understanding the durability of alkali- activated and ... · geopolymer materials, this paper aims...

Concrete in Australia | Vol 45 No 4 33PB Concrete in Australia | Vol 45 No 4

TECHNICAL PAPERS

1 INTRODUCTIONAlkali-activated and geopolymer binders have great development potential and are widely considered as a promising alternative to Portland cement (PC) binders from both the scientific and commercial point of view. While these materials have shown great potential for various construction applications, their commercial adoption has been relatively slow.

The main factors hindering the mass production of alkali-activated and geopolymer binders can be identified as:1. Properties of the final product are highly sensitive to the

properties of precursors. For instance, in the case of fly ash-based geopolymers, the properties of fly ash (FA)as an industrial by-product can vary from one power station to another and even if the fly ash is sourced from the same power station, it can vary with time.

2. Compared to Portland cement-based materials where the main variable affecting the properties of the final product is the water-to-cement ratio, the design of alkali-activated and geopolymer binders involves understanding a dynamic and interdependent set of variables. This is a major obstacle for general practitioners and small concrete suppliers.

3. The long-term performance and durability of these binders require further investigation as they are comparatively new construction materials compared to Portland cement-based binders.

To address these issues, a systematic investigation of various parameters involved in the durability of alkali-activated and geopolymer systems in chloride-contaminated or carbon dioxide-concentrated environments was carried out at the University of New South Wales (UNSW). While readers are encouraged to refer to our recent published articles [1–6] to obtain an in-depth scientific insight into the chemistry, pore structure and various phenomena involved in the durability of alkali-activated and geopolymer materials, this paper aims to deliver a brief summary of our past research to develop a better understanding of the durability of alkali-activated and geopolymer materials among both the research community and the broader community of construction industry, engineers and practitioners.

1.1. WHAT ARE ALKALI-ACTIVATED AND GEOPOLYMER MATERIALS?Alkali-activated and geopolymer binders are produced by the reaction of solid precursors such as metakaolin, fly ash, and ground granulated blast-furnace slag (GGBS) with an alkali metal hydroxide or silicate solution; less carbon emission is attributed to their production. Depending on the precursor used, the final products can have considerably different nano/microstructures and properties. Alkali activation of aluminosilicate sources such as metakaolin and FA results in an amorphous aluminosilicate network, also called geopolymer, as first coined by Davidovits [7]. On the other hand, alkali activation of precursors with a considerable amount of calcium in the matrix (such as GGBS) leads to the production of some form of calcium silicate hydrates with partial substitution of aluminium, also known as C-(A)-S-H [8–11]. Despite the considerable difference in the properties of the final product fabricated from low-calcium aluminosilicate sources or calcium-rich precursors, this paper uses the generic term of ‘alkali-activated materials’ to refer to both these materials. This nomenclature mostly refers to their production process where the application of an alkaline solution is an essential part.

1.2. CORROSION OF REINFORCEMENT AND THE VARIOUS MECHANISMS INVOLVEDOne of the main causes of shortening the service life of concrete structures is corrosion of the reinforcement through either chloride contamination or carbonation. Accumulation of chloride ions on the concrete-reinforcement interface leads to a breaking down of the passive layer formed around the reinforcement and eventually corrosion. In carbonation-induced corrosion, carbonation of concrete provides an acidic environment, leading to depassivation of reinforcement and consequently corrosion. Although there are some well-established electrochemical test methods to evaluate the corrosion rate, which have been under development for almost 60 years, the inherent complexities associated with a multi-physics process in which different transfer mechanisms (moisture, chloride, carbon dioxide, oxygen, and charge transfer) are involved and interact with each other has been a major obstacle to predicting the service life of structures based on experimental data or field observation.

