Introduction

The construction industry is a massive producer of greenhouse gases. For every ton of Portland cement manufactured, one ton of CO₂ is released (Rifaai et al., 2019; Singh; Middendorf, 2019). Over the last decades alkali-activated binders have emerged as an alternative to replace Portland cement in many applications due to their high mechanical strength and low CO₂ emissions (Singh; Middendorf, 2019; Palomo et al., 2014).

Materials alkali-activated are inorganic polymers composed of tetrahedral units of aluminum and silicon condensed at ambient temperature, formed after of the dissolution of amorphous source materials (which can be industrial by-products or heat-treated clays) in the presence of highly alkaline solutions (Provis; Van Deventer, 2009). The incorporation of alkali ions such as Na⁺, K⁺ and Ca²⁺ balances the electrical charges of the structure. The condensation of polymeric structures results in the formation of CaO-Al₂O₃-SiO₂-H₂O (C-A-S-H) gels when high calcium precursors are used, or Na₂O-Al₂O₃-SiO₂-H₂O (N-A-S-H) gels when low calcium precursors are used. This cementitious material increases its mechanical properties over time with the hardens process. That encourages the use of the material as a binder for composite materials such as concrete (Provis; Van Deventer, 2009).

Materials that contain aluminosilicates in their chemical compositions are used as precursors for alkali-activated binders. The most consolidated precursors are fly ash, slag and metakaolin (Xie et al., 2020; Zhang et al., 2020). However, these are not the only kinds, since some of them are poorly explored, becoming waste from industrial processes and without an adequate way of consumption. One of these prominent by-products are the several other kinds of steel slag, generated in the process of transformation of iron into steel, which, according to Nunes and Borges (Nunes; Borges, 2021). New studies indicate that this material can be used as a precursor for alkali-activated binder and thus mitigate the problem (Costa, 2022; Girish; Shetty; Nayak, 2022). The most used activators in alkali-activation process are sodium hydroxide (NaOH), potassium hydroxide (KOH) and sodium silicate (Na₂SiO₃). Sodium hydroxide is the main activator and acts as a catalyst for the reactions, as well as providing OH^- ions, while sodium silicate, which provides soluble silica, participates as the structure-forming element (Xie et al., 2020).

There are no specific standards for the production and durability analysis concerning the criteria and the methods for performance evaluation of these new concretes. Parameters that are established for ordinary Portland cement concretes (PCC) are often used. Alkali-activated concrete (AAC) is a recently developed material, which intends to use waste constituent materials, as well as to reduce the use of Portland cement. Then, there is a need for creation or adaptation of existing durability parameters for it. According to Amorim Júnior (2020), studies that can classify AACs as to their quality for their various applications are still limited, mainly due to the lack of standardization of formulation methodologies and a deficit of research in durability.

Several tests such as: water penetration under pressure (ABNT, 2011), electrical resistivity (ABNT, 2012) and accelerated chloride migration (ASTM, 2019) are used to evaluate durability properties of ordinary PCC. The last one indirectly measures chloride migration from the electric charge passing through the specimen at a pre-set electric potential difference. An alternative to directly evaluate chloride migration would be to measure concentrations before and after the migration process used in the classical test, such as the ion chromatography test (Amorin Júnior, 2020; Cabral, 2000; Yang; Cho; Huang, 2002).

Based on this background, this study aimed to verify the suitability as durability characterization tests for chloride migration, electrical resistivity, and water penetration under pressure in AAC based on fly ash and BOF (Blast Oxygen Furnace) steel slag. Although common in conventional concretes, they might not be suited for AAC. Besides that, a comparison with a Portland cement-based concrete (PCC) was done. It is expected to contribute to the characterization of AAC, which may allow their future application on the field.

Materials

Two precursors were used to produce AAC. Class F fly ash (FA) used unprocessed, from the Pecém Thermoelectric Complex and steel slag (SS) obtained from Basic Oxygen Furnace process. For PCC, a Portland cement CPIII (PC), which is a Portland cement type III by Brazilian standard that contains slag, was used. Table 1 shows the compositions in oxides of the precursors (for ACC) and the Portland cement (for PCC) through the X-Ray Fluorescence (XRF) test. The specific gravity of SS, FA and PC are 3.13, 2.34 and 3.06, respectively. As for particle size, the average particle sizes of SS, FA and PC were 35.4 μm, 5.6 μm and 13.2 μm, respectively.

