Use of sugarcane bagasse ash sand (SBAS) as corrosion retardant for reinforced Portland slag cement concrete

Use of sugarcane bagasse ash sand (SBAS) as corrosion retardant for reinforced Portland slag cement concrete

Construction and Building Materials 226 (2019) 72–82 Contents lists available at ScienceDirect Construction and Building Materials journal homepage:...

2MB Sizes 1 Downloads 87 Views

Construction and Building Materials 226 (2019) 72–82

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Use of sugarcane bagasse ash sand (SBAS) as corrosion retardant for reinforced Portland slag cement concrete Fernando C.R. Almeida a,b,c, Almir Sales a,⇑, Juliana P. Moretti a,d, Paulo C.D. Mendes e a

Department of Civil Engineering, Federal University of São Carlos, Via Washington Luís, km 235, São Carlos, SP 13565-905, Brazil Institute for Technological Research. Av. Prof. Almeida Prado, 532, Cidade Universitária, São Paulo, SP 05508-901, Brazil c Department of Materials Engineering and Construction, Federal University of Minas Gerais, Av. Presidente Antônio Carlos, 6627, Belo Horizonte, MG 31270-901, Brazil d Sea Institute, Federal University of São Paulo, Rua Dr. Carvalho de Mendonça, 144, Santos, SP 11070-100, Brazil e São Carlos Institute of Chemistry, University of São Paulo, Av. Trabalhador São-carlense, 400, São Carlos, SP 13560-970, Brazil b

h i g h l i g h t s

g r a p h i c a l a b s t r a c t

 SBAS increases the compressive

strength and chloride penetration resistance of concrete.  Hydration of slag cements can be activated by SBAS.  SBAS delays the depassivation of reinforcing steel (corrosion retardant).  Incorporation of SBAS up to 30% leads to a corrosion propagation similar to that observed for the reference concrete.

a r t i c l e

i n f o

Article history: Received 12 January 2019 Received in revised form 4 July 2019 Accepted 19 July 2019

Keywords: Supplementary cementitious materials Sugarcane bagasse ash sand (SBAS) Corrosion initiation Corrosion propagation

a b s t r a c t Despite their growing popularity in the construction industry, cement-based materials containing blastfurnace slag may be subject to various deteriorative processes including corrosion susceptibility owing to the reduced alkalinity of the cementitious matrix. Sugarcane bagasse ash sand (SBAS), the main agroindustrial waste in sugar and ethanol production, has the potential to enhance concrete durability. This study aims to evaluate the effect of SBAS as a fine aggregate on the corrosion of reinforced slag cement concrete. Mixes with different proportions were analysed in terms of mechanical characteristics, chloride penetration, electrolytic conductivity of pore solution, corrosion potential (by open circuit potential), and corrosion rate (by polarisation curves and Tafel plot technique), as well as the analysis of the corroded steel bars (by scanning electron microscopy technique). This paper argues the efficiency of SBAS in retarding corrosion up to an optimal content of approximately 30%. SBAS can refine micropores and act as an alkaline activator of slag cements, leading to increased compressive strength, reduced chloride penetration depth, and delayed corrosion initiation. Above 30%, the corrosion rate increases, which could be owing to the modified microstructure and higher electrolytic conductivity of SBAS-concrete pore solutions, that control the susceptibility of chloride attack on the steel bars. Ó 2019 Elsevier Ltd. All rights reserved.

1. Introduction

⇑ Corresponding author. E-mail address: [email protected] (A. Sales). https://doi.org/10.1016/j.conbuildmat.2019.07.217 0950-0618/Ó 2019 Elsevier Ltd. All rights reserved.

The durability of reinforced concrete used in major infrastructure such as bridges, tunnels, ports, and offshore constructions can be severely compromised by corrosion processes. Consequently,

73

F.C.R. Almeida et al. / Construction and Building Materials 226 (2019) 72–82

such damage reduces the service life of the overall structure and significantly increases maintenance and repair costs. In the United States, it is estimated that more than $500 billion (US) is necessary for rehabilitation of bridges and highway systems, in which the majority of degradation problems are related to corrosion of reinforced concrete [1,2]. This problem refers not only to the corrosion of reinforcing steel but also to the conditions of protection that the concrete itself creates around a steel rebar [3,4]. According to Tuutti’s model [5], the corrosion process in concrete structures comprises two main stages: initiation and propagation. The first is determined by changes in the concrete cover as a result of chemical and physical actions. This cover is responsible for protecting the reinforcement against neutralising or activating substances with concentrations favourable for corrosion. Such corrosion begins when deleterious agents such as CO2 and chlorides penetrate the concrete, usually through diffusion processes, until the passive environment closest to the steel is counteracted [6]. When this occurs, the corrosion mechanisms of the reinforcing steel are initiated, and the propagation stage begins. In turn, the rate of attack is determined both by the rate of anode-cathode reactions and the physical contact between reaction areas [7,8]. As a result, corrosion-induced damages (e.g. concrete stain, cracking, spalling, delamination, and cross-section reduction of the reinforcement) may cause aesthetic deficiency or a decrease in the structural load-bearing capacity which can lead to the collapse of the structure [1,9]. The durability of reinforced concrete can be optimized through adequate selection of sustainable materials [10]. It is widely accepted that supplementary cementitious materials, such as blast-furnace slag, reduce energetic costs and CO2 emissions from cement production, improve concrete performance and reduce wastes from other industries [11,12]. Because of the latent hydraulic property activated by alkalis/lime, slags can effectively replace up to 95% of ordinary Portland cement [13,14]. However, slags generally contain fewer alkali compounds and lower hydroxide concentrations than pure clinker, leading to reduced pH in cementitious systems [15,16]. Thus, Portland slag cements can have a strong effect on the alkalinity of concrete, reducing the protecting ability against corrosion initiation [17]. In this context, incorporation of alternative materials as fine aggregates can lead to enhanced durability by means of pore refinement and increased resistance to the penetration of deleterious agents [18]. Sugarcane bagasse ash sand (SBAS), a by-product of sugar and ethanol production, is a potential substitute for natural sand in cementitious materials [19,20]. Brazil is the world’s largest sugarcane producer and exporter with an estimated SBAS generation of 4 million ton/year [21,22]. Previous studies have shown the potential benefits of SBAS in reducing macropores larger than 0.1 lm and significantly increasing resistance to chloride penetration in cementitious materials [20,23]. Moreover, the carbonation depths of mortars with up to 30% of SBAS (with respect to the total sand content) were statistically equivalent to reference samples [20]. However, the influence of SBAS on concrete durability in terms of corrosion development, particularly in combination with supplementary cementitious materials is still unclear and deficient. Therefore, this study aims to evaluate the effect of SBAS as a fine aggregate in the corrosion process of reinforced Portland slag cement concrete. Mixes with different SBAS contents were evalu-

