texto em 
Inglês (pdf)
Artigo em XML
Referências do artigo
Tradução automática
Enviar este artigo por email
Citado por SciELO 
Similares em
SciELO 
Permalink
1. Introduction
The alkali-aggregate reaction (AAR) is one of the most harmful deterioration mechanisms that affects the durability and service life of concrete structures in the world. AAR can be divided into two main types of reaction: alkali-silica reaction (ASR) and alkali-carbonate reaction (ACR) (Fournier and Bérubé, 2000). ASR, most commonly found in concrete structures, is characterized by the chemical reaction between the unstable silica minerals present in the aggregates and the alkaline hydroxides (Na+, K+, OH-) arising primarily from the hydration of Portland cement while dissolved in the concrete pore solution. On the other hand, ACR is a much less common type of reaction, whose mechanism is still largely unknown. ACR is seen by many researchers as a reaction that occurs between alkaline hydroxides and certain types of dolomitic limestones (Fournier and Bérubé, 2000; Beyene et al., 2013). In general, it is ordinarily accepted that ACR is accompanied by a process of dedolomitization and calcite and brucite formation (Thomas and Folliard, 2007), although different studies consider ACR as a form of ASR (Katayama et al., 2012, 2016). However, the microscopic characteristics of expansion and deterioration induced by ACR were separated from the ASR because they are quite different (Sanchez et al., 2017), which justifies the distinction of terms. Independently, both deterioration mechanisms (ASR and ACR) are present in concrete structures all over the world and are responsible for the generation of secondary products that cause expansions in the presence of moisture. With the expansion, which occurs as a consequence of AAR, the durability and mechanical properties of concrete are impaired, with the appearance of microcracks in the affected structure (Fournier and Bérubé, 2000; Bérubé et al., 2005; Fournier et al., 2010; Naar, 2010; Kubo and Nakata, 2012; Zhu et al., 2021; Fiset et al., 2021).
Studies related to this type of expansive reaction seek to assess the degree of reactivity of the aggregates present in different regions (Lu et al., 2006a; Lu et al., 2006b, Drolet et al., 2017a and Drolet et al., 2017b, Santos and Ribeiro, 2021). Others seek to identify mineral additions that can mitigate the effects of this reaction (Duchesne and Berubè, 2001 and Grattan-Bellew et al., 2003; Gallo et al., 2007; Adorno et al., 2018; Adorno et al., 2020 ) and evaluative methodologies in order to correlate techniques of accelerated tests and field tests (Ideker et al., 2010; Sanchez et al., 2014; Sanchez et al., 2015; Sanchez et al., 2016; Lindgard et al., 2016; Sanchez et al., 2017; Martin et al., 2017). In addition, there are studies aimed at correlating the alkalinity of the cementing matrix with the development of AAR (Shehata et al., 2000; Shehata et al., 2002; Shehata et al., 2006; Shehata et al., 2010).
Evaluating and studying AAR is of paramount importance in order to characterize and identify factors that influence the useful life of concrete structures. Researchers in several countries have sought to understand the effects of AAR since it was first identified. Among the countries that most publish articles in journals of high scientific impact on AAR is Canada. Brazilian researchers have an important history of partnerships and research interactions on this topic with researchers working in Canada. In this context, the present research consists of carrying out a bibliographical survey in order to analyze the main themes of studies related to AAR developed by Canadian and Brazilian authors, as well as the evaluative methodologies and rising themes. The importance of developing this work is to produce an overview in order to better understand trends related to alkali-aggregate reaction studies in the scientific environment of the two countries involved. In addition, the work presents an important discussion about the use of the accelerated method of mortar bars for AAR research, in comparison with longer-term methods such as the concrete prism test, which is considered more realistic, and with methods of damage assessment such as the Damage Rating Index (DRI) and the Stiffness Damage Test (SDT).
2. Methodology
The research consists of carrying out a qualitative, integrative and systematic review in order to compare the studies and evaluation methods of the alkali-aggregate reaction in both countries, Canada and Brazil.
First, the search term "Alkali-aggregate reaction" was defined and later the keyword was searched in the Web of Science (WoS) and Scopus databases. Both present an excellent temporal coverage, with records from the years 1945 and 1970, respectively, until the present. In addition, these databases include a large number of journals, events and publishers with a peer review system, which contributes to the variety and quality of publications. The search period considered was the entire available temporal space of the databases, considering all registered research areas. Data collection was carried out on January 13, 2022.
Related analyses were carried out regarding the number of publications in each country, main authors on the subject and number of citations. Finally, we sought to detail the publications and validate the evidence based on co-citation analyses and bibliographic coupling, which allows identifying the relationship between authors and references, as developed by Mariano and Rocha (2017). The VOSviewer 1.6.10 software was used to form the networks, which reads the data and uses clustering algorithms, separating them into groups according to the study areas.
2.1 Number of Publications - Brazil and Canada
Figure 1 identifies the number of publications found in the Web of Science (WoS) and Scopus databases in high-impact journals using the term "Alkali-aggregate reaction" and the contribution of each country on the topic. Figure 1 shows the publications of the 10 most prominent countries in the world's scientific production on the topic of aggregate alkali reaction.