Alkali-activated and geopolymer materials as low-embodied-carbon alternatives to conventional Portland cement-based materials have been at the forefront of recent academic research. While these materials have shown great potential for various construction applications, their commercial adoption has been relatively slow, mainly due to concerns over their long-term performance and durability.This paper provides an overview of some of the main aspects of durability in alkali-activated and geopolymers materials. It also presents a summary of previously published research by the authors on various topics related to corrosion of reinforcement in these materials. The aim of this paper is to develop a better understanding of the durability of alkali-activated and geopolymer materials among both the research community and the broader community of construction industry, engineers and practitioners

UNDERSTANDING THE DURABILITY OF ALKALI-ACTIVATED AND GEOPOLYMER MATERIALS

MAHDI BABAEE1

ARNAUD CASTEL2

1 ASSOCIATE, TAYLOR THOMSON WHITTING (TTW), SYDNEY, NSW, AUSTRALIAEMAIL: [email protected], [email protected]

2 PROFESSOR, SCHOOL OF CIVIL AND ENVIRONMENTAL ENGINEERING, UNIVERSITY OF TECHNOLOGY SYDNEY (UTS), SYDNEY, NSW, AUSTRALIA

Concrete in Australia | Vol 45 No 4 3534 Concrete in Australia | Vol 45 No 4

TECHNICAL PAPERS

Corrosion is composed of two main stages: the initiation phase, which is the required time for aggressive agents (e.g. chloride ions) to reach the reinforcement and depassivate it, and the propagation phase, which is the active phase of corrosion [12] (Figure 1). As illustrated in Figure 1, during the initiation phase of chloride-induced corrosion, during the initiation phase the moisture transport properties of concrete define the rate of moisture adsorption and consequently the penetration of chloride ions through a convective process during wetting/drying cycles in non-saturated concrete. Once the concrete is saturated, however, chloride diffusivity is the main controlling factor. When enough chloride ions reach the interface between concrete and reinforcement, the passive layer around the bars starts to deteriorate and the corrosion begins. Not all the penetrated chloride ions contribute to the depassivation process, as a percentage of them will be chemically bound to the matrix. The amount of chloride that leads to breaking down the passive layer and the onset of corrosion is called the chloride threshold, and it is highly dependent on the chemical composition of the binder. After depassivation of the reinforcement, and during the propagation phase of corrosion, moisture adsorption and desorption again play a significant role, as the electrical resistivity of concrete depends on the moisture content. Electrochemical parameters, such as the corrosion potential and polarisation resistance, are other parameters defining the kinetic of corrosion during the propagation phase of corrosion. Finally, a lack of oxygen, which depends on the pore structure, can decrease the rate of corrosion as the corrosion process becomes concentration-controlled.

Although the transfer mechanisms and the parameters defining them have been the subject of much research in the past for PC concretes, the amount of literature for new and alternative binders such as alkali-activated and geopolymer binders is relatively small. Due to the very different chemistry of these binders compared with traditional PC binders, all governing parameters defining different mechanisms involved in the corrosion of reinforcement in alkali-activated and geopolymer binders were investigated in this study. Water-vapour-sorption-isotherms (WVSI) and moisture diffusion coefficients were studied and determined for both low and high calcium binders. Through these studies, pore structure, which plays a major role in all involved transfer mechanisms, including carbonation and chloride transfer, were investigated. Chloride diffusion coefficients and the binding capacity of alkali-activated mortars were also systematically investigated. Chloride threshold values were determined as index values to show the stability of the passive layer around the bars. Corrosion kinetic parameters were also assessed for both passive (before corrosion) and active states of corrosion. The passivity of the reinforcement in carbonated geopolymer concretes and the chemical adsorption of carbon dioxide into the pore solution of these systems were modelled and studied. Finally, the alkali cation leaching and its probable effects on the loss of alkalinity and stability of the passive layer, as well as the risk of efflorescense, were investigated.

2. INVESTIGATED MATERIALSTo ensure a systematic assessment of various influential factors on the pore structure development and durability, a wide range of mix proportions were investigated. Parameters such as the percentage of FA and GGBS, alkaline solution type, alkali concentration, the molar ratio of SiO2 to Na2O and curing regimes (heat and ambient-curing) were investigated. While a detailed discussion of the effect of these parameters on various durability properties of alkali-activated and geopolymer materials can be found on [1–6], the results for the two main categories are presented in this paper:

- Low-calcium alkali-activated (LCAA) binders: these binders were fabricated of more than 75% FA and at high alkalinity levels (Na2O% = 5-11%).- High-calcium alkali-activated (HCAA) binders: these binders were fabricated of more than 75% GGBS and at low alkalinity levels (Na2O% = 3-5%). Generally, the same results and properties are expected for systems with more than 50% GGBS.