ated in terms of mechanical characteristics and electrochemical analysis such as corrosion potential and corrosion rate under an accelerated chloride attack. 2. Materials and methods In this study, SBAS was employed as a fine aggregate for partial replacement of natural quartz sand to produce slag cement concrete. Three SBAS levels of substitution, in mass, were considered: 0% (reference), 30%, and 50%. SBAS samples were collected from sugarcane plants in the state of São Paulo, Brazil, and were submitted to a homogenisation treatment by sieving with a #4.8 mm sieve and grinding for 3 min in a mechanical mill [19,20]. The chemical compositions of SBAS samples before and after homogenisation treatment are listed in Table 1. Although the equivalent alkali value (Na2O + 0.658 * K2O) for SBAS (2.8%) is relatively high [24], this content is not harmful for deleterious expansions triggered by aggregate alkali reactions [25]. Cementitious matrices with 30% and 50% of SBAS had expansions in the same order of the reference samples, with values far below the limits required by the Brazilian standard [25,26]. Even after the homogenisation treatment, SBAS had a predominantly crystalline structure of SiO2 a-quartz, as shown by the Xray diffraction (XRD) (Fig. 1), previously analysed in [19,20]. The XRD pattern presents the characteristic peaks of the crystalline phase and the absence of amorphous halo, justifying the technical motivation to replace quartz sand by SBAS. The scanning electron microscopy (SEM) micrographs (Fig. 2) illustrate the size difference between both aggregates studied. Table 2 lists the physical characterisation of fine and coarse aggregates used for concrete production. All samples were prepared using Portland composite cement with 34%, in weight, of blast-furnace slag (CPII-E32 [28] similar to CEM II/B-S [13]). Physical and chemical characterisation of Portland slag cement is listed in Table 3. CA-60 steel bars (5 mm diameter and 100 mm length) were used as working electrodes for electrochemical analysis. For cleaning purposes, the steel bars were immersed for 10 min in a 1:1 HCl aqueous solution with 3.5 g/L of hexamethylenetetramine to control the acid attack. Then, the bars were brushed, rinsed with deionised water and immersed in acetone for 2 min. Immediately after drying under a heat gun, the bars were coated with galvanic tape to delimit the electrode-exposed area (40 mm length cylinder side) for chloride corrosion (Fig. 3). The compositions of concrete mixtures are listed in Table 4, from which three series of concrete with different SBAS contents were produced [19]. The water/cement ratio (w/c) was slightly adjusted for each composition to maintain the same consistency [29], because incorporation of SBAS requires more water in the mixture [18,30]. This procedure avoided the use of superplasticizer, which could interfere the electrochemical analysis considered in this study, since techniques to evaluate corrosion may be very sensitive to different salts present in superplasticizer’s compositions [31]. Prismatic samples (50  70  100 mm3) were prepared for electrochemical analysis, considering two treated steel bars (Fig. 3) embedded in each specimen (Fig. 4). Also, cylindrical samples were

Table 1 Chemical composition of SBAS samples in mass, % Component

SiO2

Fe2O3

K2O

Al2O3

CaO

MgO

P2 O5

TiO2

Na2O

MnO

SO3

Loss on ignition

SBAS before homogenisation SBAS after homogenisation

80.2 80.8

5.6 5.8

4.0 3.9

2.6 2.5

1.8 1.6

1.6 1.5

1.4 1.4

1.4 1.3

0.2 0.2

0.2 0.1

0.1 0.1

0.8 0.7

74

F.C.R. Almeida et al. / Construction and Building Materials 226 (2019) 72–82

Fig. 1. XRD of SBAS used in this study (blue) after homogenisation treatment (red, Q = quartz) [29]. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 2. SEM micrographs of fine aggregates: (a) quartz sand; (b) SBAS.

Table 2 Characterisation of aggregates used for concrete production. Property

SBAS

Quartz sand

Basalt gravel

Specific gravity Water absorption (%) Maximum dimension (mm) Fineness modulus Quantity of powdery material finer than 75 mm (No. 200 sieve) after washing (%) pH (measured in aqueous solution 1:1, using a pH-meter, according to [27] – Method A)

2.57 0.9 1.18 1.15 16.2

2.45 0.5 6.3 2.32 0.35

2.63 0.3 19 1.49 –

10.4

6.2



prepared for mechanical and chloride penetration tests. All concrete specimens were compacted by using a vibrating table, demoulded after 24 h, and cured in climate chamber for 28 days at a temperature of 21 ± 2 °C and relative humidity of 95 ± 5% [32,33]. Each concrete composition was tested in triplicate for all tests. The concrete samples were subjected to aging cycles for 22 weeks. Each cycle comprised a week which was divided into a wetting semi-cycle in 3.5% NaCl solution for 2 days and a drying semi-cycle in an oven at 50 ± 5 °C for 5 days [33]. Table 5 lists the adopted experimental programme according to specimen type, exposure conditions, and age of testing. Analysis of

Table 3 Physical and chemical characterisation of Portland slag cement – CPII-E32 (Source: manufacturer). Property

CPII-E32

Component

CPII-E32 (in mass, %)

Specific gravity Blast-furnace slag content (%) Initial setting time (min) Final setting time (min) Fineness: % retained #200 Blaine specific surface (cm2/g) Compressive strength (fcj): 3 days (MPa)

3.02 34.0 197 279

SiO2 Al2O3 Fe2O3 CaO MgO SO3 Na2O K2O Carbonic anhydride – CO2 Insoluble residue LOI 1000 °C

24.1 7.4 3.1 52.1 3.4 2.2 0.3 1.1 1.5

7 days (MPa) 28 days (MPa)

3.60 4,141 18.5 26.8 39.5

2.5 4.1

variance (ANOVA) and Student’s t-test at a significance level of 5% were considered for all testing results. The compressive strength test was performed at 28 days (after curing) and 182 days (after aging cycles), and carried out by a servo-mechanical press, at a loading rate of 0.45 ± 0.15 MPa/s, according to [34]. The compressive strength was obtained from the average of the ratios between the maximum load of rupture and the cylindrical cross-sectional area of the specimen.

75

F.C.R. Almeida et al. / Construction and Building Materials 226 (2019) 72–82

Fig. 3. Steel bar preparation: (a) immersion in hydrochloric acid and hexamethylenetetramine solution; (b) brushing in deionised water; (c) immersion in acetone; (d) coating with galvanic tape.