Observing Figure 1, adding the number of publications from the two databases, and excluding repeated works, it is possible to rank the countries. Thus, it is noted that the country that most contributed to research related to AAR was Canada with 215 publications, followed by the United States and China with 170 and 167 publications, respectively. Brazil ranks seventh, contributing with 50 publications.
A total of 727 records on AAR were found in the WoS database, with 113 of these publications being Canadian and 33 being Brazilian. In the Scopus database, 1096 records were obtained, where Canada contributed with 173 publications and Brazil with 46. In general, based on the information obtained, it appears that Canada is the country that most contributes to studies related to AAR, representing about 22% of the total, while Brazil represents 5.7%. Thus, the number of studies published on "Alkali-aggregate reaction" in Canada is almost four times greater than in Brazil.
2.2 Co-citation analysis - Brazil and Canada
Co-citation analysis is based on how often two authors or papers are cited together in subsequent articles. This indicates a proximity between the research themes addressed and the researchers. Thus, it is possible to distinguish a domain of knowledge through iterative analyses (Hjorland, 2013).
The interconnection of citations characterizes the research according to its social, historical and dynamic nature and helps to identify the dependence on academic literature, which allows the determination of the main research groups related to the study topics and their publications with the greatest impact (Glänzel, 2003).
Figures 2 show the links between the main research groups and authors related to AAR in Canada and Brazil, extracted from the WoS and Scopus databases, respectively.
Figure 2. Link between the main research groups focused on investigating the alkali-aggregate reaction, with data obtained from WoS (a) and Scopus (b).
Analyzing Figure 2, one can observe three major research themes led by Fournier, Thomas and Berubè with 29, 21 and 14 publications by each author, respectively, for the WoS database and Scopus, excluding repeated works.
The three major groups of researched themes are led by Canadian authors, and in these lines of research, Brazilian authors such as Sanchez and Monteiro share similar themes, relating mainly to AAR assessment techniques. The first author is a professor at the University of Ottawa, Canada, and has interaction with the research of Professor Fournier's group, located at the University of Laval, Quebec City, also in Canada. The second author is currently a professor at the University of California in Berkeley, USA. He is also part of a research group led by Hasparyk at Furnas Centrais Elétricas, Department of Concrete & Quality Control in Aparecida de Goiânia, GO, Brazil and with partnerships with Professor Dal Molin from the Federal University of Rio Grande do Sul in Porto Alegre, RS, Brazil. The approaches researched by Monteiro significantly interact with the subjects studied by Thomas.
The following chapter seeks to go deeper into the subject of the works developed by the authors cited above, based on a selection of the articles with the highest number of citations.
3. Results and discussions
3.1 Themes of greatest relevance among the most cited articles
The connection network between the authors identifies the similarity of the themes studied. Thus, publications released from the year 2000 onwards were chosen, with more than five citations of each author listed above, for analysis of the lines of research.
Several studies carried out together with other researchers encompass a series of subjects related to the alkali-aggregate reaction, becoming an international reference on the subject.
Studies related to AAR evaluative methodologies were carried out by Ideker et al. (2010), Lindgard et al. (2012), Sanchez et al. (2014), Sanchez et al. (2015a), Sanchez et al. (2015b), Sanchez et al. (2016a), Sanchez et al. (2016b), Lindgard et al. (2016), Sanchez et al. (2017) and Martin et al. (2017).
Studies have indicated that the Stiffness Damage Test (SDT) can be used as a way to assess AAR and evaluate the damage related to the expansion of the concrete (Martin et al., 2017). However, parameters such as environmental conditions, sample characteristics and sample conditioning history seem to significantly influence the results of the cyclic loading test. On the other hand, different types of surface preparation and sample size do not seem to modify the test analyses (Sanchez et al., 2014; Sanchez et al., 2015a; Sanchez et al., 2016a and Sanchez et al., 2017 ).
The Damage Rating Index (DRI), a microscopic and semi-quantitative petrographic tool, also proved to be effective for the assessment of AAR (Martin et al., 2017), and from the analyses with various reactive aggregates, a quantitative model was developed for this assessment (Sanchez et al., 2015b and Sanchez et al., 2016b).
In addition, studies evaluating the reliability and comparisons between expansion tests in mortar bars and field tests indicate that the accelerated method in mortar, despite being used for more than 15 years worldwide, still needs to be better calibrated to transpose or approach the reality of the structures. This is true for all types of aggregates and binders, as they can present false negative and false positive results. In addition, parameters such as temperature, pH and concentration of alkali in the exposure solution must be considered, as they influence the performance of the tests (Jason et al., 2010; Lindgard et al., 2012; Lindgard et al., 2016).
As for the reactivity tests of the aggregates, studies indicate that the finer the particles and the higher the temperature, the greater and faster the release of alkalis. Furthermore, thermal activation has the greatest effect on gneiss and a smaller and similar effect on granite and feldspar (Lu et al., 2006a, Lindgard et al., 2013a and 2013b). It was also observed that, when exposed to Ca(OH)2, NaOH and KOH solutions, the test on mortar bars showed more alkalis released in the solutions (with NaOH and KOH) compared to the solution saturated with Ca(OH)2 (Berubè et al., 2002).