A summary of the range of mixed variables for both categories is presented in Table 1.

3. WATER-VAPOUR-SORPTION-ISOTHERMS, PORE STRUCTURE, AND MOISTURE TRANSPORT CHARACTERISTICSWVSI establish the relationship between the moisture content and the environmental relative humidity at a fixed temperature [13]. They also provide valuable information regarding the structurally-bound water, the pore-blocking effect, the cavitation, and the volume of pores involved during the adsorption/desorption processes from dry/saturated conditions respectively. Correlation of the many transport phenomena (e.g. chloride transport during wetting/drying cycles) with moisture transport,

FIGURE 1. Corrosion phases and various transfer mechanisms involved in each phase

TABLE 1. Mix composition of investigated samples

Precursor SiO2/Na2O(molar) Na2O% (1)

FA/(FA+GGBS)

%

Water/binder

LCAA 1.0-1.5 5-11 ≥75% 0.32-0.35

HCAA 1.0-1.5 3-5 ≤25% 0.40(1) Na2O%=Na2O/ (FA+GGBS) (%)


TECHNICAL PAPERS

and the dependency of other parameters such as the bulk electrical resistivity and the oxygen diffusivity on the moisture content make water/water vapour the ideal probe to investigate the pore structure and permeability of porous binders [5,14,15].

Figure 2 depicts WVSIs for: (a) an LCAA paste sample, (b) an HCAA paste sample, and (c) a PC paste. Analysis of these WVSIs reveals several interesting aspects of the pore structure of alkali-activated and geopolymer materials. Traditionally for Portland cement-based binders, a layered structure at the scale of tens of nanometres is considered with the ability to hold ‘interlayer’ water molecules in the spaces within the layers [8,16–18]. The hysteresis behaviour at low relative humidities (RH<11%) in WVSIs can be regarded as a qualitative index representing the ability to retain one water molecule in the interlayer spaces. As can be seen in Figure 2, unlike the PC sample, both the LCAA and HCAA paste samples lack a notable low-relative-humidity hysteresis between the adsorption and desorption branches. While the LCAA sample does not have enough calcium to have significant C-S-H components, some form of C-S-H formation is expected for HCAA. A lack of retained interlayer water for the HCAA sample can be in part due to a different nanostructure of the C-(A)-S-H gel in the alkali-activated slag binders than the C-S-H gel formed in Portland cement-based binders, consistent with the considerably lower Ca/Si ratios of these matrixes. The GGBS-rich sample showed larger hysteresis between the adsorption and desorption branches of WVSIs compared to both the low-calcium LCAA and PC samples, suggesting higher pore blocking1 in the network due to the formation of finer pores.

According to Figure 3, a totally different pore structure for aluminosilicate-dominated and calcium-rich binders was revealed through pore size distribution calculations by MIP test. The pore structure of an aluminosilicate-dominated sample (LCAA) was comprised of an almost uniformly distributed volume of pores over the whole mesopore (2 nm< pore dimeter <50 nm) and macropore (pore dimeter> 50 nm) ranges. For the calcium-rich binder (HCAA), on the other hand, a considerably finer pore structure was observed with a relative lack of large macropores.

The time-dependent sorption kinetics was also studied for both samples. Results demonstrated a considerably lower permeability for the HCAA sample (up to 10 times less permeable than LCAA sample) and highlighted the importance of the presence of calcium to develop a fine pore structure with high tortuosity, which is critical to developing durable engineering binders.

