1 presents the accelerated carbonation results of concretes based on cements CEM I (panel a) and CEM II/A (panel b), after wet cures of 28 days and 365 days.

3.1.2.2. Cure effect. The effect of a longer wet cure was visible on the results of both the water porosity and the gas permeability. On the one hand, the development of hydration reactions between 28 and 365 days caused a decrease or stagnation in the water porosity volume and a matrix densification (except for CEM II/A-LL, CEM III/A and CEM II/ALL 15). On the other hand, the longer wet cure led to a decrease in the gas permeability (except for stabilization in CEM II/A-V 15 and a slight increase in CEM II/A-V and CEM III/A).

3.2.1.1. After 28 days of wet cure. In the case of cement CEM I, using more metakaolin in the system caused a progressive increase in the carbonation depth after 70 days of accelerated carbonation (4% of CO2): + 73%, + 85% and + 120% for MK contents of 15, 20 and 25%, respectively. The replacement of 15% CEM II/A-V by MK led to a similar trend but to a lesser extent, as the increase in carbonation depth reached only 27%.

3.1.2.3. Metakaolin effect. In the case of CEM I, the replacement of cement by 15 to 25% of metakaolin caused a decrease in water porosity 21

Cement and Concrete Research 99 (2017) 18–29

3.6 ± 0.9

It is noteworthy that the combination of MK and limestone filler (CEM II/A-LL with 15% MK) showed very good behavior against carbonation, as it led to a decrease of 15% in carbonation depth when compared to the reference without MK. 3.2.1.2. After 365 days of wet cure. A long period of curing was beneficial to the MK concretes most of the time, although the carbonation depths remained greater than those of CEM I and CEM II/A-V concretes. This benefit was particularly noticeable for CEM I 20 (20% of MK), which seemed to be an optimal replacement rate to limit long-term carbonation. This replacement rate of cement by MK had already shown good durability behavior in other studies [2,13]. Very good behavior was still observed for the concrete CEM II/A-LL 15 (combination of 15% of MK and 16% of limestone filler), with a carbonation depth almost equivalent to that of the CEM I samples. This means that the use of a combination of clinker, limestone filler and MK could behave as well as CEM I alone with respect to carbonation. This could provide a real opportunity to decrease the clinker content in concretes (on other concretes than very high strength ones) without sacrificing the carbonation durability. The results of permeability and porosity were not used because it was not possible to establish a relation between carbonation kinetics and porosity or permeability.

15.1 ± 2.1 20.2 ± 1.8 5.3

3.6 ± 0.7

0.8 1.1 1.3 1.5 1.1 1.6 1.3 ± ± ± ± ± ± ± 2.8 3.6 4.3 4.1 4.7 4.9 2.4 1.0 1.8 1.1 1.0 1.4 1.5 1.3 ± ± ± ± ± ± ±

3.2.2. Comparison with a reference: natural carbonation Fig. 2 summarizes the carbonation depths of concretes with CEM I, CEM II/A-V and CEM II/A-LL (with and without MK), after 1 and 2 years of exposure to natural conditions. Globally, the same trends were obtained as with accelerated carbonation, but with higher relative dispersion of the results: – An increase in the carbonation depth when MK was used with CEM I and CEM II/A-V (fly ash). – Good behavior of a clinker-limestone filler-metakaolin system, with a decrease of the carbonation depth compared to CEM II/A-LL, and carbonation equivalent to that of CEM I. 4. Discussion

/

± ± ± ± ± ± ±

1.9 2.0 1.3 2.1 1.4 1.9 1.2

10.8 13.7 10.7 14.8 11.8 21.1 10.4

± ± ± ± ± ± ±

2.0 1.7 1.3 2.6 2.2 2.5 1.8 15.8 16.4 21.4 16.5 17.6 21.0 12.8 7.2 / 6.3 13.1 8.6 4.5 6.8 126 ± 2 372 ± 56 67 ± 1 125 ± 2 161 ± 29 148 ± 27 42 ± 11