Berubè et al. (2003a), Rivard et al. (2007) and Shehata et al. (2010) investigated the effects of alkaline content in cement and found that the alkali content in cement has a significant effect on the expansion of concrete prisms and mortars and depends on the proportion and aggregate used. Furthermore, Berubè et al. (2003b) indicated that low-alkali concretes are a good choice for marine exposure conditions or de-icing salts.
To mitigate the effects of AAR in concrete, researchers evaluated the effect of mineral additions to it and found that the presence of some additions significantly reduces the expansions generated by AAR (Duchesne and Berubè, 2001; Grattan-Bellew et al., 2003). Other studies show that cement-silica fume mixtures were not efficient in retaining alkalis in their hydration products, although cement mixtures with silica fume and fly ash have the ability to agglomerate and retain alkalis (Shehata et al., 2000; Shehata et al., 2002 and Shehata et al., 2006, Xue et al., 2021, Andrade Neto et al., 2021).
Skolu et al. (2007) and Feng et al. (2010) investigated the influence of lithium nitrate on expansions in mortars and concrete. The results obtained indicated that the lithium salt reduced the occurrence of expansions due to the formation of late ettringite, forming a protective barrier that prevents the reactive silica from being attacked by alkalis.
Munhoz et al. (2021) evaluated the effect of self-healing of improved autogenous mortar on the alkali-aggregate reaction (AAR). The results showed that the reference mortar and the mixture with 1% polypropylene microfiber showed the highest and lowest levels of expansion, respectively.
The themes developed by Brazilian researchers deal with the investigation of the characteristics of materials that influence the development of AAR (Tiecher et al., 2012; Tiecher et al., 2017), as well as the characterization of the gel produced by AAR (Johnson et al., 2017), from the investigation of works with signs of AAR (Hasparyk et al., 2009; Prado et al., 2019) and analysis of the behavior of new materials subjected to AAR (de Carvalho et al., 2009).
Tiecher et al. (2012) studied the mesostasis material present in the interstices of volcanic rocks, with results that indicated that the mesostasis in the interstitials of volcanic rocks generally consists of micro to cryptocrystalline mineral phases of quartz, feldspar and clay. Mesostasis has been identified as having different characteristics and therefore this new characterization calls for a reassessment of its influence on the reactivity of volcanic rocks. Tiecher et al. (2017) presented research on the influence of quartz deformation in aggregates on the development of the alkali-silica reaction in concrete and its relationship with the dissolution of silica. The results indicated that the presence of slightly deformed quartz indicates the presence of internal pathways available to react with alkaline solutions in the concrete pores and points to the potential development of an alkali-silica reaction. Using the visible spectrophotometer test, it was possible to evaluate the dissolution potential of different samples of deformed quartz, which confirmed that the reactivity of quartz increases as the deformation of the crystalline structure increases.
Johnson et al. (2017) evaluated the fracture properties of the gel produced by the alkali-silica reaction in concrete. The researchers used the scanning electron microscopy technique and a fracture mechanics model and estimated that the fracture energy of the gel is Gf=11.2 J/m2.
At the Furnas Hydroelectric Plant in Brazil, already identified with the presence of AAR, Hasparik et al. (2009) investigated the properties of the affected concrete. The ultrasonic results showed that the pulse velocity is sensitive to different degrees of deterioration due to AAR and correlates with the modulus of elasticity of the investigated concretes. Furthermore, the modulus of elasticity was the main property affected by AAR.
Carvalho et al. (2009) investigated the effect of using steel fibers in mortars subjected to AAR, and the results showed that the addition of steel fibers reduced the expansion by AAR. The most expressive benefit corresponded to the addition of 13.0 mm fibers in the mixture containing 2% fibers.
Most works (13 articles) are related to methodologies for assessing damage caused by alkali-aggregate reactions. This refers to the attempts to standardize the tests used worldwide and the correlation between accelerated and field methods (Lindgard et al., 2016). Furthermore, most of these investigations were led by Fournier's Canadian research group, with significant contributions from Brazilian Sanchez, Professor at the University of Ottawa.
Research related to the use of pozzolans totals 10 works, while the subjects reactivity of aggregates and alkalinity of Portland cement present proportions of similar studies, 7 and 6, respectively.
3.2 Testing methods
Accurately analyzing samples of concrete and mortar subject to AAR provides reliability and safety related to the conditions of use of certain materials in reinforced concrete structures. Thus, the standardization of the methods of carrying out tests allows the efficient comparison of the conditions of accelerated and field tests, aiming to simulate reality more accurately. Furthermore, understanding how much deterioration in existing structures has affected or will affect their performance is another difficulty to be mastered in the field of alkali-aggregate reaction research.
Within this context, the main evaluation methods for diagnosing AAR and the degree of concrete deterioration as a result of the generated expansion are presented below.
3.2.1 Petrographic Analysis
Petrographic analysis is used to identify and classify the aggregates according to their mineralogical, structural and chemical characteristics, in order to determine their possible reactive potential.