Table 4 Mixture composition of concrete samples. Sample

SBAS content (%)

CR C30 C50

0 30 50

Mixture proportion (in mass)

Slump test (mm)

Cement

Sand

SBAS

Gravel

Water

1.000 1.000 1.000

2.010 1.407 1.005

– 0.603 1.005

2.860 2.860 2.860

0.520 0.530 0.540

90 ± 5 87 ± 5 85 ± 5

Fig. 4. Prismatic moulds of reinforced concrete for electrochemical analysis: (a) positioning of steel bars; (b) concrete-filled moulds with embedded steel bars; (c) open circuit potential measurement.

Table 5 Experimental programme for concrete samples. Testing

Standard

Specimen type

Exposure condition

Age of testing

Compressive strength

[34]

Tensile strength Elastic modulus – Method A Chloride penetration by colorimetric method with AgNO3 and fluorescein Electrolytic conductivity of pore solution

[35] [36] [37]

Cylindrical Cylindrical Cylindrical Cylindrical Cylindrical

Curing Curing Curing Curing Curing

At At At At At

Corrosion potential by open circuit potential

[39]

Aqueous solution extracted from powder of concrete specimens Prismatic

Curing followed by drying of concrete specimens Curing followed by aging cycles

Corrosion rate by polarisation curves Pitting depth by SEM analysis

[40] [41]

Prismatic Extracted steel bars

Curing followed by aging cycles After aging cycles embedded in concrete

[38]

For the tensile strength test, cylindrical specimens were subjected to diametrical compression according to [35] at 28 days (after curing). The test was performed using a servo-mechanical press at loading rate of 0.05 ± 0.02 MPa/s until the rupture. The tensile strength was determined by Eq. (1).

ft ¼

2P

pDh

;

ð1Þ

only followed by aging cycles only only followed by aging cycles

28 days 182 days 28 days 28 days 182 days

At 28 days During 22 weeks after curing At 182 days –

where P is the maximum load of rupture (N), D is the diameter of the cylindrical specimen (mm), and h is the height of the cylindrical specimen (mm). The elastic modulus by compressive stress was determined according to [36] (Method A) at 28 days (after curing). Extensometers were positioned on cylindrical specimens for the strain measurement. The load was applied by a servo-mechanical press at a loading rate of 0.25 ± 0.05 MPa/s, until the stress corresponding

76

F.C.R. Almeida et al. / Construction and Building Materials 226 (2019) 72–82

to 30% of the concrete compressive strength (rb) and kept for 60 s. After that, the load was reduced to a stress of 0.5 MPa (ra) and kept for 60 s. This process was repeated twice, and, in the last cycle, the strain was checked for ra and rb, resulting ea and eb. The elastic modulus was determined by Eq. (2).



rb  0:5  103 eb  ea

3. Results and discussion

ð2Þ 3.1. Mechanical properties

where rb is the stress corresponding to 30% of the concrete compressive strength (MPa), ea is the strain at ra = 0.5 MPa, and eb is the strain at rb. The chloride penetration test was conducted using the colorimetry method by sprinkling silver nitrate and fluorescein solution [37]. At 182 days (after 28 days of curing followed by 22 aging cycles), concrete samples were dried at 105 ± 5 °C and then broken. The surface obtained by diametrical rupture of the specimen was treated with 0.1 M silver nitrate (AgNO3) solution and fluorescein alcoholic solution (1% fluorescein salt, 29% distilled water, 70% ethanol). A white precipitate (AgCl) was formed after the sprayed surface dried. This whitish region, in contrast to the darker area, indicates the presence of chloride ions. The depth of this whitish area was measured using a digital pachymeter from the outer edge of the specimen, which indicates the chloride penetration depth in the concrete specimen. Electrolytic conductivity was analysed for the pore solution extracted from finely macerated concrete samples. The extractions were prepared with a 1:1 proportion (in mass) between the solid samples and ultrapure water (2.37 mS/cm), during 24 h with occasional agitation, followed by vacuum filtration. The coarse aggregate was not included, since it is expected to release a negligible quantity of ions [38]. A lab bench conductivity meter, with cell constant of 0.75 cm1, was used for the measurement at room temperature. The corrosion potential was estimated through the open circuit potential, which was measured in triplicate based on [33,39]. Three reinforced concrete (prismatic specimens with two embedded steel bars each) were continuously monitored by means of the open circuit potential (Eoc) using a high impedance input voltmeter (Minipa ET-2076). Measurements started at 28 days after the curing period and were conducted twice a week for both wetting and drying conditions at the end of each semi-cycle. It was assumed that the standardisation of the testing condition was obtained due to the steel bars preparation (Fig. 3) [33]. At the end of 22 cycles, or 182 days after casting, all working electrodes were in an active state of corrosion, i.e. the Eoc values were inferior to 254 mV vs. the reference electrode of Ag/AgCl/KClsat [39]. At this stage, the polarisation curves were obtained by using a potentiostat/galvanostat PGSTAT mod. 20 Autolab, using the Ag/ AgCl/KClsat reference electrode and an auxiliary titanium electrode. This technique is a perturbation method with linear scanning around Ecorr ± 350 mV at 0.5 mV/s [4,40]. The corrosion current density (icorr), in lA/cm2, which is directly proportional to the rate of corrosion, was calculated by Eq. (3) considering the corrosion potential (Ecorr) and corrosion current (Icorr) obtained from Evans diagrams and Tafel plots [42,43]:

icorr ¼

Icorr S

vacuum mode, accelerating voltage (EHT) of 15 kV, and working distance (WD) of 8 mm. The level of corrosion was characterised by measuring the deepest point of the pitting penetration of each working electrode [44].