Figure 1 shows a comparison of the diffraction patterns of SS and FA. The SS showed characteristic peaks of Gypsum (CaSO₄), calcium hydroxide (Ca(OH₂)), berlinite (AlPO₄) and Katoite (Ca₃Al₂O₆(H₂O)₆), while the FA showed crystalline phases of SiO₂ (quartz), Fe₃O₄ (magnetite), Mullite (Al(Al_1.272Si_0.728O_4.864), Brushite (CaPO₃(OH)2H₂O) and Monetite (CaPO₃(OH)). The FA sample also showed an amorphous pattern between 2θ 20 and 35°.

NaOH and Na₂SiO₃ solutions were used as alkali activators (Xie et al., 2020). Sodium hydroxide solution at 10 mol/L concentration (Rafeet et al., 2017) composed by mass of 31.3% NaOH and 68.7% H₂O. The sodium silicate solution is composed by mass of 14.98% Na₂O, 31.83% SiO₂ and 53.19% H₂O.

Silica sand and granitic gravel were used as fine and coarse aggregates. The sand used had a specific gravity of 2.57, a bulk density of 1470.5 kg/m³ and a fineness modulus of 2.66. The gravels used presented specific mass of 2.62 and 2.63 and bulk density of 1406.7 kg/m³ and 1487.2 kg/m³, for the 4.75-12.50 mm (gr. 1) and 9.5-25 mm (gr. 2), respectively.

Aggregate packing curves, both experimental and theoretical (with the assistance of the software EMMA - Elkem Materials Mixture Analyzer - version 3.5.2.11) were elaborated to determine the proportion of aggregates that presented the lowest voids content. The final mass proportions obtained by both methodologies were 35% of sand to 65% of gravel (26% - 1 and 39% - gravel 2). Paste contents were tested to obtain a self-compacting concrete, and the ideal content obtained was 38.4 % (Araújo, 2023).

To compare the test results, a PCC with silica fume (SF) as a partial substitute for cement was produced (concrete which was under development at the same period in the laboratory). The smaller size of the silica fume (0.1 µm and 5.5 µm) in its amorphous form and the SiO₂ content present give to the concrete a reduction in permeability (Mehta; Monteiro, 1994). This comparison was necessary due to the lack of data in the literature for comparison according to the adopted methodology. The mix proportions of the concretes produced is shown in Table 2. The AAC presented a liquid/binder ratio of 0.50 and the PCC of 0.56 with addition of a superplasticizer SP 3100 (SP).

The concrete mixing procedure was performed in a free fall mixer as it follows:

1. mixing the gravel with 10% of activator solution or water for 1 minute;

2. adding binders (precursors or cement) and 80% of activator solution or water followed by mixing for 4 minutes;

3. adding sand and mixing for 4 minutes; and

4. adding the rest of the activator solution or water and mixing for 2 minutes.

Then, the specimens were molded with 10 cm in diameter and 20 cm in height. All the tests for both concretes were performed after 28 days of curing. For conventional concrete, wet curing was performed for 28 days. For alkali-activated concrete, thermal curing was performed for 24 hours in an oven at a temperature of 65 °C (Araújo, 2023; Soutsos et al., 2016), and then ambient curing (24 °C to 31 °C and humidity between 65%-85%, typical in Fortaleza, Ceará) for 27 days (Araújo, 2023). Wet curing was not applied to the AAC as it is not suitable for the material. Immersion curing in water is avoided for alkali-activated materials (AAM) because it can lixiviate the activator and inhibit strength development (Athira et al., 2021; Provis, 2018). As they are materials with different reaction mechanisms, the ideal curing method for each type of concrete was applied. The AAC and PCC presented compressive strenght of 77.1 MPa and 35.3 MPa, respectively.