Based on the petrographic analysis of the rock, it is possible to classify the aggregate according to the AAR as potentially innocuous or potentially reactive. The maximum limits of reactive phases generally adopted to classify an aggregate as potentially innocuous are: 5% (deleterious quartz), 3% (chalcedony), 1% (tridymite or cristobalite), 3% (volcanic glass) and 0.5% (opal). However, based on the experience of the petrographer, as well as on the present features, slightly different limits can be admitted. In Brazil, the standard that establishes a method for performing petrographic analysis of representative samples of rocks for use as aggregates in concrete, with emphasis on the conditioning factors for the occurrence of the alkali-aggregate reaction (AAR), is NBR 15577 (2018) part 3.
In recent decades, researchers have sought to develop this technique, increasing diagnostic evaluation procedures and potential for expansion in structures subject to the development of AAR (Sanchez et al., 2015). In this context, several petrographic methods were developed worldwide (Blight et al., 1981; Sims and Miglio, 1992, Salomon and Panetier, 1994; Clemena et al., 2000; Rivard et al., 2000; Broekmans, 2002; Lindgard et al., 2004), but the most used is the Damage Rating Index (DRI), improved by Grattan-Bellew and colleagues (Grattan-Bellew and Danay, 1992; Dunbar and Grattan-Bellew, 1995; Grattan-Bellew and Mitchell, 2006; Sanchez, 2014).
In 2017, a Brazil-Canada collaboration work (Sanchez et al., 2017) was published to disseminate the AAR assessment by the Damage Rating Index (DRI), which is an important milestone for the insertion of the method in Brazil. Currently, the Damage Rating Index (DRI) test is registered as a FURNAS Technical Instruction document, authored by renowned Brazilian researchers Nicole Hasparyk and Leandro Sanchez (Hasparyk and Sanchez, 2021). This fact is a breakthrough such that Brazilian research should start to use this technique more frequently to investigate AAR and other expansive reactions.
3.2.2 Damage Rating Index (DRI)
The Damage Rating Index (DRI) technique is a semi-quantitative petrographic analysis performed with the aid of a stereomicroscope with 15 to 16x magnification (Figure 3), where the deterioration characteristics of the samples are accounted for through small areas with dimensions of 1 cm x 1 cm marked on the polished surface of the sample (Grattan-Bellew, 1995; Villeneuve et al., 2012). It is important to point out that in addition to quantifying the cracks found, their petrographic characteristics are also identified (Figure 3 and Table 1), which are multiplied by important factors, seeking to balance the influence of the types of cracks with respect to the resulting impact on losses of mechanical properties (Table 1) (Sanchez et al., 2015). The number obtained by the DRI is normalized to an area of 100 cm2 (Villeneuve, 2011; Sanchez, 2014; Waidner, 2016; Sanchez et al., 2017).
Figure 3 Image of the Damage Rating Index (DRI) technique performed at the Materials and Structures Laboratory of the Federal University of Paraná (LaME-UFPR).
Table 1 Weighted characteristics for DRI analysis.
| Petrographic Characteristics | Abbreviation | Weight | Group |
|---|---|---|---|
| Closed cracks in the aggregate | CCA | 0.25 | I |
| Open cracks in the aggregate | OCA | 2.00 | II |
| Cracks in the aggregate with the presence of gel | OCAG | 2.00 | II |
| Disaggregated aggregate | CAD | 3.00 | - |
| Disaggregated/shattered aggregate particle | DAP | 2.00 | II |
| Cracks in the cement paste | CCP | 3.00 | III |
| Cracks in the cement paste with the presence of gel | CCPG | 3.00 | III |
Classification of Groups:
Group I: closed cracks (opening ≤ 0.1 mm) in coarse aggregate particles or in the coarse fraction of sand particles (1 to 5 mm) (CCA);
Group II: open cracks (opening ≥ 0.1 mm) in the coarse aggregate particles or in the coarse fraction of sand particles (1 to 5mm), with or without reaction products (OCA + OCAG);
Group III: cracks in the cement paste, with or without reaction products (CCP + CCPG) (Sanchez et al., 2015) adapted from (Villeneuve, 2011).
The method proposes the classification of cracks into groups. Group I presents great variability in the number of cracks according to the aggregate used, showing values proportional to the increase in expansion. However, it should be noted that the aggregate crushing process also causes variations in the results, since the energy applied in the process can lead to the formation of small cracks inside the aggregate particles. On the other hand, groups II and III present open cracks that tend to propagate towards the transition zone of the aggregate with the cement paste. In addition to the varying appearance of cracks, it is interesting to add that the density and amount of them per cm2 follow the increasing trend due to the development of expansion of the concrete samples (Sanchez et al., 2015).
Based on the various DRI observations, Sanchez et al. (2015) proposed an interesting model that relates concrete expansion with the development of cracks generated by AAR, as shown in Figure 4 and exemplified below. At the beginning of the physical-chemical reaction (concrete expansions smaller than 0.05%), cracks caused by AAR, type A or B, are commonly found within the reactive aggregate particles. As expansion progresses to moderate levels (approximately 0.12%), new cracks continue to develop within the reactive particles. However, there is also progress in the development of pre-existing cracks, which are able to cross the transition zone and progress towards the cement paste. For high expansion levels (approximately 0.20%), most previously formed cracks in the aggregates tend to be found already propagated within the cement paste. Furthermore, at this stage of deterioration, the “law of least energy” tends to govern crack progression. In other words, there is a greater tendency for existing cracks to continue to propagate, rather than a continuous generation and appearance of new cracks. Finally, at very high expansion levels (equal to or greater than 0.30%), the cracking is seen as very intense, since there is a tendency for interconnection between the cracks formed in different aggregate particles. At this point, it is worth mentioning that the mechanical properties of concrete are extremely reduced and affected, especially the compressive strength of concrete.