ð3Þ

where Icorr is the corrosion current intensity (lA), and S is the exposed surface area of the working electrode (15.83 cm2). Afterwards, the steel bars were extracted from the concrete specimens and were cleaned following the same procedure, as shown in Fig. 3. The pitting depths of extracted corroded bars were analysed through the Scanning Electron Microscopy (SEM) technique (EVO 50, Carl Zeiss AG, Oberkochen, German), according to [41]. The samples were not pre-coated, and observations were carried out in high

The results of compressive strength at 28 and 182 days, before and after aging cycles, as well as those of tensile strength and elastic modulus at 28 days are listed in Table 6. At 28 days, the incorporation of SBAS increased the compressive strength by around 8% and 11% for concretes with 30% and 50% of quartz sand replacement, respectively. At 182 days, the addition of SBAS resulted in increments of 4% and 14% (absolute values) for C30 and C50, respectively, relative to the reference concrete (CR). The considerable increases in strength between both ages must be attributed to prolonged cement reaction owing to the blastfurnace slag in the cement composition in addition to the continuous drying/wetting cycles in the saline solution. The differences between the compressive strength results at 28 days (no aging cycles) are statically significant. The SBAS incorporation indeed increased this property even with a higher w/c ratio for concretes with SBAS. This increment can be attributed to the packing effect of SBAS in the mixture. According to previous studies [20,23], SBAS particles have filled pores in size ranges of 60–80 lm and 100–300 lm, resulting in a denser microstructure and hence increased compressive strength. The increased strength can also be related to the interaction of SBAS and Portland composite cement with blast-furnace slag [45]. The reaction of slag itself with Portland cement and water is a complex process. It is well known that the main constituents of ordinary Portland cement (PC) are calcium silicates (Ca3SiO5 and Ca2SiO4), aluminate (Ca3Al2O6), and ferrite (Ca4(AlxFe1-x)4O10) which can be abbreviated to C3S, C2S, C3A, and C4AF. A number of other minerals such as calcium sulphates (present as gypsum, anhydrite and/or hemihydrate), calcite, calcium oxide, magnesium oxide, Na- and K-sulphates are also usually present. These components react with water to form various hydration products such as calcium silicate hydrate (C–S–H), portlandite (CH), ettringite, calcium monosulphoaluminate or calcium monocarboaluminate [14,46]. Slag, due to its high alumina and silica content, produces somewhat more complex hydrates than PC. In hydrated PC-slag systems, the presence of C–S–H, CH, ettringite, AFm (monosulphate and monocarbonate), and a hydrotalcite-like phases is observed. However, hydration of PC-slag cements produces less portlandite (at long term), less ettringite (at high slag content) and less AFm and AFt phases (as more Al is bound in C–S–H) than in pure PC hydration. Usually, the C–S–H formed in PC–slag blends has a lower Ca/Si ratio and a higher Al/Si ratio than pure PC [10]. Introducing alumina into the C–S–H phase, to form C–A–S–H, markedly increases its alkali-binding capacity and, hence, reduces alkalinity of the pore solution [47]. In the presence of water, slag hydrates to a limited degree. A protecting film deficient in calcium is quickly formed, which slows down further reaction. If pH is kept sufficiently high in presence of activators, slag hydration can be accelerated and intensified. The most common activator of slag is PC clinker, but also other alkaline materials [48]. In this context, SBAS can act as an alkaline activator, considering its relatively high equivalent alkali value (Na2O + 0.658 * K2O = 2.8%, from Table 1) and higher pH (Table 2) compared to quartz sand. Thus, secondary C–S–H phases can be formed as hydration products of SBAS-PC-slag concrete, which are responsible for lower permeability and higher strength.

77

F.C.R. Almeida et al. / Construction and Building Materials 226 (2019) 72–82 Table 6 Results of mechanical tests of concrete. Sample

Compressive strength (at 28 days) Average (MPa)

CR C30 C50 ANOVA and ttest (1) (2)

(1)

SD (MPa)

CV (%)

36.54 1.77 5 39.36 1.53 4 40.65 0.67 2 CR/C30 & CR/C50 statistically significant

(2)

Compressive strength (at 182 days) Average (MPa)

(1)

SD (MPa)

CV (%)

51.26 1.53 3 53.51 2.94 5 58.74 1.95 3 CR/C30/C50 not statistically significant

(2)

Tensile strength (at 28 days) Average (MPa)

(1)

SD (MPa)

CV (%)

3.75 0.11 3 3.70 0.52 14 4.21 0.58 14 CR/C30/C50 not statistically significant

Elastic modulus (at 28 days) (2)

Average (GPa)

SD (1) (GPa)

CV (2) (%)

36.96 0.23 1 39.26 1.73 4 37.76 0.65 2 CR/C30/C50 not statistically significant

SD: standard deviation. CV: coefficient of variation.

Therefore, when slag is mixed with water, its hydration products form a thin Si-rich layer on the surface of the slag grains, which stifles further hydration [49,50]. In turn, together with portlandite (CH) and alkali hydroxides from Portland cement hydration, SBAS can act as slag’s activator. Owing to its higher alkalinity compared with quartz sand (Table 2), SBAS raises the pH in the vicinity of the slag and prevents the formation of the Si-rich layer. As a result, SBAS can improve the hydration degree of Portland slag cement and increase the compressive strength of the concrete. However, according to ANOVA results, the null hypothesis cannot be rejected for the results of compressive strength at 182 days (F-statistic < f-critical). Although C30 and C50 reached greater absolute values than CR, the specimens subjected to degradation cycles had statistically equivalent compressive strength regardless of the SBAS content. NaCl salts were crystallised in the concrete pores, leading to increased compactness in all samples, as shown in Fig. 5, which led to an increase in the masses of specimens over the cycles. This effect was caused by the gradual accumulation of salts into the concrete microstructure in addition to the hygroscopic nature of NaCl, which led to increased water retention. However, the mass increment ceased, and the mass variation stabilised after approximately 16 weeks. This can indicate a ‘saturation’ of the salt crystallisation effect, in which the microstructure clogging by NaCl potentially overcame the SBAS packing. Thus, the compressive strength values were statistically equivalent for concretes with and without SBAS. The results of tensile strength and elastic modulus at 28 days were not affected by the SBAS incorporation with F-statistic < fcritical (p-value > 0.05). However, the evidence was insufficient for rejecting the null hypothesis by ANOVA test. Thus, the concrete

with SBAS had similar performance for those properties when compared with the reference sample. Therefore, SBAS enhanced the mechanical properties by increasing compressive strength at 28 days. Tensile strength and elastic modulus, as well as compressive strength after degradation cycles were not influenced by the SBAS incorporation. 3.2. Chloride penetration Fig. 6 shows the results of colorimetric testing of chloride penetration in concrete after 22 aging cycles of drying and wetting in NaCl solution. This qualitative method indicates the presence of free chloride ions which indeed represent a risk for susceptibility of reinforcement corrosion. Analyses through ANOVA and Student’s t-test indicated that all averages were statistically different. The incorporation of SBAS had a significant influence on the chloride penetration depth in concrete such that a higher SBAS level in the mixture corresponded to a lower chloride front depth. This result is attributed to the physical and chemical effects of SBAS. The physical effect can be related to the filling of micropores with SBAS particles, particularly pores with dimensions smaller than 150 mm, and the reduction of capillary interconnectivity by adding a finer material in the mix [20,51]. Moreover, ionic diffusion can be considered as the main transport mechanism of chloride ions in concrete with SBAS. The presence of other ions in the pore solution, such as Na+, K+, Ca2+, and OH, affects the presence of Cl, accelerating or slowing its penetration via mechanisms of repulsion or attraction among the electrical charges [52,53]. Thus, incorporation of finer particles of SBAS affects the transport mechanisms and thus acts as a barrier against chloride ion penetration.