2.7 3.4 3.8 3.0 3.1 4.1 1.6

2.6 ± 0.7 9.9 ± 1.4 15.1 ± 2.0 9.3 35 ± 15

1.7 ± 0.9

2.2 ± 1.5 2.3 ± 1.5 8.9 ± 1.5 13.3 ± 2.0 9.5 ± 2.3 18.0 ± 2.1 13.6 17.9 134 ± 25 /

1.9 ± 1.4 2.3 ± 1.4

Accelerated carbonation after 365 days of cure (mm) Accelerated carbonation after 28 days of cure (mm) Portlandite content (%) Gas permeability at 365 days (× 10− 12 m2/ s)

Natural carbonation at 365 days (mm)

Natural carbonation at 730 days (mm)

R. Bucher et al.

86 ± 13 11.6 ± 0.1 0.53 42.5

12.0 ± 0.1

83 ± 28 339 ± 51 86 ± 21 153 ± 20 195 ± 51 157 ± 42 43 ± 8 0.4 0.2 0.4 0.2 0.3 0.4 0.2 ± ± ± ± ± ± ± 10.7 14.0 11.8 12.7 12.7 13.7 11.8 0.2 0.4 0.7 0.2 0.6 0.3 0.3 ± ± ± ± ± ± ± 0.53 0.60 0.53 0.60 0.60 0.60 0.53 42.5 52.5 42.5 52.5 52.5 52.5 42.5

11.2 13.5 12.1 13.1 13.8 13.6 11.1

98 ± 12 11.8 ± 0.6 0.53 42.5

11.4 ± 0.7

191 ± 53 206 ± 13 13.4 ± 0.3 12.7 ± 0.4 0.60 0.60 52.5 52.5

14.5 ± 0.7 12.8 ± 0.4

The literature is quite clear on the fact that SCMs used in replacement of a fraction of Portland cement lead to a deterioration of the carbonation performance of the concrete most of the time [21]. However, it is not rare to find commercially available blended cements around the world (especially in Europe) that could be used in concrete intended for construction, without worrying about the carbonation issue. To the authors' knowledge, these cements have not led to major failures of structures as long as the curing of the concrete and the minimum thickness of concrete cover are respected. The aim of this discussion is to argue the use of metakaolin directly in the concrete mix design of concretes in a CO2 environment, in comparison to commercially available cements which have already proved their suitability on the market. Fig. 3 presents the carbonation depths of all twelve concrete mixes of this study, according to the conditions used in the carbonation tests (accelerated or natural, short and long curing times). The four parts of the Figure refer to: accelerated carbonation after a wet cure of (a) 28 days or (b) 365 days, and natural carbonation after a wet cure of 28 days and an exposure of (c) 365 days or (d) 730 days. On each figure, the upper and lower limits represent the extreme values of carbonation obtained on concretes made of commercial blended cements (in conformity with EN 197-1 [15]), accepted in concrete design for all exposure classes of durability, including the carbonation classes XC1 to XC4 [14]. These limits allowed us to situate metakaolin concretes in comparison to blended cements that could be used in

CEM I CEM I PMES CEM II/ALL CEM II/A-V CEM III/A CEM V/A CEM I 15 CEM I 20 CEM I 25 CEM II/ALL 15 CEM II/A-V 15

Water porosity at 365 days (%) Water porosity at 28 days (%) W/B Class

Table 4 Physical, chemical and durability properties of all mixtures.