3.2.3 Stiffness Damage Test (SDT)
The SDT method was initially developed by Walsh (1965); the author detected an interesting correlation between the internal cracking density of rock samples and the stress/strain behavior provided by their cyclic loading. In the early 1990s, in research led by Crouch, the use of SDT was proposed for the evaluation of concrete samples affected by the alkali-aggregate reaction. As a result, Chrisp et al. (1993) predicted the application of a fixed cyclic load (five cycles), from 0 to 5.5 MPa, at a rate of 0.10 MPa/s. The authors indicate that the modulus of elasticity and the area of hysteresis (J/m3) are good parameters for verifying the AAR, in addition to not being a destructive test and being able to reuse the samples in other analyses. The results show that the modulus of elasticity decreases, while the hysteresis area increases due to the plastic deformation generated when the cyclic load is applied (Smaoui et al., 2004a).
This method has already been studied by several authors, who aimed to identify test characteristics that influenced the results (Smaoui et al., 2004a; Sanchez et al. 2014; Sanchez et al. 2015; Sanchez et al. 2017a; Sanchez et al., 2017b). Smaoui et al. (2004a) tested other load application rates such as maximum, 5, 10 and 15 MPa, so that the load of 10 MPa showed the best results in general.
Sanchez et al. (2014a), Sanchez et al. (2015a), Sanchez et al. (2017) studied the application of techniques in various concrete mixtures with different aggregates and also in testimonies extracted from structures affected by AAR. The results showed that the use of an application load of 40% of the concrete strength at 28 days showed better performance to diagnose the degree of AAR damage.
Sanchez et al. (2016a and 2017) indicate a strong correlation between the SDT and DRI tests, as both measure damage caused by AAR in a similar way and can complement each other and show the same trends in relation to expansion.
The SDT numbers express the mechanical characteristics of the concrete; that is, as the expansions caused by the AAR progress, the cracks also increase, decreasing the modulus of elasticity and tensile strength (Sanchez et al., 2017). Figure 5a shows the SDT test being performed at LaME-UFPR, while Figure 5b shows an output result of the SDT test, where the parameters called Stiffness Damage Index (SDI) and Plastic Deformation Index (PDI) are identified. They are translated as the ratio of dissipated energy over the total energy applied to the sample and the ratio of plastic deformation to the total strain applied to the system, respectively. Finally, the modulus of elasticity and the Non-Linear Index (NLI) can also be obtained from the test.
In Brazil, this evaluation technique was published by Sanchez et al. (2017) in a Brazil-Canada collaboration and, currently, the Stiffness Damage Test (SDT) is registered as a Technical Instruction document by FURNAS (Hasparyk and Sanchez, 2021). This is yet another national advancement such that Brazilian research should apply SDT more frequently in research on expansive reactions in concrete.
Figure 5 a) performance of the SDT test at the Laboratory of Materials and Structures of the Federal University of Paraná (LaME-UFPR); b) stress-strain behavior for concrete samples affected by AAR (Sanchez et al., 2017)
3.2.4 Accelerated Mortar Bar Method
The accelerated mortar bar test (AMBT) is standardized in Brazil by NBR 15577: 2018 - part 4 and in Canada by CSA A23.2-25A.
This is one of the most widespread tests in the world and consists of evaluating the variation in length of the mortar bars in order to verify a potential expansion due to the occurrence of the alkali reaction in a short period of time (16 and 30 days in the case of NBR 15577:2018 - part 4 and 14 days in the case of CSA A23.2-25A). Because it is an accelerated test, it occurs under severe exposure conditions (high temperature, 80ºC, and immersion in an alkaline 1N NaOH solution), which generates several criticisms regarding its use and variations in results.
Figure 6a shows the mortar specimens that are submitted to the expansion test by AAR and Figure 6b shows the measurement of the dimensional variation being performed in AMBT.
Several authors use this method to assess reactivity in aggregates (Fournier and Bérubé, 2000; Thomas et al., 2006; Alaejos et al., 2014; Golmakani; Hooton, 2016) and indicate that it can generate false positive and false negative results. Aggregates that result in false negatives may initially appear as harmless, but over time have reactive potential. It is also possible to obtain results of aggregates that can be used in concrete without damage, but the AMBT results discard them, generating false positives.
Several researchers seek to improve the AMBT method in order to remedy existing problems (Shon et al., 2002; Thomas et al., 2006; Sanchez et al., 2011; Alaejos et al., 2014), but it remains a highly criticized test due to its drastic conditions of acceleration of expansions by AAR, as reported (Thomas et al., 2006; Alaejos et al., 2014; Golmakani; Hooton, 2016; Fournier; Bérubé, 2000).