Fig. 5. Variation in mass of concrete specimens during aging cycles. Higher (lower) values refer to saturated (dried) conditions.

78

F.C.R. Almeida et al. / Construction and Building Materials 226 (2019) 72–82

Fig. 6. (a) Results of chloride penetration depth and (b) colorimetric treatment applied to concrete specimens.

The chemical effect, in turn, can be attributed to slag activation by SBAS. The higher pH of SBAS (compared to quartz sand, Table 2) contributes to activate hydration of Portland-slag cement, resulting in the creation of secondary C–S–H. This leads to refinement of the pore structure and, hence, reduced permeability in the concrete [53,54]. Chlorides bind to aluminate phases or are adsorbed on C–S–H [6]. Therefore, in comparison with the reference concrete, the incorporation of SBAS had a positive effect on the resistance to chloride penetration.

portions of aggregates, cement consumption, the presence of chemical admixtures, degree of hydration, microstructure and pore density, rebar passivity, alkalinity of the pore solution); environmental characteristics (temperature and relative humidity) and the action of aggressive agents (chloride penetration and carbonation) [4,33]. Thus, all properties should be carefully evaluated in conjunction before reaching a reliable conclusion regarding the performance expected with SBAS incorporation.

3.3. Electrolytic conductivity of pore solution

3.4. Corrosion initiation

To improve the understanding of the solution that can be present in the pores of the concrete samples, the electrolytic conductivities of solutions extracted from finely macerated samples (as an indirect approximation to the pore solution) were compared. The results are shown in Fig. 7. There is positive correlation between the conductivity of the extracted solution and the content of SBAS in the concrete; the conductivity increased by 14% and 23% for a SBAS incorporation of 30% and 50%, respectively, with respect to the reference without SBAS. The conductivity is a summation of contributions from all ions [38]; however, the increase can be attributed mostly to the higher concentration of OH expected in the presence of SBAS, according to the higher associated pH compared with quartz sand (Table 2). In fact, OH are among the most conductive ions found in concrete pore solutions [55,56]. The higher conductivity could be expected to facilitate corrosion due to the increased mobility of reactive species, i.e., allowing the flow of electrons and hence corrosion of the reinforcement bars. However, it is not possible to ascertain the performance of these materials solely by this measurement, since many other factors affect the permeability of species and corrosion, such as the characteristics of the concrete (water/cement ratio, type and pro-

The results of open circuit potential (Eoc) measured during the degradation cycles in the saline solution are shown in Fig. 8. Because the system was allowed time to stabilise before each measurement, the Eoc could be a satisfactory approximation to the corrosion potential for each condition of, specifically the anodic and cathodic reactions should have the same rate. After 20 weeks, the probability of corrosion of all working electrodes was expected to be greater than 90% for all samples with Eoc values below 254 mV versus Ag/AgCl (saturated KCl) [39]. The outliers observed during the first week can be attributed to the dramatic change in environment from a humidity chamber with curing at 95% relative humidity to aging cycles in a saline solution. However, during the following weeks, all measurements stabilised, indicating a trend related to the wetting and drying cycles to which the concrete samples were subjected. The variations in the Eoc measurements for each semi-cycle were attributed to the moisture content of the specimens, directly affecting the ohmic drop in the concrete. Specifically, less negative Eoc values were obtained after drying semi-cycles in which the electrical resistivity or the difficulty of ion movement was higher, whereas the Eoc was more negative when the samples were saturated with the solution. This is because the minimized ohmic drop under saturation condition led to lower electrical resistivity, resulting in more negative Eoc values [57]. Overall, during the passivation period, working electrodes presented Eoc values between 160 mV and 30 mV. At the time of depassivation (Eoc < 254 mV), a sharp decline in Eoc values was verified for all samples, indicating the beginning of corrosion propagation in the steel bars [33]. The working electrodes in the reference concrete were the first bars to begin corrosion propagation, at week 14, followed by concretes C50 and C30 at weeks 16 and 18, respectively. The delay in corrosion activation in the steel bars may be related to the packing effect of SBAS. The substitution proportion of 30% must have contributed as an ideal condition to close or interrupt the connectivity between pores [20,23]. This created a physical barrier against the attack of chloride ions [58]. Moreover, the increment of alkalinity in the concrete by SBAS may have contributed to delay corrosion initiation. Concrete samples CR, C30

Fig. 7. Results for the conductivity of the aqueous solution extracted from the cementitious matrix simulating pore solution.

F.C.R. Almeida et al. / Construction and Building Materials 226 (2019) 72–82

79

Fig. 8. Evolution of the open circuit potential of working electrodes versus Ag/AgCl/KClsat in concrete specimens.

and C50 respectively had pH values of 11.93, 12.32 and 12.33, measured by this research group from the solubilised extracts [19,25] obtained according to the Brazilian standard [59]. The susceptibility to pitting corrosion in an acid solution containing primarily chlorides increased with the decreased alkalinity provided by the passivation cover [6,60], i.e. the balance between alkalinity and acidity, given by the activity of OH and Cl, respectively, is responsible for maintaining the passivation of the reinforced rebar. In summary, the beginning of the corrosion process in reinforced concrete was delayed by substituting 30% of the quartz sand with SBAS, owing to its physical influence in refining micropores and its higher alkalinity. 3.5. Corrosion propagation The linear polarisation curves are shown in Fig. 9, where the current values (I) are plotted against the overpotential (g), obtained by sweeping at 0.5 mV/s, from the lower limit around the Eoc following the relation g = Eoc ± 350 mV. Analogous to the interpretation of Eoc, the Ecorr indicates the electrochemical dynamic equilibrium, in which both the anodic and cathodic currents have the same magnitude and thus the resultant total current is null. From this point, the anodic and cathodic terms are unbalanced by the application of overpotentials, where the polarisation resistance and current propagation can be interpreted. Although ohmic drop is one of the main error sources of this technique [4,57], it was minimised by employing samples which were already corroded after conditioning cycles (Fig. 8) and were fully saturated with NaCl solution electrolytes. Active dissolution of the working electrode was noted for all samples, as indicated by the upward sloping curve in relation to the overpotential axis. In particular, sample C50 presented the highest current values according to its highest slope, both in the anodic branch (positive values of g) and cathodic branch (negative values of g) indicating greater dissolution of the metal. Specimens CR and C30 had profiles similar to each other when compared with C50, although their current was significantly lower. It is crucial to keep in mind that the current is proportional to the electrode area. Although the exposed geometric area was the same for all samples, the unique corrosion process that occurred on each material greatly affects this measurement, hence, a higher active area could be expected for C50 to explain the higher currents. The corrosion