4. CHLORIDE BINDING, CHLORIDE DIFFUSIVITY, AND CHLORIDE THRESHOLDTo have a comprehensive assessment of the corrosion initiation, chloride binding, chloride diffusivity, and chloride threshold should be investigated. As depicted in Figure 4, unlike in the PC control sample, a comparison of acid and water-soluble chloride contents showed no bound chloride in the aluminosilicate-dominated (LCAA) and calcium-rich (HCAA) mortar samples. This is characteristic of a lack of chemical binding in the alkali-activated and geopolymer materials studied here. As a result, any discussion regarding the chloride diffusivity in the alkali-activated and geopolymer binders should be focused on the role of the pore structure in physical chloride binding and

FIGURE 2. Water vapour sorption isotherms of aluminosilicate-dominated LCAA, calcium-rich HCAA, and a cement paste [5]

FIGURE 3. Pore volume distribution for LCAA and HCAA samples

1 Pore blocking occurs if a pore has restricted access to the external gas phase via narrow constrictions. Ink-bottle-shaped pores are a good example of pores with limited access due to narrow constrictions [27, 28].


TECHNICAL PAPERS

encapsulation; this highlights the importance of developing a fine pore structure.

In agreement with the content of the previous section, where the presence of calcium in the mix resulted in developing a very fine pore structure, calcium played a vital role in reducing the chloride diffusivities as well, as can be seen in Figure 5. The chloride diffusion coefficient of calcium-rich HCAA sample in Figure 5 is considerably smaller than in both the LCAA and PC samples, emphasising the critical role of calcium in the matrix to deliver durable alkali-activated binders in marine environments. According to Figure 5, the LCAA samples figures were substantially influenced by the modulus ratio (i.e. the molar ratio of SiO2 to Na2O) and showed considerably lower chloride diffusion coefficients at lower modulus ratios. Although the diffusion coefficient at lower modulus ratios for aluminosilicate- dominated samples is comparable to those of the Portland cement-based samples, the difficulties associated with the production of such

highly alkaline products that are corrosive, highly viscose and have a potentially high risk of efflorescence discourage the commercial use of such materials. Although results are not presented here, the low-calcium aluminosilicate-dominated samples showed systematically higher diffusion coefficients compared to the Portland cement-based and calcium-rich samples [1].

To assess the chloride threshold, a wide range of depassivated reinforced samples was investigated as the first systematic attempt at developing a chloride threshold database for alkali-activated and geopolymer materials. Chloride thresholds were obtained by chloride content measurement of the powder sampled from the concrete-reinforcement interface, after observing signs of corrosion initiation (i.e. after observing a considerable drop in the half-cell potential and polarisation resistance values). Results ranged between 0.2-0.3 (wt.% binder mass) for calcium-rich binders fabricated at low alkalinity levels and 0.5-0.7 for aluminosilicate-dominated samples manufactured with highly alkaline solutions. It seems that the suggested threshold values of 0.2% or 0.4% (wt. % by binder mass) commonly used for PC binders are suitable for low-calcium aluminosilicate-dominated materials that require high alkalinity levels to provide acceptable mechanical strengths; however, they are not conservative enough for calcium-rich binders fabricated at low (to medium) levels of alkalinities.

5. THE PASSIVITY OF EMBEDDED REINFORCEMENT IN CARBONATED LOW-CALCIUM GEOPOLYMER CONCRETEAs one of the main threats to the durability of reinforced concrete, carbonation plays a detrimental role in concrete by reducing the alkalinity of the pore solution in the vicinity of the reinforcement. This may lead to the depassivation of reinforcement, leaving it prone to corrosion. For PC concrete, the hydration product portlandite (Ca(OH)2) provides a buffer effect in which the continuous dissolution of portlandite maintains a high pH level of pore solution in case of neutralisation of the OH- ions during the carbonation process. Unlike the PC binders, low-calcium aluminosilicate-dominated binders do not contain

FIGURE 4. Comparison of total and free chloride profiles (wt. % by sample mass) in (a) LCAA, (b) HCAA, and (c) PC mortars after 45

days of immersion in 16.5% NaCl solution [1]

FIGURE 5. Chloride diffusion coefficients of PC, LCAA and HCAA mortar samples [1]

TABLE 2. Chloride threshold values for alkali-activated mortar samples

Precursor Cl (wt.% by binder mass1)