3.2.5 Concrete Prism Method
The Concrete Prism Test (CPT) is generally considered the best indicator of field performance when studying the alkali-aggregate reaction, and consists of measuring expansions of concrete prisms subjected to an environment of high humidity (95%) at a temperature of 380C. However, it requires an evaluation period of one year for studies of aggregate reactivity and two years for the use of reaction mitigating materials (Ideker et al., 2010). The fact that it only generates results in the long term is one of the reasons why many studies and companies tend not to follow this test method, even though it is considered more appropriate than the AMBT (accelerated mortar bar test) method. Figure 7a shows the concrete prisms used for the expansion test and Figure 7b shows the test being carried out.
In Brazil, NBR 15577-6 (2018) establishes the necessary procedures for the evaluation of AAR in concrete prisms. It is indicated to determine the reactive potential of the aggregates in comparison with concrete with standard cement, with the addition of sodium hydroxide, for a period of 365 days. As for the evaluation of supplementary materials or AAR inhibitors, the investigation period extends over 2 years.
It is worth noting that although the CPT is considered to be the most reliable test procedure in existence, a major disadvantage of the method is its long test period (1-2 years). In an attempt to solve this problem, accelerated CPT procedures were proposed in the 1990s (Ranc and Debray, 1992). Notable among them was an accelerated CPT version (i.e., ACPT) performed at 60oC, showing promising results worldwide. It was found that reliable results can be achieved within 3-4 months or 6-8 months to assess the potential reactivity of aggregates or the efficiency of preventive measures, respectively.
This method has been standardized in Europe by RILEM (Lindgård, 2011), although in North and South America it is currently not a standard, as some problems related to alkali leaching and important variability in test results have been observed (Kermit, 2017).
The literature also suggests a test period of 3 months for the ACPT method to determine the reactivity of the aggregates (Degrosbois and Fontaine, 2000; Touma et al., 2001; Fournier et al., 2004) and six months for the determination of effectiveness of supplementary cementitious materials (Touma et al., 2001). In Brazil, NBR 15577-7 (2018) indicates the use of the accelerated concrete prism test (ACPT), which establishes a temperature of 60oC for a period of 140 days with humidity above 95%.
It is worth mentioning that Sanchez (2008) and Sanchez, Kuperman and Helene (2011) studied another accelerated method with the objective of reducing the problems resulting from leaching that exist in the ACPT method and to seek a balance between the internal environment (test specimen) and the external one (solution in which the specimens are immersed); that is, a solution with less NaOH is used at a temperature of 80ºC and carried out for a period of four weeks. It is called the accelerated Brazilian concrete prism test (ABCPT). The results obtained using this method were satisfactory compared to the CPT (89% correlation).
3.2.6 Comparison between the evaluation methods of expansion by AAR
Many AAR research tests were identified, some standardized and others not, some more consistent with field conditions and others not. Thus, in this section we chose to show an overview of the frequency of use of the test methods covered between the years from 2000 to 2022. In this context, Figure 8 presents a comparison of the frequency of use of the types of expansion monitoring methods applied to assess AAR. The survey was carried out from the systematic search of studies published in the WoS and SCOPUS databases from the year 2000 onwards, with more than 5 citations from both countries (Brazil and Canada).
Figure 8. Methods for evaluating the expansion caused by AAR applied in research in Canada and Brazil.
It is observed that the accelerated mortar bar test (AMBT) is the one that predominates in studies related to AAR in Brazil. It should also be noted that, in Brazil, the other methods (CPT, SDT and DRI) were not used in the articles consulted in the survey carried out for this research. This fact is explained by the time and ease of execution of AMBT (results in 4 weeks) compared to concrete prisms (results in 1 to 2 years). It is worth mentioning that this research does not consider Brazilian journals with great relevance in the country, so works that use other evaluation methods may have been published nationally. In this context, the publication by Sanchez et al. (2017) and the FURNAS Technical Instruction document (Hasparyk and Sanchez, 2021) are important recent works that should make these methods of assessing damage by AAR applied more in research carried out in Brazil.
Thus, the little use of SDT and DRI analyses is due to the little knowledge and diffusion related to them in Brazil. In addition, they are methods for evaluating the state of deterioration in concrete test specimens that were not found in the articles in the survey of this research, with the published works coming from Brazilian research based only on mortars (AMBT).
On the other hand, in Canada, the concrete prism test is used more frequently than in Brazil. Figure 8 also shows that the SDT and DRI are frequently used tools to measure the extent of damage caused by AAR in Canada. The frequent use of these methods (CPT, SDT and DRI) in research in Canada is based on the fact that the AMBT has a dubious reliability. In addition, the SDT and DRI are useful to understand the present deterioration of concrete by AAR and to determine whether the expansion by AAR has affected the concrete structure and its mechanical properties.
3.3 Overview of AAR research in Brazil
In Brazil, the first registered studies addressing the AAR problem occurred in 1963, when the characterization of the aggregates that would be used in the construction of the Jupiá dam, located between the states of São Paulo and Mato Grosso do Sul, was carried out. The study was carried out by Gitahy and Ruiz in partnership with IPT - Instituto de Pesquisa e Tecnologia (Research and Technology Institute), and Centrais Elétricas de Urubupungá S.A. and found that the aggregates had reactive elements in their composition that could react like the alkalis present in cement, as Stanton published in 1940 (Lima, 2009; Priszkulnik, 2005).