Gas permeability at 28 days (× 10− 12 m2/ s)

4.1. MK concretes vs. commercially available blended cements

22

Cement and Concrete Research 99 (2017) 18–29

R. Bucher et al.

(a)

(b) Fig. 1. Accelerated carbonation (4% CO2, 55% RH) of mixtures based on (a) CEM I and (b) CEM II/A.

approach to the calculation of service life is presented later (Section 4.4).

common concrete intended for structures subject to CO2 ingress. It can be seen that the upper limit of the commercial blended cements was often due to CEM V or CEM III, while the lower limit was mostly set by CEM I. CEM III and CEM V contained high proportions of GGBS and fly ash-GGBS, respectively, and the significant carbonation was likely due to the low portlandite content (Table 4) as the lime reserve was too low to slow down the CO2 ingress. The general trend of MK in concretes was that the carbonation depth was often inside or not too far beyond the range of blended cements, except for the highest MK content (CEM I with 25% MK). Sometimes, carbonation depths were around the maximum values (slightly higher or slightly lower) but the high dispersion of the results prevents us from affirming that the concretes with 15 and 20% of metakaolin were statistically worse than the other concretes. It seems that the combination of MK and CEM I (or CEM II/A-V) led to the least suitable mixtures in terms of carbonation. However, the ternary blend composed of clinker, limestone filler and metakaolin was, as already mentioned, very efficient against carbonation. The limitation of SCMs due to carbonation issues is thus open to discussion, and only long-time behavior in terms of service life time could provide valid arguments about whether or not the use of large amounts of SCMs, such as MK, is appropriate regarding carbonation. An

4.2. Particular case of the combination of MK and limestone filler The only mixture based on metakaolin that showed a carbonation depth less than its reference (CEM II/A-LL) was the mixture CEM II/ALL 15. Results of this work showed a beneficial synergy between metakaolin and limestone filler, as confirmed by the very high strength of this concrete. These results are in agreement with Antoni et al.'s work [22], where the good behavior is explained by an increase in the degree of metakaolin reaction. The better compressive strength of the CEM II/ A-LL 15 compared to the CEM II/A-LL could mean a denser porous network that would slow down CO2 diffusion. Antoni et al. [22] showed that the reaction between limestone filler and metakaolin formed hemicarboaluminates. This reaction could be a second explanation of the decrease in CO2 penetration, as Damidot et al. [23] stated that, when the CO2 concentration increased in the pore solution, hemicarboaluminates were destabilized and firstly changed to monocarboaluminates and then into calcite. The hemicarboaluminates could thus be considered as a CO2 sump (like C-S-H and portlandite). 23

Cement and Concrete Research 99 (2017) 18–29

R. Bucher et al.

(a)

(b) Fig. 2. Natural carbonation of mixtures based on (a) CEM I and (b) CEM II/A.

resistance to CO2 ingress in natural conditions generally also performed well under accelerated carbonation at 4% CO2. The exceptions concerned mainly concrete based on SCMs having slow kinetics of reaction (e.g. GGBS or fly ash) and where the curing time had a significant effect on the results. This was the case for the concrete made of CEM V/A, which performed much better in accelerated conditions after a cure of 365 days.

4.3. Correlation between natural and accelerated carbonation for MK and other blended cements Fig. 4 presents the relationships between the different carbonation conditions (accelerated at 4% CO2 vs. natural). A trend seemed to emerge but the correlations were not clear enough to systematically validate the accelerated carbonation test as representative of natural carbonation. Natural carbonation should always be used because acceleration could completely alter the reality of the phenomenon. However, it is sometimes difficult to wait for a long time before obtaining results. In order to verify whether the classification of the different concretes subjected to accelerated carbonation was representative of reality (for instance is a concrete that is efficient in accelerated carbonation conditions still efficient in the case of natural carbonation?), the twelve designs were classified according to the carbonation depth for each carbonation condition (ranked 1 for the smallest carbonation depth and 12 for the greatest carbonation depth). Then, correlation coefficients were calculated based on carbonation depth rankings between accelerated and natural conditions: the higher the correlation between the rank of each formulation according to the carbonation type, the higher the confidence that the accelerated test could be used to compare the different concretes. Table 5 gives the rankings of designs according to the carbonation type, and the correlation coefficients calculated when comparing accelerated and natural conditions. It can be seen that the correlations were fairly good, meaning that the formulations having a good