Fig. 9. Polarisation curves of working electrodes in concrete (linear scale).

rates are directly proportional to the corrosion current, i.e. higher values of electrical current passing through the system indicate that the corrosion is more severe. Thus, by a qualitative analysis, it can be inferred that the corrosion propagation was greater for C50 and was similar for CR and C30. To determine the corrosion current, the polarisation curves were plotted in a logarithm scale (Evans diagram, Fig. 10). From the intersection of the extended cathodic and anodic branches on the Tafel slope, the corrosion potential (Ecorr) and corrosion current, or more accurately, the logarithm of the corrosion current (log |Icorr|) were found. This result, in turn, can be extrapolated to corrosion rate analysis by considering the following factors: (i) generalised corrosion, which is perfectly uniform along the surface; (ii) simple dissolution as the only anodic reaction, written as Fe0 ? Fe2+ + 2e; and (iii) Faraday’s constant. The results of the corrosion current density (icorr) of each working electrode are listed in Table 7. The results emphasise the similarities between CR and C30 samples, which had the same order of corrosion current density. Because they were in the range of 0.5–1.0 lA/cm2, both rates could be classified as moderate levels but not severe [44]. However, C50 clearly had the greatest corrosion propagation compared with the

80

F.C.R. Almeida et al. / Construction and Building Materials 226 (2019) 72–82

Fig. 10. Evans diagrams and Tafel plots: polarisation curves of working electrodes in concrete (logarithm scale).

other two samples, with a rate about 10 times greater than the rate of the reference sample. C50 can be characterised as have a high level of corrosion because its icorr value was greater than 1.0 lA/ cm2 [44]. It is worthwhile to note that the approximation of the corrosion rate is affected by how localised the corrosion is. As mentioned above, this corrosion rate analysis is extrapolated by considering a generalised corrosion, which is perfectly uniform along the steel surface. However, the presence of pitting, cracks and other surface

Fig. 11. SEM micrographs for typical pitting depth measurements. The inserts show the pointed cavities of each working electrode (reinforcing steel bars) in detail.

Table 7 Results of corrosion parameters of working electrodes in concrete samples. Sample

Ecorr (mV)

log |Icorr|

Icorr (mA)

icorr (mA/cm2)

Pitting depth by SEM (mm)

Classification by [44]

CR C30 C50

600.2 663.2 686.6

4.9 4.9 3.9

11.6 13.7 121.2

0.7 0.9 7.7

242.4 238.0 460.4

moderate moderate high

F.C.R. Almeida et al. / Construction and Building Materials 226 (2019) 72–82

irregularities need to be evaluated carefully to further understand the results. In this context, the typical pitting depths of corroded steel bars were characterised by the SEM micrography technique (Fig. 11). The steel bars embedded in concrete C50 had the greatest surface irregularities (pitting depth) in comparison with the other samples (Table 7), indicating a more localised corrosion. If the remaining areas of the steel bar are considered passive, a high corrosion current divided by a smaller area (more localised) results in a higher corrosion current density. This technique completes the understanding of the corrosion profile, because the analysed working electrodes did not incur uniform and generalised degradation. As expected, the corrosion was localised, which is characteristic of the chloride action and indicated by the cavities on the bar surface. Pitting corrosion can be more dangerous than uniform deterioration because it progressively reduces the cross-sectional area of the rebar to a point at which it can no longer withstand the applied load, leading to severe structural failure [4]. Increments of corrosion kinetics in concrete with SBAS, particularly at contents above 30%, can be related to its increased electrolytic conductivity (Fig. 7). Moreover, the presence of some salts in the matrix, such as NaCl from saline solution during the aging cycles, accelerates the electrochemical reactions on the steel bars during corrosion progress. This is owing to their capacities for depolarising and, consequently, increasing the electrical charge passing through the concrete as an effect of dissolved salts. In this context, particularly C50 had the greatest mass variation on the wetting and drying cycles (Fig. 5), indicating a higher water absorption capacity (including saline solutions) and, hence, a greater ability of NaCl crystallisation. Thus, incorporation of higher SBAS levels of more than 30% can contribute to faster propagation of the corrosive process, especially due to the modified microstructure of the cementitious matrix and increased ionic mobility in the mix which leads to greater electrolytic conductivity. Although SBAS may retard the beginning of the corrosive process, as indicated by Eoc measurements, this by-product contributed to accelerating the corrosion propagation as soon as it began. This negative effect was even more pronounced in concrete with 50% SBAS. Therefore, incorporation of SBAS at levels up to 30% can reduce corrosion probability, as a retardant of corrosion initiation, and keep corrosion propagation at the same rate as the reference sample. Moreover, it has been well identified that not only the corrosion rate is important but also how localised it is. Due to the limited number of samples to deduce a ‘‘Tuutti’s prediction” [5] and adverse conditions in field, the outcomes presented herein may not occur in practice. However, this study is representative of general trends that can be expected and is important for basing the advances of the current knowledge on reinforcement corrosion of SBAS concrete. 4. Conclusion On the basis of the results obtained in this experimental study, the following conclusions were drawn:  The substitution of quartz sand by SBAS increases both the compressive strength and resistance to chloride penetration of concrete. This positive outcome can be attributed mainly to a packing effect and the ability of SBAS to act as an alkaline activator to Portland-slag cement. The tensile strength and elastic modulus are not significantly affected by the incorporation of SBAS.  The incorporation of SBAS leads to decreased probability of corrosion, owing to the delayed depassivation of the reinforcing steel. This effect is related to increased concrete alkalinity and decreased pores connectivity promoted by SBAS. Concrete with