LCAA 0.50-0.70

HCAA 0.20-0.30

(1) Binder = FA + GGBS + anhydrous activator


TECHNICAL PAPERS

a considerable amount of portlandite as a reaction product; as a result, they might be more prone to the loss of alkalinity [19]. During the carbonation of fly ash-based geopolymers, the main process is carbonation of the highly alkaline pore solution in which the alkali carbonate/bicarbonate salts precipitate from the pore solution. No significant change in the nanostructure of aluminosilicate matrix has been reported [20]. On the other hand, calcium-rich alkali-activated binders experience carbonation of the pore solution, followed by carbonation of secondary products and decalcification of calcium aluminosilicate hydrate (C-(A)-S-H) gel, which can lead to structural strength degradation and an increase in porosity, particularly when a water glass is used for activation [21,22]. Therefore, the pore solution alkalinity and composition play a major role during the carbonation process of calcium- rich alkali-activated and low-calcium geopolymer binders: a pore solution with enough alkalinity can keep the reinforcement passivated; it can also prevent the structural alteration of the binder (in particular Ca-rich binders), which occurs only after the initial exhaustion of the pore alkali content during carbonation.

To assess the carbonation of low-calcium geopolymer concretes, an analytical chemisorption model was developed to predict the pH drop and carbonate phase distributions based on the equilibrium of different coexisting species in an aqueous sodium hydroxide solution (as representative of the pore solution of low calcium geopolymer materials). Various alkali contents and partial pressure of the ambient carbon dioxide were investigated. As illustrated in Figure 6(a), accelerated carbonation tests in laboratory conditions using high carbon dioxide concentration (higher than the ambient pCO2=0.03%) led to an artificial pH decline compared with natural carbonation, as it promoted the formation of bicarbonates over carbonates. Nevertheless, based on long-term monitoring of corrosion properties such as polarisation resistance, the artificially decreased pH level (accelerated carbonation in this study was pCO2=1%) did not lead to the depassivation of reinforcement as the pH remained at rather high levels after carbonation (Figure 6(a) and (c)).

Of interest also is the application of phenolphthalein (as the traditional indicator for assessing the carbonation front) to carbonated low-calcium geopolymers. In Figure 6(a), a colour change chart of phenolphthalein indicators over the considered range of pH values is schematically presented. Figure 6(b) also shows a carbonated sample after 10 weeks of carbonation in the carbonation chamber at pCO2=1%. In agreement with the calculated pH graph of Figure 6(a), the pH of the carbonated zone is close to 9.5, resulting in a faintly discernible pink colour. This is of great importance, as the faint pink colour is not actually representative of a completely carbonated concrete which can lead to corrosion and is not illustrative of the carbonation front.

Furthermore, based on the developed model, even low levels of alkali concentrations in the pore solution ([Na]~0.2 mol/l) appear to provide enough protection for the reinforcement during natural carbonation. On the other hand, a very low alkali concentration in the pore solution ([Na]<0.2 mol/l) will increase the risk of depassivation considerably. These findings generate more confidence in the industrial application of low-calcium geopolymer concretes where carbonation-induced steel corrosion is a concern, as these binders were shown to be capable of providing the

required passivity for the reinforcement. More research is however required for high-calcium systems.

6. ELECTROCHEMICAL ASPECTS OF CORROSION OF REINFORCEMENTDepassivation of reinforcement is traditionally considered as the end of the service life of concrete structures. To detect depassivation and active corrosion, corrosion parameters such as corrosion potential (Ecorr) and polarisation resistance (Rp) are measured on-site or in the laboratory. Measured values then need to be compared against some index figures to be able to identify a passive sample from a corroding one. The index value traditionally used for Portland-cement based systems are summarised in Table 3.

FIGURE 6. (a) pH of the aqueous alkali solution at different alkali concentrations and CO2 levels, (b) application of phenolphthalein

on geopolymer concrete, and (c) polarisation resistance value showing the same sample is still in the passive condition [3]