4.4. Modelling of the long-term carbonation On the one hand, accelerated tests allow a quick comparison of CO2 penetration to be made among several concrete designs but may not be compatible with the estimation of the service life of a structure (e.g. the time before the CO2 reaches the steel bars). On the other hand, tests in natural conditions are slow and give very low carbonation depths within the duration of a research project. Thus, considering that an experimental approach is not sufficiently complete to conclude on the suitability of using SCMs such as MK in carbonated concretes, a carbonation model fitted on our experimental results was used to evaluate the time for the CO2 to reach the steel bars in concrete with or without SCMs. 4.4.1. Model for the calculation of CO2 ingress in concretes In order to evaluate the time before the CO2 reached the steel bars in a concrete, a predictive model, based on natural carbonation results was used. This physical carbonation model [16,24,25] takes the 24

Cement and Concrete Research 99 (2017) 18–29

R. Bucher et al.

(a)

(b)

(c)

(d)

Fig. 3. Comparison of carbonation depth between reference concretes and metakaolin-based concretes. (a) 70-d accelerated carbonation (4% CO2, 55% RH), after an initial wet cure of 28 days; (b) 70-d accelerated carbonation (4% CO2, 55% RH), after an initial wet cure of 365 days; (c) natural carbonation at 365 days, after an initial wet cure of 28 days; (d) natural carbonation at 730 days, after an initial wet cure of 28 days.

25

Cement and Concrete Research 99 (2017) 18–29

R. Bucher et al.

a

b

c

d

Fig. 4. Relationships between the different carbonation conditions (square: samples without metakaolin; triangle: samples with metakaolin).(a) 70-d accelerated carbonation (28-d wet curing) vs. 365-d natural carbonation (28-d wet curing).(b) 70-d accelerated carbonation (28-d wet curing) vs. 730-d natural carbonation (28-d wet curing).(c) 70-d accelerated carbonation (365-d wet curing) vs. 365-d natural carbonation (28-d wet curing).(d) 70-d accelerated carbonation (365-d wet curing) vs. 730-d natural carbonation (28-d wet curing).

the carbonate area (m2/s) (parameters fitted with the experimental results by using a least squares method)P0: CO2 pressure on the concrete surface (Pa) (P0 = 3039 Pa)t: time (s),R: universal constant of perfect gases (J/ mol/K)(R = 8.31 J/mol/K)T: temperature (K) (T = 293 K)β: fitted parameter independent of the materials (β = 7.76)C2: calcium content in the CS-H (mol/l of cement paste), calculated with the chemical composition of the cement and the C/S ratio of the C-S-HPatm: Atmospheric pressure (Patm = 101,325 Pa)φp: volumetric fraction of paste in concrete (liters of cement paste/l of concrete)(φp = 0.27)n: fitted parameters independent of the materials (n = 0.67)Q1: calcium content in portlandite, AFt and AFm (mol/l of cement paste), calculated from the chemical composition and

microstructure and the concrete chemical composition data into account. However, it does not consider cracks in the concrete, which could favor CO2 ingress. As shown by Eq. (2), the model depends on the square root of the time, the CO2 partial pressure and the quantity of calcium that could be affected by CO2 (e.g. Ca in portlandite, C-S-H and AFm).

xc (t ) =

0 2. DCO . P0. t 2

(

R. T . 1 + β. C 2.

P0 n Patm

( ) ). (

φp . C 2 n+1

.

P0 n Patm

( )

+ Q1

)

(2)

With:xc(t): carbonation depth (m)DCO20: coefficient of diffusion of CO2 in Table 5 Rankings of concretes in terms of resistance to carbonation (1: most resistant; 11: least resistant), for each of the carbonation conditions. Correlation index between the different types of carbonation (based on the classification). Cement

CEM I

% of MK

0

15

CEM II/A-LL

CEM II/A-V

CEM I PMES

CEM III/A

CEM V/A

20

25

0

15

0

15

0

0

0

Classification of carbonation depths for each formulation, by type of carbonation Accelerated carbonation (4% CO2, 55% RH) 28 days 1 6 7 Accelerated carbonation (4% CO2, 55% RH) 365 days 1 9 6 Natural carbonation 365 days 3 6 7 Natural carbonation 730 days 1 8 10