Methods

The following tests were performed to characterize the durability of the concretes: water penetration under pressure, electrical resistivity, accelerated chloride migration and ion chromatography. The ion chromatography test was performed from the resulting solution of the chloride migration test. A relationship between both tests was established, since the first one calculates the chloride migration indirectly (by integrating the electric current over time), while the second one measures the concentrations of the solutions and produces a direct measurement of this migration. The results of all AAC tests were analyzed and then compared to the chosen PCC. Despite differences in the liquid/binder ratio, powder and aggregate content, and curing methods, which have influence on the durability results of the material, the goal of this work is not just compare the numerical results between the AAC and PCC. Instead, it aims to analyze whether the tests for water penetration under pressure, electrical resistivity, accelerated chloride migration, and ion chromatography present adequate and valid relationships for AAC, as they do for PCC. This aspect will be explored in the results.

Water penetration under pressure

The test is described by the NBR 10787 (ABNT, 2011) and was performed at 28 days after molding. Three specimens with dimensions of 10 cm in diameter and 20 cm in height previously air dried for a period of 24 hours before the start of the test were used. Figure 2 shows the representation of the test, which consisted of applying successive increments of pressure. Initially, it was applied water at a pressure of 0.1 ± 0.01 MPa during the first 48 hours. Then, a pressure of 0.3 ± 0.03 MPa was applied for 24 hours, and finally, water was applied at a pressure of 0.7 ± 0.07 MPa for 24 hours. Then, the specimens were broken by diametrical compression to allow the measurement of the maximum depth of water penetration and the distribution profile of the penetrated water.

Electrical resistivity

The electrical resistivity measures the difficulty of electric current to pass through an object. It can be obtained through the ratio between the applied voltage and the resulting electric current, multiplied by a constant factor (related to geometry). However, resistivity is seen as one of the main parameters to evaluate the corrosion of reinforcement bars inserted in concrete (Hornbostel; Larsen; Geiker, 2013). According to Halliday, Resnick and Walker (2009), resistivity is a property that reflects the ability of the material to carry electrical charges inside the concrete. Thus, this parameter is also defined as the inverse of the material conductivity.

The principle for direct resistivity measurement in the volume of a specimen consists in applying an electric potential difference (voltage) between two or more electrodes, positioned on opposite sides, aligned, and pressed against the opposed surfaces. Four specimens at 28 days were tested. The property is determined from the measured current and the dimensions (diameter and height) of the specimens after rectification. The equipment used for electrical resistivity analysis was the Resipod from Proceq SA (Figure 3).

The relation between the applied voltage and the measured current provides the resistance to the electrical current, according to Ohm's law. The electrical resistivity (ρ) is determined by multiplying the resistance (R) by a correction factor called cell constant (K). This factor depends on the dimensions of the specimen in which the resistivity measurements were taken, being the ratio of the section area by the length of the specimen (RIBEIRO, 2010), as in Equation 1.

Where, ρ represents the electrical resistivity (kΩ.cm); K is the cell constant (cm); R is the reading of the Resipod device and Radjusted = R/2πa (kΩ), where “a” is a distance parameter from the equipment. The values obtained can be compared based on Table 3, which establishes criteria for evaluating the risk of reinforcement bar corrosion in concrete with the measured resistivity (CEB, 1989).

Accelerated chloride migration

For the accelerated chloride migration test, an apparatus with two chambers was used, one with 3% sodium chloride (NaCl) solution and the other with 0.3N sodium hydroxide (NAOH) solution, as suggested by Andrade (1993) and developed by Ribeiro (2010), based on C1202 (ASTM, 2019). Four cylindrical specimens of 10 cm in diameter and 5 cm in height were used. The test consisted in accelerating a chloride flow through an electric field, with voltage of 60 V, applied between the two chambers (one with sodium chloride and the other with sodium hydroxide) of the diffusion cell with the help of two metallic electrodes, as shown in Figure 4. The result was obtained by measuring the electrical current intensity in Amperes every 30 min for 6 hours. The objective is to calculate a charge passed in Coulomb associated with the migration of these chlorides through the specimen, with a time integration.