TECHNICAL PAPERS

Figure 7 illustrates the corrosion parameters for passive PC-based samples along with LCAA and HCAA samples. Samples have been submerged in water only with no exposure to chloride contamination. On each graph, the index values of corrosion parameters from Table 3 are shown in dashed lines. As can be seen, the measured corrosion parameter values for PC samples lie conveniently within the passive zone identified by the index values (i.e. Ecorr >-200 and Rp>250). For geopolymer and alkali-activated samples, however, the traditional index values can be misleading and will require re-calibration. For LCAA samples, Ecorr values are generally between -150 and -400. The addition of calcium shifted the Ecorr values toward a more negative range, up to -500 and even more. For calcium-rich samples, the more negative potential values can be attributed to consumption of dissolved oxygen by sulphides, which are sourced by slag particles; this process impedes the formation of the passivating iron oxide layer during the early stages of matrix development [1]. Based on conventional index values, such negative potential numbers would falsely be a sign of corrosion. Furthermore, the scattering of data for LCAA and HCAA compared to the PC samples could be due to variations in the alkaline solution type and concentration, which would add another layer of difficulty to interpreting the potential values. As shown, the polarisation values were less susceptible to scattering and remained generally above 100 kΩ.cm2. These results suggest that the corrosion potential is not a reliable corrosion detection parameter for alkali-activated and geopolymer systems, and, depending on the calcium content, a modified Rp value of around 100 kΩ.cm2 would suit these systems better as an index value for the passive condition. Further research is required to develop a larger data base for statistical review of this recommended index value.

After complete depassivation of reinforcement, due to chloride contamination or carbonation, the rate of corrosion is governed by the electrical resistivity of concrete or mortar. The electrical resistivity of alkali-activated and geopolymer materials studied by the authors was found to be significantly influenced by the alkaline solution concentration, i.e. the concentration of alkali ions in the matrix. The second influential parameter, although of comparatively subordinate importance, was the total porosity and tortuosity of the binder. Thus, high-calcium binders provide a better corrosion performance, i.e. a lower corrosion rate, after depassivation of reinforcement. The high-calcium binders are generally fabricated at low alkali concentration levels as strength development of the matrix depends to a large extent on the hydration of calcium components, which does not need very high pH compared to low-calcium fly ash-based systems. High-calcium binders also have a noticeably finer pore structure, as explained in Section 3.

7. ALKALI CATION LEACHING AND EFFLORESCENCEConsumption of concentrated alkaline solution to achieve high mechanical strengths can lead to the availability of free and mobile alkali ions in the pore solution of alkali-activated and geopolymer materials that cannot be incorporated into the matrix. Moreover, various degrees of reactivity of the solid precursors, in particular, fly ash, may lead to overestimation of the absorbable alkalis and partial incorporation of the designated alkali content. As a result, a fraction of alkalis will be not structurally bound and can diffuse out of the matrix. Alkali leaching may lead to serious problems, such as efflorescence or loss of alkalinity of the binder.

To assess the effect of various mix composition parameters and of the curing condition on the efflorescence and alkali leaching potential, a wide range of samples were tested. The results showed that although the alkali content was the most important parameter

FIGURE 7. Corrosion potential (Ecorr) and polarisation resistance (Rp) for (a) Portland-cement based samples, (b) LCAA samples, and

(c) HCAA samples

TABLE 3. Index values of corrosion parameters for PC-based systems

Ecorr (mV) Rp (kΩ.cm2)

Passive condition(>90 % probability of having no corrosion)

Ecorr >-200 Rp>250

Very low-to-moderate corrosion rate -200< Ecorr <-300 25<Rp<250

High corrosion rate(>90 % probability of active corrosion)

Ecorr <-300 Rp<25


TECHNICAL PAPERS

influencing the ion leaching and the efflorescence, the amount of leached-out alkali ions and consequently the efflorescence severity was not proportional to the alkali content introduced into the mixes. The presence of calcium ions, the degree of matrix development, which depends on the initial alkalinity of the activator solution, and the curing regime (heat or ambient) can affect the amount of leaching ions and efflorescence. In general, higher calcium and alkali content, lower silicate content, and ambient curing increased the ion leaching and risk of efflorescence [6].