10 11 11 11

3 2 2 4

2 3 1 3

5 8 8 6

4 5 5 5

9 10 9 7

8 7 4 2

11 4 10 9

Correlation index between the different types of carbonation (based on the classification) Between accelerated carbonation at 28 days and natural carbonation at 365 days Between accelerated carbonation at 28 days and natural carbonation at 730 days Between accelerated carbonation at 365 days and natural carbonation at 365 days Between accelerated carbonation at 365 days and natural carbonation at 730 days

0.88 0.73 0.78 0.65

26

Cement and Concrete Research 99 (2017) 18–29

R. Bucher et al.

C = 0.6. – The CEM II/A-LL 15 (clinker-limestone filler-metakaolin) formulation gave a service life with > 300 years, almost 50 years more that the CEM II/A-LL alone. – The service life of the CEM III/A, CEM I 15 and CEM I 20 was considerably lower than the CEM I, at around 100 years. – The time for C02 to reach 30 mm was the shortest with CEM I 25, the service life being about 70 years.

stoichiometry of the cements) The values of Q1 and C2 were calculated from Eqs. (3) and (4):

Q1 = (CH + 4 AFm + 6 AFt ) × φp

(3)

C 2 = 1.65 CSH

(4)

The quantities of hydrates were obtained by solving the equation system 5 [14]:

⎧ CH + 1.65 CSH + 4 AFm + 6 AFt (or 3 C3 AH6 ) = α CaO ⎪ ⎪ CSH = α SiO2 ⎨ 2 AFt (or 2 C3 AH6 ) + 2 AFm = 2α Al2 O3 ⎪ 3 AFt (or 0 C AH ) + AFm = α SO 3 6 3 ⎪ ⎩ α = 1 − exp (−3.3 × W C )

Despite the significant decrease in the time necessary for CO2 to reach 30 mm in the cases where MK was used in replacement of CEM I, it must be said that the ages calculated remained far above the actual service life of common structures.

(5)

In the case of concrete based on metakaolin (Al2O3·2SiO2, i.e. AS2), the system in Eq. (5) was coupled with Murat's equation [26] (Eqs. (6) to (8)). According to the CH/AS2 ratio value, the quantities of C-S-H and calcium aluminate (in mol) formed by the pozzolanic reaction of the metakaolin were calculated.

CH AS2 = 1 CH AS2 = 1.67

CH AS2 = 2

4.4.4. Modelling vs. accelerated carbonation The carbonation results calculated with the model were compared with the accelerated carbonation tests in order to verify that the two progressed in the same direction. To do this, it was necessary to find the time taken for natural carbonation to reach the carbonation depth measured in the accelerated test after an exposure of 70 days at 4% CO2 and a wet cure of 28 days (9.5 mm in the CEM I case). An exposure time of 32 years under natural carbonation was found to correspond to the same carbonation depth as that after 70 days of accelerated carbonation. Then, all carbonation results in accelerated conditions were compared with the carbonation depth in natural conditions (calculated) after 32 years' exposure (Table 7). It can be seen that the results of the calculation are in good agreement with the measured ones. Only concretes based on CEM II/A-LL gave different results, probably due to the fact that hemicaboaluminate formation was not taken into account. Generally speaking, it seemed that 70-d accelerated carbonation carried out in a 4% CO2 environment were equivalent to 32 years' carbonation in natural conditions. Results showed a linear relationship between the accelerated carbonation and the calculated natural carbonation, as it was the case in Section 4.3, giving the comparison based on experimental results. This linear tendency can also be found in the literature [28,29]. Yoon et al. showed that it exist a linear relation between the carbonation depth and [CO2∗ temps and this relation is true with three rate of CO2. More the CO2 rate is important more the increase of the carbonation depth is important also. The results of this present work are similar than the Rozière et al. work. They showed a linear relation between the carbonation depth in natural conditions and the carbonation depth in accelerated conditions. Sisomphon et al. [30] showed that the accelerated carbonation was approximately 10 times faster than the natural carbonation, while in our case, the increase was about 170 times. This important difference could be explained by the exposition of the sample during the test of natural carbonation. In this study, the samples were exposed to an outdoor environment (for the natural carbonation) with many variations of climatic conditions. So, during natural carbonation tests, the saturation degrees of the samples could reach high values or a low values. These maximum and minimum values correspond to conditions where the carbonation kinetic is very low.