81

approximately 30% of SBAS seems to indicate an appropriate mixture for retarding the initiation of corrosive processes.  Once corrosion began, the propagation rate is higher for concrete with SBAS contents greater than 30%, due to the modified microstructure and increased conductivity of its pore solution. It appears that a SBAS content of about 30% may result in a physical and electrochemical balance between closing micropores (or interrupting pores connectivity) and satisfying an adequate conductivity of the pore solution that controls the susceptibility of chloride attack on the steel bars. This equilibrium prevents the propagation of higher corrosion currents and creates a propagation condition comparable to the behaviour for a sample without SBAS. For future studies, evaluating different SBAS concentrations around 30% is recommended in order to accurately identify the optimal content that leads to an enhanced concrete performance. Therefore, in addition to environmental benefits of reusing an agro-industrial by-product, the incorporation of SBAS at levels up to 30% can improve the mechanical properties and enhance the durability of reinforced concrete. The capacity of SBAS to retard corrosion initiation in an environment susceptible to chloride attack adds value to this new eco-friendly construction material. Declaration of Competing Interest There is no conflict of interest. Acknowledgments The authors acknowledge CAPES 001 and CNPq (Grant numbers 309892/2013-9 and 409685/2017-8) for financial support, Cosan for SBAS supplies, and Prof Lucia H. Mascaro from Laboratório Interdisciplinar de Eletroquímica & Cerâmica (LIEC/UFSCar) for electrochemical measurement support. References [1] A. Michel, M. Otieno, H. Stang, M.R. Geiker, Propagation of steel corrosion in concrete: Experimental and numerical investigations, Cem. Concr. Compos. 70 (2016) 171–182. [2] AASHTO, Transportation Bottom Line Report - Executive Version, American Association of State Highway and Transportation Officials, Washington, D.C, 2015. [3] D.V. Ribeiro, J.C.C. Abrantes, Application of electrochemical impedance spectroscopy (EIS) to monitor the corrosion of reinforced concrete: A new approach, Constr. Build. Mater. 111 (2016) 98–104. [4] S. Ahmad, Reinforcement corrosion in concrete structures, its monitoring and service life prediction - a review, Cem. Concr. Compos. 25 (2003) 459–471. [5] K. Tuutti, Corrosion of steel in concrete, Swedish Cement and Concrete Research Institute, Stockholm, 1982. [6] L. Bertolini, B. Elsener, P. Pedeferri, E. Redaelli, R.B. Polder, Corrosion of Steel, Concrete Prevention, Diagnosis, Repair, 2nd ed., 2013. [7] C.L. Page, Degradation of reinforced concrete: Some lessons from research and practice, Mater. Corros. 63 (2012) 1052–1058. [8] C. Andrade, A. Cesetti, G. Mancini, F. Tondolo, Estimating corrosion attack in reinforced concrete by means of crack opening, Struct. Concr. 17 (2016) 533– 540. [9] Y.S. Ji, G. Zhan, Z. Tan, Y. Hu, F. Gao, Process control of reinforcement corrosion in concrete. Part 1: Effect of corrosion products, Constr. Build. Mater. 79 (2015) 214–222. [10] B. Lothenbach, K. Scrivener, R.D. Hooton, Supplementary cementitious materials, Cem. Concr. Res. 41 (2011) 1244–1256. [11] K.L. Scrivener, B. Lothenbach, N. De Belie, E. Gruyaert, J. Skibsted, R. Snellings, A. Vollpracht, TC 238-SCM: hydration and microstructure of concrete with SCMs, Mater. Struct. 48 (2015) 835–862. [12] N. De Belie, M. Soutsos, E. Gruyaert, eds., Properties of Fresh and Hardened Concrete Containing Supplementary Cementitious Materials: State-of-the-Art Report of the RILEM Technical Committee 238-SCM, Working Group 4, 1ed., RILEM & Springer, 2017. [13] BSI, BS EN 197-1. Cement. Part 1: Composition, specifications and conformity criteria for common cements, 2011. [14] K.L. Scrivener, P. Juilland, P.J.M. Monteiro, Advances in understanding hydration of Portland cement, Cem. Concr. Res. 78 (2015) 38–56.

82

F.C.R. Almeida et al. / Construction and Building Materials 226 (2019) 72–82

[15] A. Vollpracht, B. Lothenbach, R. Snellings, J. Haufe, The pore solution of blended cements: a review, Mater. Struct. 49 (2016) 3341–3367. [16] F.C.R. Almeida, A.J. Klemm, Effect of GGBS on water absorption capacity and stability of Superabsorbent polymers partially crosslinked with alkalis, J. Mater. Civ. Eng. 30 (2018) 1–11. [17] V.K. Ortolan, M. Mancio, B.F. Tutikian, Evaluation of the influence of the pH of concrete pore solution on the corrosion resistance of steel reinforcement, J. Build. Pathol. Rehabil. 1 (2016) 1–7. [18] P. Shafigh, H. Bin Mahmud, M.Z. Jumaat, M. Zargar, Agricultural wastes as aggregate in concrete mixtures - a review, Constr. Build. Mater. 53 (2014) 110– 117. [19] A. Sales, S.A. Lima, Use of Brazilian sugarcane bagasse ash in concrete as sand replacement, Waste Manag. 30 (2010) 1114–1122. [20] F.C.R. Almeida, A. Sales, J.P. Moretti, P.C.D. Mendes, Sugarcane bagasse ash sand (SBAS): Brazilian agroindustrial by-product for use in mortar, Constr. Build. Mater. 82 (2015) 31–38. [21] J.P. Moretti, A. Sales, F.C.R. Almeida, M.A.M. Rezende, P.P. Gromboni, Joint use of construction waste (CW) and sugarcane bagasse ash sand (SBAS) in concrete, Constr. Build. Mater. 113 (2016) 317–323. [22] OECD/FAO, OECD-FAO Agricultural Outlook 2015-2024, OECD Publishing, Paris, 2015. [23] J.P. Moretti, A. Sales, V.A. Quarcioni, D.C.B. Silva, M.C.B. Oliveira, N.S. Pinto, L.W. S.L. Ramos, Pore size distribution of mortars produced with agroindustrial waste, J. Clean. Prod. 187 (2018) 473–484. [24] ABNT, NBR 12653 - Materiais pozolânicos - Requisitos. (Pozzolanic materials Requirements) [in Portuguese], 2015. [25] S.A.L. Bessa, Utilização da cinza do bagaço da cana-de-açúcar como agregado miúdo em concretos para artefatos de infraestrutura urbana. (Utilization of sugarcane bagasse ash as fine aggregate in concretes for urban infrastructure artifacts), Universidade Federal de São Carlos (2011). PhD Thesis [in Portuguese]. [26] ABNT, NBR 15577-1 - Agregados - Reatividade álcali-agregado. Parte I. (Aggregates - Alkali-aggregate reactivity. Part 1: Guide for the evaluation of potential reactivity of aggregates and preventive measures for its use in concrete) [in Portuguese], 2008. [27] ASTM, D 4972/01 - Standard test method for pH of soils, 2007. [28] ABNT, NBR 16697 - Cimento Portland - Requisitos (Portland cement Requirements) [in Portuguese], 2018 [29] ABNT, NBR NM 67 - Concreto - Determinação da consistência pelo abatimento do tronco de cone (Concrete - Slump test for determination of the consistency) [in Portuguese], 1996 [30] M.V. Madurwar, R.V. Ralegaonkar, S.A. Mandavgane, Application of agro-waste for sustainable construction materials: a review, Constr. Build. Mater. 38 (2013) 872–878. [31] E. Arif, M.W. Clark, N. Lake, Sugar cane bagasse ash from a high-efficiency cogeneration boiler as filler in concrete, Constr. Build. Mater. 151 (2017) 692– 703. [32] ABNT, NBR 5738 - Concreto - Procedimento para moldagem e cura de corpos de prova (Concrete - Procedure for molding and curing concrete test specimens) [in Portuguese], 2015 [33] D.V. Ribeiro, J.A. Labrincha, M.R. Morelli, Effect of the addition of red mud on the corrosion parameters of reinforced concrete, Cem. Concr. Res. 42 (2012) 124–133. [34] ABNT, NBR 5739. Concreto - Ensaios de compressão de corpos-de-prova cilíndricos (Concrete - Compression test of cylindric specimens - Method of test) [in Portuguese], 2007. [35] ABNT, NBR 7222 - Argamassa e concreto - Determinação da resistência à tração por compressão diametral de corpos-de-prova cilíndricos (Mortar and concrete - Determination of the tensile stregth of cylindrical specimens subjected to diametrical compression) [in Portuguese], 2011. [36] ABNT, NBR 8522. Concreto – Determinação do módulo estático de elasticidade à compressão (Concrete - Determination of the elastic modulus by compression) [in Portuguese], 2008.