Considering all the interacting variables mentioned above, establishing a relationship between the mix variables and the risk of efflorescence seems unfeasible. Thus, a practical performance-based approach was developed to assess the risk of efflorescence based on the amount of leached out ions that reacted with the ambient carbon dioxide to produce white efflorescence products. Paste cylinders (100(H) x 50(D) mm) were fabricated and were kept in contact with the water at the bottom 28 days after casting. The water depth was maintained at a constant 1 to 3 mm. During the test period, samples were stored in a controlled temperature of 23 ± 2 °C and 50% relative humidity for 90 days. At day 90, any white efflorescence products formed on the side and top surfaces of the samples were scraped into a container and weighed using a milligram precision scale. The density of the efflorescence product was determined by dividing the mass of the scraped efflorescence product in milligrams by the total volume of the sample in cubic centimetres (mg/cm3). By visually comparing the paste samples with their concrete counterparts, a performance-based guideline was established, as shown in Table 4.

It should be noted that the test method developed was an accelerated approach, and the corresponding concrete sample would experience significantly less efflorescence, as shown in Figure 8. This is due mainly to the different paste content and permeability of concrete samples compared to their paste counterparts.

Furthermore, the loss of alkalinity did not lead to depassivation of reinforcement, which was studied through long-term monitoring of the electrochemical parameters of reinforced mortar samples immersed in deionised water. This finding is also consistent with the results of the chemisorption model in Figure 6(a), where a significantly low level of alkali concentration in the pore solution was enough to keep the pH at rather high levels and therefore enough to protect the reinforcement.

8. CONCLUSIONThis paper contributes to developing a better understanding of the durability of alkali-activated and geopolymer materials with a focus on the corrosion of the reinforcement. One of the main outcomes of this work is that it shows the strong dependency of the investigated phenomena/parameters on the mix composition and chemistry of the binders. All the investigated materials are generally called alkali-activated binders or geopolymers; however, they are a rather diverse family of materials that can present vastly different properties depending on the composition of their mix. This variability introduces some difficulties for general practitioners and for the commercial application of these materials, since developing durable materials requires a proper understanding of the nano/microstructure and its dependency on various factors. However, this variability also presents interesting opportunities to produce alternative binders that can be tailored to particular engineering applications.

Developing a low permeable and highly durable binder that is capable of protecting a reinforcement against corrosion is certainly achievable. The mix proportions and the type of alkaline solution will depend, however, on the application and the environmental conditions to which the concrete/mortar will be exposed. Generally, the addition of calcium into the matrix (in the correct quantities to meet the specific need) will reduce the permeability and require less alkalinity for strength development compared to low-calcium systems, thus providing a durable and practical alternative for Portland cement-based systems in marine environments.

TABLE 4. Performance-based criteria to assess the risk of efflorescence

Density of scrapped efflorescence product (mg/cm3)

Risk level Example

<1 No risk

1.0< <2.5 Low-to-medium risk

2.5< <10 Medium-to-high risk

>10 High risk

FIGURE 8. An example of a sample with a medium-to-high risk of efflorescence: concrete sample on the left and paste sample on

the right

Concrete in Australia | Vol 45 No 4 PB40 Concrete in Australia | Vol 45 No 4

TECHNICAL PAPERS

REFERENCES[1] M. Babaee, A. Castel, Chloride diffusivity, chloride threshold,

and corrosion initiation in reinforced alkali-activated mortars: Role of calcium, alkali, and silicate content, Cem. Concr. Res. 111 (2018) 56–71.

[2] A. Noushini, M. Babaee, A. Castel, Suitability of heat-cured low-calcium fly ash-based geopolymer concrete for precast applications, Mag. Concr. Res. 68 (2015) 1–15.

[3] M. Babaee, M.S.H. Khan, A. Castel, Passivity of embedded reinforcement in carbonated low-calcium fly ash-based geopolymer concrete, Cem. Concr. Compos. 85 (2018) 32–43.

[4] M. Babaee, A. Castel, Chloride-induced corrosion of reinforcement in low-calcium fly ash-based geopolymer concrete, Cem. Concr. Res. 88 (2016) 96–107.

[5] M. Babaee, A. Castel, Water vapor sorption isotherms, pore structure, and moisture transport characteristics of alkali-activated and Portland cement-based binders, Cem. Concr. Res. 113 (2018) 99–120.

[6] M. Babaee, A. Castel, The Effect of alkali concentration and the binder composition on the alkali leaching in geopolymer binders, in: 28th Bienn. Natl. Conf. Concr. Inst. Aust., 2017.