(6)

AS2 + 3CH + 6H → C2 ASH8 + C‐S‐H

(7)

AS2 + 5CH + 3H → C3AH 6 + C‐S‐H

(8)

AS2 + 6CH + 9H → C4 AH13 + C‐S‐H

4.4.2. Determination and calculation of the parameters of the model In order to compare the impact of metakaolin as a Portland cement substitute with a reference concrete and with commercial blended cements, seven designs were used for the modelling: CEM I, CEM I 15, CEM I 20, CEM I 25, CEM III/A, CEM II/A-LL and CEM II/A-LL 15. All parameters for these concretes are summarized in Table 6. 4.4.3. Prediction of the carbonation behavior 4.4.3.1. Carbonation depth after 50 years. The model was first used to calculate the state of carbonation of concretes after 50 years of exposure, as this age is often used as the design service life of common structures [27]. The evolution of the carbonation depth calculated up to 50 years is given in Fig. 5 for the seven mixtures with and without MK. As expected and in conformity with the experimental results, the replacement of fractions of CEM I (w/ b = 0.6) by MK led to the highest carbonation depths (between 21 and 26 mm). CEM III/A (w/b = 0.6) containing GGBS followed the same trend. The combination of clinker, limestone filler and metakaolin confirmed its interest, the carbonation depth of the CEM II/A-LL 15 reached the same depth than the CEM II/A-LL (w/b = 0.53). It can be noted that, whatever the design considered, the carbonation did not exceed 30 mm. Knowing that, in usual buildings, the steel bars are placed at a depth of about 50 mm [4], it can be concluded, according to the application of the model, that the concrete designs of this study would not be deteriorated before the end of the building's service life. 4.4.3.2. Age to reach a carbonation depth of 30 mm. The model was applied to determine, for the seven concretes, the time needed for the CO2 to reach 30 mm (typical concrete cover). The results showed that: – The CEM I gave the longest service life of the mixtures with a W/ Table 6 Parameters of the model.

DCO20 R C2 Q1

CEM I

CEM I 15

CEM I 20

CEM I 25

CEM III/A

CEM II/A-LL

CEM II/A-LL 15

11.9 · 10− 8 8.31 1.37 1.51

18.30 · 10− 8 8.31 1.51 0.73

16.40 · 10− 8 8.31 1.53 0.56

20.90 · 10− 8 8.31 1.54 0.51

22.5 · 10− 8 8.31 1.74 0.91

10.16 · 10− 8 8.31 1.05 1.34

5.61 · 10− 8 8.31 1.07 0.84

27

Cement and Concrete Research 99 (2017) 18–29

R. Bucher et al.

Fig. 5. Application of the natural carbonation model for a life span of 50 years for concrete with a W/B = 0.53 (cement CEM II/A) and with a W/B = 0.6 (cement CEM I and CEM III/A). Table 7 Comparison between experimental 70-d accelerated carbonation (4% CO2) and calculated 32-year natural carbonation. Composition

CEM CEM CEM CEM CEM CEM CEM

I I 15 I 20 I 25 III/A II/A-LL II/A-LL 15

•

Carbonation depth (mm) After 70 days of carbonation at 4% of CO2 (measured)

After 32 years of natural carbonation (calculated)

9.5 16.5 17.6 21.0 16.4 15.1 12.8

9.4 16.6 18.1 20.6 15.9 10.1 9.4

•

5. Conclusion This study aimed at evaluating the effect of substituting flash metakaolin for cement on the carbonation kinetics of concrete. The following conclusions can be drawn:

References

• Metakaolin caused an increase in the carbonation kinetics when the

•

•

very high strength ones) without sacrificing the carbonation durability. The modelling of CO2 ingress into the concretes showed that, when MK was used to replace a fraction of Portland cement (CEM I), the carbonation depth doubled after 50 years (which represents the service life of common structures). However, the carbonation depth did not exceed 30 mm, far less than the 50 mm depth of concrete cover generally used to protect the steel bars. So the structures including MK would not be deteriorated within the service life of the building. According to the model, the time necessary for CO2 to reach a depth of 50 mm lies between > 800 years (for Portland cement alone, or a combination of clinker, limestone filler and metakaolin) and 200 years (for the mixture of Portland cement and 25% of metakaolin). Despite the significant decrease in the time necessary for CO2 to reach 50 mm when MK was used in replacement of CEM I, it must be said that the ages calculated remain far above the actual service life of common structures.

[1] J.P. Ollivier, A. Vichot, La durabilité des bétons, Ponts et chaussées, (2008). [2] R. San Nicolas, Approche performantielle des bétons avec métakaolins obtenus par calcination flash, PhD thesis in French, Toulouse (2011). [3] H.S. Kim, S.H. Lee, H.Y. Moon, Strength properties and durability aspects of high strength concrete using Korean metakaolin, Constr. Build. Mater. 21 (2007) 1229–1237. [4] D.O. McPolin, P.A.M. Basheer, A.E. Long, K.T.V. Grattant, T. Sun, New test method to obtain pH profiles due to the carbonation of concretes containing supplementary cementitious materials, J. Mater. Civ. Eng. 19 (2007) 936–946. [5] R. Mejia de Gutierrez, C. Rodriguez, E. Rodriguez, J. Torres, S. Delvasto, Metakaolin concrete: carbonation and chloride behavior, Rev. Fac. Eng.-Univ. Antioquia 48 (2009) 55–64. [6] A. Bouikni, R.N. Swamy, A. Bali, Durability properties of concrete containing 50% and 65% slag, Constr. Build. Mater. 23 (2009) 2836–2845. [7] E. Gruyaert, P. Van Den Heede, M. Maes, N. De Belie, A comparative study of the durability of ordinary Portland cement concrete and concrete containing (high) percentages of blast-furnace slag, International RILEM Conference on Material Science, RILEM Publications, Bagneux, 2010, pp. 241–251. [8] V.M. Malhotra, M.H. Zhang, R.H. Read, J. Ryell, Long-term mechanical properties and durability characteristics of high-strength/high-performance concrete incorporating supplementary cementing materials under outdoor exposure conditions, ACI Mater. J. 97 (2000) 518–525. [9] P. Sulapha, S.F. Wong, T.H. Wee, S. Swaddiwudhipong, Carbonation of concrete containing mineral admixtures, J. Mater. Civ. Eng. 15 (2003) 134–143. [10] K.H. Khayat, P.C. Aitcin, Silica fume in concrete—an overview, SP-132: Fly Ash, Silica Fume, Slag and Natural Pozzolans in Concrete, 1993, pp. 835–872. [11] M.D.A. Thomas, J.D. Mattews, Carbonation of fly ash concrete, Mag. Concr. Res. 44 (1992) 217–228. [12] J. Bai, S. Wild, B.B. Sabir, Sorptivity and strength of air-cured and water-cured PC–PFA–MK concrete and the influence of binder composition on carbonation depth, Cem. Concr. Res. 32 (11) (2002) 1813–1821.