The test considers a direct relationship between the amount of Coulumbs and the chloride flow. Then a great amount of calculated Coulumbs passing represents a high "permeability" of the concrete to chlorides (Andrade, 1993). Based on this, with the electric charge passed it was possible to classify, according to Table 4, the penetration rate of chlorides for concrete. However, it was noted that such classification was adjusted for concretes that do not have the presence of reinforcing steel, other electrically conductive materials incorporated or specimens with reinforcement positioned longitudinally, since they provide a continuous electrical route between the two edges of the specimen (ASTM, 2019), and this may be the case similar for concretes containing steel slag as discussed in this paper.

Ion chromatography

According to Meyer (2013), ion chromatography is a modern and prominent method for the separation and determination of ions, due to the simplicity to separate, identify and quantify chemical substances. The ion chromatography tests in this paper consisted in using the solutions from the chloride migration test to directly quantify the passage of chloride ions from one side to the other after the test. For this, initial and final concentrations of chlorides in each solution were determined. Samples of NaCl and NaOH solutions were collected before and after the chloride migration test. A 0.5 ml portion of the respective solutions was dissolved in 250 ml of ultra-pure water in a beaker due to the equipment's need for high electrical conductivity correction. The equipment that quantifies these concentrations and allows the direct measurement of chloride migration is the Ion Chromatography System Dionex ICS - 1100 (Figure 5), which identifies the ions based on the retention time of each analyte, determined by integrating the peak area displayed in a graph of results obtained with the chromatograph (Marques, 1999).

Results

Water penetration under pressure

The results of the water penetration under pressure test are presented in Figure 6. The average of water penetration for AAC was 40.7 mm and for PCC was 24.0 mm. Despite the lack of homogeneity among the data between the specimens, considering the average or individually, all results are considered low (Neiville; Brooks, 1997), since concretes with penetration measurements until 50 mm are classified as being "impermeable", and if less than 30 mm, it is classified as "impermeable under aggressive conditions" (Neiville; Brooks, 1997). Helene (1993) also classifies as durable concrete those with void contents less than 10% and absorption less than 4.2%, which is observed in the characterization data of AAC, which presented 9.4% and 4.1% of voids and absorption index, respectively, according to NBR 9778 (ABNT, 2009) and C642 (ASTM, 2021). This fact is a positive indicator of the durability of both tested concretes, since they have, in addition to low void contents, low ability to be crossed by fluid even under pressure. This prevents the action of damaging external agents to the material.

Electrical resistivity

The results are presented in Figure 7. The average electric resistivity for AAC was 3.6 kΩ.cm and for PCC was 109.6 kΩ.cm. The average electrical resistivity of the PCC is 30 times higher than AAC. The result suggests that the AAC composition affects the electrical resistivity, since the values indicates high electrical conductivity AAC and negligible electrical conductivity for PCC. This behavior may be explained by two factors. First by the high Fe₂O₃ contents found in the composition of the SS and FA (Ahmed; Kamal, 2022). According to the results of the XRF test (Table 1), these contents are 34.40 and 26.98%, respectively. Secondly, by the physical structure of the pores that aligns with the chemical characteristics of the solution, producing an easy passage of the current (Hu et al., 2019).

In concrete, the electrical resistivity is expected to relate to permeability (Gjorv, 2009). So, associated to the low permeability of the AAC, the electrical resistivity test presents a contradiction. According to Andrade (1993), a composite with lower electrical resistivity is associated with a pore system with greater interconnectivity, which means a higher permeability. This fact differs from the result of the test of water penetration under pressure. Studies also show that in concretes with a smaller connection between the pores (low permeability), an increase in resistivity values is observed (RIBEIRO, 2010). Such opposition can be further reinforced by the comparison between the different concretes (Table 6), in which the PCC presented electrical resistivity results far higher than the AAC, despite both having low permeability values. Besides that, the AAC presented higher compressive strenght results, a lower liquid/binder relation and more powders materials content than PCC which should also lead to a higher electrical resistivity (Medeiros Junior; Munhoz; Medeiros, 2019). The high iron oxide content of the AAC may be interfering on these electrical based results (Ahmed; Kamal, 2022).