[7] J. Davidovits, Geopolymers and geopolymeric materials, J. Therm. Anal. 35 (1989) 429–441.

[8] R.J. Myers, S.A. Bernal, R. San Nicolas, J.L. Provis, Generalized Structural Description of Calcium–Sodium Aluminosilicate Hydrate Gels: The Cross-Linked Substituted Tobermorite Model, Langmuir. 29 (2013) 5294–5306.

[9] 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.

[10] S.-D. Wang, K.L. Scrivener, Hydration products of alkali activated slag cement, Cem. Concr. Res. 25 (1995) 561–571.

[11] C.K. Yip, J.S.J. van Deventer, Microanalysis of calcium silicate hydrate gel formed within a geopolymeric binder, J. Mater. Sci. 38 (2003) 3851–3860.

[12] K. Tuutti, Corrosion of steel in concrete, CBI Forsk. 824. (1982).

[13] J.M. de Burgh, S.J. Foster, Influence of temperature on water vapour sorption isotherms and kinetics of hardened cement paste and concrete, Cem. Concr. Res. 92 (2017) 37–55.

[14] J.M. De Burgh, I.M.A. Al-Damad, S.J. Foster, Experimental investigation of water vapour sorption isotherms and microstructure of low calcium geopolymer binders, (2017) 978–1.

[15] V. Baroghel-Bouny, Water vapour sorption experiments on hardened cementitious materials.

Part II: Essential tool for assessment of transport properties and for durability prediction, Cem. Concr. Res. 37 (2007) 438–454.

[16] J.M. de Burgh, S.J. Foster, H.R. Valipour, Prediction of water vapour sorption isotherms and microstructure of hardened Portland cement pastes, Cem. Concr. Res. 81 (2016) 134–150.

[17] S. Grangeon, F. Claret, Y. Linard, C. Chiaberge, X-ray diffraction: a powerful tool to probe and understand the structure of nanocrystalline calcium silicate hydrates., Acta Crystallogr. B. Struct. Sci. Cryst. Eng. Mater. 69 (2013) 465–73.

[18] P. Pourbeik, J.J. Beaudoin, R. Alizadeh, L. Raki, Dimensional stability of 1.4 nm tobermorite, jennite and other layered calcium silicate hydrates, Adv. Cem. Res. 27 (2015) 2–10.

[19] R.R. Lloyd, J.L. Provis, J.S.J. van Deventer, Pore solution composition and alkali diffusion in inorganic polymer cement, Cem. Concr. Res. 40 (2010) 1386–1392.

[20] S.A. Bernal, J.L. Provis, B. Walkley, R. San Nicolas, J.D. Gehman, D.G. Brice, A.R. Kilcullen, P. Duxson, J.S.J. van Deventer, Gel nanostructure in alkali-activated binders based on slag and fly ash, and effects of accelerated carbonation, Cem. Concr. Res. 53 (2013) 127–144.

[21] F. Puertas, M. Palacios, T. Vázquez, Carbonation process of alkali-activated slag mortars, J. Mater. Sci. 41 (2006) 3071–3082.

[22] S.A. Bernal, J.L. Provis, R. Mejía de Gutiérrez, J.S.J. van Deventer, Accelerated carbonation testing of alkali-activated slag/metakaolin blended concretes: effect of exposure conditions, Mater. Struct. 48 (2015) 653–669.

[23] M. Babaee, M.S.H.S.H. Khan, A. Castel, Passivity of embedded reinforcement in carbonated low-calcium fly ash-based geopolymer concrete, Cem. Concr. Compos. 85 (2018) 32–43.

[24] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol, T. Siemieniewska, IUPAC (Recommendations 1984), Reporting physisorption data for gas/solid systems with special reference to the determination of surface area and porosity, Pure Appl. Chem. 57 (1985) 603–619.

[25] M. Thommes, Physical Adsorption Characterization of Nanoporous Materials, Chemie Ing. Tech. 82 (2010) 1059–1073.

understanding the durability of alkali- activated and ... · geopolymer materials, this paper aims...

Documents