comparison was made with a concrete based on CEM I (> 95% of clinker): the higher the metakaolin content, the greater the carbonation depth. The increase in the carbonation kinetics was due to the consumption of portlandite by the pozzolanic reaction of metakaolin. The general trend detected for the use, in a CO2 environment, of concretes having metakaolin directly in the mix design in comparison to commercially available blended cements (including fly ash, GGBS or limestone filler) which had already proved their worth on the market, was that the carbonation depth was often within, or not too far outside, the range for blended cements, except for the highest MK content (25%). The combination of MK and limestone filler (in CEM II/A-LL with 15% MK) showed very good behavior against carbonation, with a carbonation depth almost equivalent to that of the CEM I samples. This means that a combination of clinker, limestone filler and MK could behave as well as CEM I alone regarding carbonation. The beneficial synergy between metakaolin and limestone filler was confirmed by the very high strength of this concrete, and was probably due to the formation of hemicarboaluminates, which might be considered as a CO2 sump (like C-S-H and portlandite) and thus limit CO2 ingress. This could provide a real opportunity to decrease the clinker content in concretes (on other concretes than 28

Cement and Concrete Research 99 (2017) 18–29

R. Bucher et al. [13] R. San Nicolas, M. Cyr, G. Escadeillas, Characteristics and applications of flash metakaolins, Appl. Clay Sci. 83-84 (2013) 253–262. [14] NF EN 206/CN, Béton—spécification, performance, production et conformité—Complément national à la norme NF EN 206, (2014). [15] EN 197-1, Cement—Part 1: Composition, Specifications and Conformity Criteria for Common Cements, (2001). [16] N. Hyvert, A. Sellier, F. Duprat, P. Rougeau, P. Francisco, Dependency of C–S–H carbonation rate on CO2 pressure to explain transition from accelerated tests to natural carbonation, Cem. Concr. Res. 40 (2010) 1582–1589. [17] S. Salvador, Pozzolanic properties of flash-calcined kaolinite: a comparative study with soak-calcined products, Cem. Concr. Res. 25 (1995) 102–112. [18] G. Dreux, Nouveau guide du béton, Eyrolles, [in French], (1985). [19] NF P18-459, Concrete—Testing Hardened Concrete—Testing Porosity and Density, (2010). [20] XP P 18-463, Concrete—Testing Gas Permeability on Hardened Concrete, (2011). [21] F. Pacheco Torgal, S. Miraldo, J A., J. De Brito Labrincha, An overwiew on concrete carbonation in the context of eco-efficient construction: evaluation, use of SCMs and/or RAC, Constr. Build. Mater. 36 (2012) 141–150. [22] M. Antoni, J. Rossen, F. Martirena, K. Scrivener, Cement substitution by a combination of metakaolin and limestone, Cem. Concr. Res. 42 (2012) 1579–1589.

[23] D. Damidot, S. Stronach, A. Kindness, M. Atkins, F.P. Glasser, Thermodynamic investigation of the CaO-Al2O3-CaCO3-H2O closed system at 25°C and the influence of NaO, Cem. Concr. Res. 24 (1994) 563–572. [24] M. Thiery, G. Villain, P. Dangla, G. Platret, Investigation of the carbonation front shape on cementitious materials: effects of the chemical kinetics, Cem. Concr. Res. 37 (2007) 1047–1058. [25] B. Bary, A. Sellier, Coupled moisture—carbon dioxide–calcium transfer model for carbonation of concrete, Cem. Concr. Res. 34 (2004) 1859–1872. [26] M. Murat, A. Bachiorrini, Corrélation entre l'état d'amorphisation et l'hydraulicité du métakaolin, Bulletin de mineralogie, Chapitre 3: Etude de la carbonatation de matrices cimentaires contenant du métakaolin, 105 1982, pp. 543–555. [27] M. Karimpour, M. Belusko, K. Xing, F. Bruno, Minimising the life cycle energy of buildings: review and analysis, Build. Environ. 73 (2014) 106–114. [28] E. Rozière, A. Loukili, F. Cussigh, A performance based approach for durability of concrete exposed to carbonation, Constr. Build. Mater. 23 (2009) 190–199. [29] I. Yoon, O. Copuroglu, K. Park, Effect of global climatic change on carbonation progress of concrete, Atmos. Environ. 41 (2007) 7274–7285. [30] K. Sisomphon, L. Franke, Carbonation rates of concretes containing high volume of pozzolanic materials, Cem. Concr. Res. 37 (2007) 1647–1653.

29