[37] AASHTO, T 259 - Standard method of test for resistance of concrete to chloride ion penetration, 2002. [38] C. Shi, Effect of mixing proportions of concrete on its electrical conductivity and the rapid chloride permeability test (ASTM C1202 or ASSHTO T277) results, Cem. Concr. Res. 34 (2004) 537–545. [39] ASTM, C876-09 - Standard test method for corrosion potentials of uncoated reinforcing steel in concrete, 2009. [40] ASTM, G5-94 - Standard reference test method for making potentiostatic and potentiodynamic anodic polarization measurements, 2011. [41] ASTM, G46-94. Standard guide for examination and evaluation of pitting corrosion, 2013. [42] V. Yegnaraman, C. Ahmed Basha, G. Prabhakara Rao, Application of the accelerated Tafel plot technique to corrosion kinetics: Role of double layer through simulation, J. Appl. Electrochem. 18 (1988) 869–875. [43] M.E. Ismail, H.R. Soleymani, Monitoring corrosion rate for ordinary portland concrete (OPC) and high-performance concrete (HPC) specimens subjected to chloride attack, Can. J. Civ. Eng. 29 (2002) 863–874. [44] C. Andrade, C. Alonso, On-site measurements of corrosion rate of reinforcements, Constr. Build. Mater. 15 (2001) 141–145. [45] V.N. Castaldelli, J.L. Akasaki, J.L.P. Melges, M.M. Tashima, L. Soriano, M.V. Borrachero, J. Monzó, J. Payá, Use of slag/sugar cane bagasse ash (SCBA) blends in the production of alkali-activated materials, Materials (Basel). 6 (2013) 3108–3127. [46] B. Lothenbach, F. Winnefeld, Thermodynamic modelling of the hydration of Portland cement, Cem. Concr. Res. 36 (2006) 209–226. [47] S.-Y. Hong, F.P. Glasser, Alkali sorption by C-S-H and C-A-S-H gels: Part II. Role of alumina, Cem. Concr. Res. 32 (2002) 1101–1111. [48] M. Thomas, Supplementary Cementing Materials in Concrete, CRC Press, Boca Raton, 2013. [49] M. Thomas, The effect of supplementary cementing materials on alkali-silica reaction: A review, Cem. Concr. Res. 41 (2011) 1224–1231. [50] F.C.R. Almeida, A.J. Klemm, Efficiency of internal curing by superabsorbent polymers (SAP) in PC-GGBS mortars, Cem. Concr. Compos. 88 (2018) 41–51. [51] J.P. Moretti, S. Nunes, A. Sales, Self-compacting concrete incorporating sugarcane bagasse ash, Constr. Build. Mater. 172 (2018) 635–649. [52] C.P. Figueiredo, F.B. Santos, O. Cascudo, H. Carasek, P. Cachim, A. Velosa, The role of metakaolin in the protection of concrete against the deleterious action of chlorides, IBRACON Struct. Mater. J. 7 (2014) 685–708. [53] R. Loser, B. Lothenbach, A. Leemann, M. Tuchschmid, Chloride resistance of concrete and its binding capacity - comparison between experimental results and thermodynamic modeling, Cem. Concr. Compos. 32 (2010) 34–42. [54] R. Siddique, Utilization (recycling) of iron and steel industry by-product (GGBS) in concrete: Strength and durability properties, J. Mater. Cycles Waste Manag. 16 (2014) 460–467. [55] P. Hewlett (Ed.), Lea’s Chemistry of Cement and Concrete, 4ed, Elsevier, London, 2003. [56] P. Azarsa, R. Gupta, Electrical resistivity of concrete for durability evaluation: A review, Adv. Mater. Sci. Eng. 2017 (2017). [57] C. Argiz, M.Á. Sanjuán, P.C. Borges, E. Álvarez, Modeling of corrosion rate and resistivity of steel reinforcement of calcium aluminate cement mortar, Adv. Civ. Eng. 2018 (2018) 1–9. [58] R.E. Núñez-Jaquez, J.E. Buelna-Rodríguez, C.P. Barrios-Durstewitz, C. GaonaTiburcio, F. Almeraya-Calderón, Corrosion of modified concrete with sugar cane bagasse ash, Int. J. Corros. 2012 (2012) 1–5. [59] ABNT, NBR 10006 - Procedimento para obtenção de extrato solubilizado de resíduos sólidos (Procedure for obtaining solubilized extracts from solid wastes) [in Portuguese], 2004. [60] A.A. Mazhar, S.T. Arab, E.A. Noor, The role of chloride ions and pH in the corrosion and pitting of Al-Si alloys, J. Appl. Electrochem. 31 (2001) 1131– 1140.