The literature suggests that the electrical resistivity parameter as a factor of great importance to indicate the corrosion velocity of steel reinforcement of concrete structures (Balestra et al., 2020; Farias; Tezuka, 1992). Associating the data obtained from the measurements with the limits suggested by CEB (1989), presented in Table 3, AAC revealed very low values of electrical resistivity, and it is classified in the very high probability range of corrosion in concrete reinforcement, while the PCC is classified in the negligible range as considered classically. As it will be seen, this creates a contradiction when looking into direct measurements of chloride contents in the solutions from accelerated chloride migration tests.

Accelerated chloride migration

Electric current over time curves for the AAC are presented in Figure 8. High values of electric current were obtained. Between six hours readings, these values exceeded the maximum current supported by the equipment. It automatically caused the electric potential difference inserted into the system to be reduced, so that during this excess current generated, the total system current (summing up currents in all 4 branches of the circuit) remained at 2.1 A. This agrees with the electrical resistivity test, AAC is a very conductive composite. Azarsa and Gupta (2017) prove that concrete resistivity is inversely related to chloride ingress (at least as indirectly measured by the passing charge), where lower resistivity would be related to faster chloride migration.

From Figure 8, followed by the calculation procedure suggested by C1202 (ASTM, 2019), Table 5 was obtained, with corrected average charge passed values. According to Andrade et al. (1994), the test assumes a direct relationship between the number of Coulombs passed through the tested concrete and the chloride flow, which means that a large amount of load represents a large penetrability of chlorides into the concrete. A high average charge passed electric current was observed after data correction, which characterizes the concrete, according to the limits established in the test standard (Q > 4000 C), as a concrete with high chloride ion penetrability.

The same test and calculation procedure were also performed for the PCC, which obtained an average charge passed value of 194.94 C. For the AAC, the value was 11133.92 C (Table 5), a value 57 times higher than the PCC. These results show a large difference between the concretes concerning chloride ions penetrability according to the chloride migration test. Meira and Ferreira (2019) suggest the chloride migration test to verify the concrete regarding the resistance to the development of reinforcement corrosion by chloride attack. Therefore, available standard and literature data suggest that the alkali-activated concrete based on mineral FA and SS, by presenting high value of average charge passed, is characterized as a concrete with high penetrability of chloride ions and consequently low resistance to the development of reinforcement corrosion by chloride attack. As it will be seen, this creates a contradiction when looking into direct measurements of chloride contents in the solutions from accelerated chloride migration tests. The PCC presented low value average charge passed and it is characterized as a concrete with very low penetrability of chloride ions.

The difference obtained between the concretes agrees with the results of the electrical resistivity test, but creates a counterpoint about the test, since the PCC presented a much lower average than the AAC, despite both have low values of permeability. Moreover, the study by Patil and Allouche (2013) suggests that the activation of fly ash, a precursor used in AAC, may play a vital role in the ability of the alkali-activated matrix to resist the entrance of chlorides. Besides that, as discussed before, the AAC demonstrated superior compressive strength, a lower liquid/binder ratio, and a higher content of powdered materials compared to PCC, which should also contribute to lower values of chloride migration (Medeiros Junior; Munhoz; Medeiros, 2019).

Amorim Júnior (2020) claims, however, that the criteria used so far, which indirectly assess the durability of concretes, including the correlation between electrical resistivity and chloride ion penetration, are used for performance analysis of concretes composed of Portland cement. Therefore, they reflect the behavior of this material more than that of alkali-activated materials. This motivated the investigation of the chloride ion contents obtained in the chloride migration tests, to verify if the charge passed is in fact associated with the passage of chloride ions through the material investigated.

Ion chromatography

The chloride concentration data is showed in Figure 9. As expected, at the beginning of the test there was a zero concentration of chloride at the NaOH solution and a very high concentration at the NaCl solution. It was also observed that the concentration of chlorides in the final NaCl sample decreased during the test, since the purpose of the chloride migration test is precisely to force the passage of chloride ions in the specimen. The final values of chloride concentration between AAC and PCC, which have low permeability, are close, with a difference of just 10.67%. It is noteworthy that the disparity in chloride concentration between the start and the final concentration of the chambers containing sodium chloride was more pronounced for AAC (2592 mg/L) than for PCC (1443 mg/L), in comparison, AAC would have 1.80 times more chloride penetrability than PCC. The chloride migration test revealed a discrepancy, with 57 times more charge passed for AAC compared to PCC. Based on the chromatography results, AAC is less resistant to chlorides, however, it is not to the same extent indicated by the calculation method of C1202 (ASTM, 2019).