According to Hasparyk (2005), since the discovery of AAR, several studies have focused on understanding the chemical process and characterization of reactive minerals, as well as the causes and consequences in works affected by this deterioration process. Haspariryk (2005) indicated that at least 19 dams in Brazil were identified with signs of AAR, which compromises the durability of the structures, in addition to the impact on the country's economy caused by the malfunction or suspension of activities.
With regard to the characterization of materials, data collected by the Brazilian Association of Portland Cement (ABCP) in 2015 indicate that of the aggregates already analyzed, 36% have reactive potential. In addition, the number of civil construction companies that seek to analyze the materials used in their works has been increasing every year, where most requests refer to the states of São Paulo and Pernambuco (Battagin et al., 2016).
Regarding the state of Pernambuco, in 2004 there was a milestone regarding the occurrence of AAR in Brazil, which was the first national proof of AAR in building constructions. This occurred in the Metropolitan Region of Recife (RMR), state of Pernambuco, with the expert investigation of the collapse of the Areia Branca residential building. Cracks due to AAR were concentrated in the foundation blocks of the Areia Branca building, but it is worth noting that the building did not collapse due to the deterioration process caused by AAR (Silva et al., 2021). In the following years, with inspection work led by Dr. Tibério Andrade, Professor at the Federal University of Pernambuco (UFPE), several residential buildings were found with foundation blocks cracked by AAR. These facts resulted in very important discussions for the technical and scientific environment, so that inspection procedures and recovery methods developed specifically for the solution of these cases emerged.
The growing concern with the subject has increased the number of studies related to AAR in Brazil. Figure 9 shows the growth of Brazilian publications related to the subject in high-impact journals, in the WoS and Scopus databases.
Figure 9 Growth of publications related to AAR in Brazil in high-impact journals present in the WoS and Scopus databases.
It can be seen from Figure 9 that publications related to AAR in Brazil, in addition to having increased over the years, show a leap from 2007. This fact coincides with the studies developed for the publication of the Brazilian regulatory standard in 2008, NBR 15577, which is divided into 7 parts and refers to the test procedures for the characterization of aggregates and binding materials subject to AAR. It was revised and updated in 2018.
Concomitantly with the publication of NBR 15577:2018, the Alkali-Aggregate Reaction Prevention Guide - Recommended Practices by IBRACON (Brazilian Institute of Concrete) was launched in 2018. The work was developed by the Alkali-Aggregate Reaction Technical Committee from IBRACON, with the coordination of researchers Cláudio Sbrighi Neto, Eduardo Quitete and Arnaldo Battagin.
With regard to Brazil's performance as a host and organization of events in the field of research on AAR, in 1997, the 1st Symposium on Alkali-Aggregate Reactivity in Concrete Structures took place, which was a landmark incentive for Brazilian research on AAR. It should be noted that 1997 also coincides with the first records of Brazilian research (article publication) in international databases, as can be seen in Figure 9.
In 2006, the 2nd Symposium on Alkali-Aggregate Reaction in Concrete Structures was held, at which time Brazil was already at a higher level in research on AAR and the verification of the reaction in the foundation blocks that occurred in buildings in Pernambuco was evidenced in Brazil. It should be noted that shortly after this event, Brazil launched its first standard on AAR, the aforementioned NBR 15577/2008. It should also be noted in Figure 9 that the participation of Brazilian researchers in publications in high-impact national journals grew a lot after the 2nd Symposium on Alkali-Aggregate Reaction in Concrete Structures.
In 2016, another Brazilian landmark on research on AAR occurred when Brazil hosted the 15th International Conference on Alkali-Aggregate Reaction (15th ICAAR), considered the most important conference in the world on research and development on AAR. This is a merit of the national researchers and the editors of the annals of this event, Dr. Haroldo de Mayo Bernardes and Dr. Nicole Pagan Hasparyk. Figure 9 also shows an increase in the participation of Brazilians in publishing articles in high-impact journals from 2016, when ICAAR was held in São Paulo, Brazil.
Thus, it is evident that Brazil plays an important global role in advancing research related to AAR, participating and encouraging national and international discussions provided by its specialists who discuss knowledge in specific committees on the subject.
3.4 Future approaches and knowledge gaps
The publications of the last three years in Brazil and Canada, referring to the researched topic, amount to 19 articles. With these works, a bibliographic coupling analysis was carried out, which measures the relationship between two articles based on the number of common references cited (correlations), in order to define similar areas of research that are on the rise, as indicated in the work of Lucas; Garcia-Zorita, Sanz-Casado (2013).
Of these 19 articles, those with at least two citations from the two databases under study were selected and the research themes were correlated between them. These works are presented in Table 2.
Among the works presented, two large groups of research can be observed, which are related according to the studied themes. The largest group identified refers to the study of alkali-aggregate reactions in general, as well as their implications for reinforced concrete structures. Within this group are the works by Campos et al. (2018) and Gorga et al. (2018) with 21 and 26 correlations, respectively. This group indicates that the theme related to AAR is still on the rise, as well as the various parameters related to tests and the forms of deterioration observed with attempts to model the behavior of concrete with AAR. It is important to highlight that the two works cited as outstanding in this group, Campos et al. (2018) and Gorga et al. (2018), present Brazilian researchers as first authors.