The results of this test contrast with the chloride migration and electrical resistivity tests, since the chloride concentration values found in the ion chromatography test showed much smaller percentage difference between the two concretes analyzed. In other words, the direct measurement of chloride concentration contrasts the electrical resistivity and electric charge passed results in chloride migration tests. Bernal, Gutiérrez and Provis (2012) indicated that a more precise explanation for this high conductivity phenomenon is related to the diffusion of Na⁺ ions through the samples as the alkalinity in the solution of pores increases, associated with higher values of the SiO₂/Al₂O₃ (S/A) ratio, which may cause interference in the results according to C1202 (ASTM, 2019) method.

Azarsa and Gupta (2017) and Lu (1997) also affirm that the direct relationship between electrical resistivity and ionic diffusivity occurs only for more porous materials, which may not be quite the case of the AAC studied. Amorim Júnior (2020) also reported a reduction in chloride percolation flow in alkali-activated concrete when compared to Portland cement concrete. Evidence of this is the fact that, at the end of the test, the chloride concentration in the anodic cell was reduced by more than 60% when comparing the values obtained for the reference concrete with the alkali-activated concrete.

The low electrical resistivity values and high values of the electric charge passed contrast with the low level of water penetration under pressure and the low concentration of chlorides passed during the migration test. This suggests an incompatibility between the tests for the classification of alkali-activated concrete based on FA and SS in terms of the criteria for durability and chloride ion passage classically used for Portland cement concretes.

Conclusions

This paper investigated different tests classically used for ordinary Portland cement concrete (PCC) and compared results for an alkali-activated concrete (AAC) based on fly ash (FA) and blast oxygen furnace (BOF) steel slag (SS). After analyzing the results, it was possible to conclude that:

1. AAC was classified as a low permeable concrete presenting low void content, low absorption, and low water penetration. However, the electrical resistivity and chloride migration by ASTM C1202 tests, indicated a high conductivity material and a high ion chloride penetrability, which strongly discourage the utilization of AAC based on BOF steel slag in scenarios involving steel reinforcement when exposed to external factors;

2. direct measurement (ion chromatography) of AAC resulted in a similar chloride migration when compared to the PCC. The found values were lower and do not agree with the results showed by the chloride migration test, even because AAC presented a greater amount of retain chloride ion in the specimen than PCC;

3. the electrical resistivity and the chloride migration tests separately do not result in reliable durability indicators for AAC based on BOF SS. This fact is supported by the observations that the water penetration under pressure and ion chromatography tests (direct method for determining chloride concentration) confirm low values of penetration, unlike the electrical resistivity and chloride migration tests (indirect methods for determining permeability to aggressive agents and chlorides);

4. due to the high conductivity of the AAC with BOF SS, tests that evaluate durability parameters from electrical charges cannot be reliable without further corrections; and

5. the combination of the tests of water penetration under pressure, followed by chloride migration and a subsequent test to measure the concentration of chloride ions as the ion chromatography test, is proposed to obtain consistent results for characterization of some durability properties of the AAC based on the materials studied.

For future research, it is important to explore additional properties of AAB (alkali-activated concrete) based on BOF (blast oxygen furnace) SS that could influence durability assessments, such as amounts of Fe₂O₃, and the physical structure of the pores in line with the chemical characteristics of the solution on it. Furthermore, different formulations of BOF SS-based concrete should be investigated to validate the results from the ion chromatography test. Additionally, novel chloride identification methods could be analyzed. With a well-established method, a new classification can be created to better characterize these materials regarding chloride migration. Such efforts will contribute to a more comprehensive classification of these emerging materials, potentially influencing their utilization in civil construction.

References

Edited by

Publication Dates

History