The other group is represented by the works by Angulo-Ramirez et al. (2018) with three correlations and Tiecher et al. (2018), with 5 correlations. These papers refer to alkali-aggregate reactions in alternative cementitious systems, such as alkali-activated cements and the use of pozzolanic additions. This study group seeks a better understanding of the durability of cements with low clinker content and sees it as a theme of greater originality with few bibliographies available so far, and which should be studied more in-depth over the next few years. These two works also present the performance of Brazilians in the development of research, which shows Brazil's participation at the forefront of research on AAR.
In general, works related to AAR deal mainly with the characterization of materials, forms of AAR mitigation, evaluation methods and identification of works with a deleterious reaction. Despite the topic being recent and most works referring to the implications of AAR in concrete structures, among the gaps in knowledge on the subject, the study of reaction mechanisms in aggregates of volcanic and dolomitic igneous origins can be mentioned, as well as the characterization of AAR in alternative cements. In addition, emphasis should be given to the development of methodologies to recover or contain the effects caused by AAR.
Table 2 Correlation analysis between the topics studied regarding AAR in Canada and Brazil.
| Themes | Authors | Year | Citations | Correlations |
|---|---|---|---|---|
| Alternative cements | ||||
| Alkali-activated Portland blast furnace slag cement mortars: Performance to alkali-aggregate reaction | Angulo-Ramirez et al. | 2018 | 17 | 3 |
| Alkali-Aggregate Reaction: A study of the influence of the petrographic characteristics of volcanic rocks | Tiecher et al. | 2018 | 2 | 5 |
| Alkali-aggregate reaction and structural implications | ||||
| Effects of an internal sulfate attack and an alkali-aggregate reaction in a concrete dam | Campos et al. | 2018 | 12 | 21 |
| Deterioration Assessment of Infrastructure Using Fuzzy Logic and Image Processing Algorithm | Pragalath et al. | 2018 | 9 | 0 |
| FE approach to perform the condition assessment of a concrete overpass damaged by ASR after 50 years in service | Gorga et al. | 2018 | 6 | 26 |
| Impact of pore solution concentration on the accelerated mortar bar alkali-silica reactivity test | Golmakani and Hooton | 2019 | 5 | 1 |
| Structural implications of internal swelling reactions in concrete: review and research needs | Noel et al. | 2018 | 2 | 12 |
| Material, structural and modelling aspects of alkali aggregate reaction in concrete | Ferche et al. | 2019 | 2 | 10 |
4. Conclusions
The analysis carried out from the systematic review shows that Canada is the country that most stands out in the publication of scientific articles in high impact journals internationally in the scope of the alkali-aggregate reaction. It is also evidenced in this work that Brazil is the seventh country in this classification ranking. The comparison between published research of Canadian and Brazilian origin indicates that Canada produces three times more works of international scope than Brazil related to the AAR theme.
It was also evident that several of these Canadian studies are carried out jointly with Brazilian authors. The interconnection between citations and authors indicated that the works carried out between the two countries are divided into three large groups of researchers led by Fournier, Thomas and Berubè, thus identifying the correlation between the topics studied.
Regarding the thematic categories, four main areas of research were identified, namely: AAR evaluation methodologies, use of cement additions, aggregate reactivity, and cement alkalinity. The topic with the highest number of publications refers to the various methodologies applied to the assessment of AAR, led by the Canadian Benoit Fournier, with significant contributions from the Brazilian Leandro Sanchez.
Regarding the AAR detection methods, it was noted that in Brazil, the concrete prism test is little used in the evaluated studies, despite being standardized. In addition, SDT and DRI analyses are not widespread in the country, although there are national publications in recent years that tend to change this context.
With the criterion and procedure of bibliographic survey applied in this research, it was demonstrated that the AAR studies developed in Brazil are based on AMBT, with no detection of the use of more reliable tests such as: concrete prism test (CPT), Stiffness Damage Test (SDT) and Damage Rating Index (DRI).
Despite this, Brazil presents an advance in the number of studies related to the theme of AAR, in order to identify, mitigate and solve the pathological manifestations resulting from it. In the country there are many specialists dedicated to the subject, having organized several specific conferences on AAR, indicating the national protagonism in this field of research.
As for future studies, there is a trend related to AAR and its structural implication and the effect of AAR in alternative cements without the use of clinker. There are efforts in search of a better understanding of the reaction mechanisms in distinct aggregates and innovative materials, as well as ways of recovering structures already affected by the reaction.
Brazil shows an increase in publications on AAR in high-impact journals over the years and showed a preponderant presence when it came to future approaches and the knowledge gap.
Through a project funded by Capes and the National Water Agency (ANA), and in partnership with the University of Ottawa (Canada), the Materials and Structures Laboratory (LaME) at the Federal University of Paraná (UFPR) began to develop research applying the mentioned methods. In the near future, it is intended to publish many studies including results of CPT, SDT and DRI in experimental programs developed in Brazil.
