Revista Ingeniería de Construcción Vol.24 N°2, Agosto de 2009 www.ing.puc.cl/ric PAG. 141-152

Evaluation methods of alkali-silica reaction in concrete with recycled aggre-gates

Miguel Barreto Santos*, Jorge de Brito**¹, António Santos Silva***

* Instituto Politécnico de Leiria, PORTUGAL

**Universidad Técnica de Lisboa, PORTUGAL

***Laboratorio Nacional de Ingeniería Civil, PORTUGAL

Corresponding author:

ABSTRACT

Alkali-silica reactions (ASR) are one of the causes of chemical degradation of concrete with natural aggregates (CNA) that compromise its durability. The introduction of recycled aggregates (RA) in concrete creates changes in their properties and differences in the results of the evaluation tests of ASR. Existing bibliography emphasizes special care in the evaluation of RA and concrete with recycled aggregate (CRA) for ASR and changes are proposed to the existing test methods. There are proposals to change the accelerated test of mortar and concrete with RA to accelerate the reactions and recommendations to prevent changes in the characteristics of the RA, during the preparation of samples. Some articles recommend the pre-saturation of the AR, due to its absorption of water, to avoid significant variations in the results of expansion of concrete samples from early ages. This article aims to briefly describe the characteristics of CRA and ASR presenting comments from existing bibliography to the evaluation methods of ASR in CRA and CNA.

Keywords: Concrete with recycled aggregates, alkali-silica reaction, durability, degradation

1. Introduction

The attack by internal sulfate reactions and by alkali-aggregate reactions may cause concrete degradation because of the so called "expansive reactions from internal origin". Alkali-aggregate reactions (AAR) include alkali-silica reactions (ASR), alkali-silicate and alkali-carbonate reactions, even though ASR are the most outstanding ones.

Changing the properties of concrete with natural aggregates (CNA) because of ASR, mainly reduces flexural and tensile capabilities, decreases the elastic modulus and at a minor scale reduces compression strength. The reaction also causes other effects in concrete such as, for example, the increase of volume or cracks on the concrete surface, which allow external aggressive agents to ingress which increase the susceptibility of structures to corrosion reinforcement effects or freezing thawing cycles, among others.

The most affected structures by this noxious reaction are the dams, the bridges, the piers, and the highway pavements, although there is evidence of ASR in other kind of concrete structures.

There is a set of international recommendations regarding the thematic of prevention and mitigation from the expansive reaction in CNA. Such recommendations are available from E461 Specification at the National Laboratory of Civil Engineering (LNEC) in Portugal, under the title "Concrete. Methodologies to prevent internal expansive reactions (LNEC E461, 2006).

The use of recycled aggregate (RA) changes concrete properties, demanding special care on traditional ASR tests methods. The following chapters briefly describe the properties of concrete using recycled aggregate (CRA), the problems arising from ASR in CNA as well as some other remarks referred to ASR test methodologies in CRA.

2. Concrete with recycled aggregates

The RA have several applications, as referred by Brito (2005), Carrijo (2005), Miranda (2000) and Hansen (1992). RA can be used in highways pavement base layer and sub-base layer or in landfills of a such diverse nature that a very rigorous selection of components is not required. The most demanding applications such as mortar and structural concrete require an expert knowledge about capabilities and restrictions of the RA to be included in the mixtures.

The grading and crushing process of grading and crushing aggregates, that provide shape and texture to particles, also influence mechanical resistance, workability and cement consumption in CRA (Brito, 2005).

Bibliography analysis indicates that characteristics of adhered mortar are the main reasons of mechanical resistance decrease in RA. Alaejos and Sánchez (2004), for example state that RA coming from concrete having compression resistance value below 25MPa, should not be used.

Bibliography references generally indicate that workability reduction of CRA is higher, as long as the replacement rate of natural aggregates (NA) by recycled aggregates (RA) is also high. This is mainly because of the higher water absorption by RA, as proven in tests made by Santos (2002), Matías and Brito (2005) and Evangelista (2007) for a constant effective water/cement (w/c) ratio. This change in the property of fresh concrete can be reduced, for instance, following recommendations from Alaejos et al. (2004). Those are the quantification of additional amount of water to put in concret during production, the aggregate pre-saturation or the use of admixtures (supplementary cementing materials or mineral addition).

Difference between compression resistance of CRA and CNA rises as per the increase of concrete type of strength. However by keeping the grading curve of the aggregate and the fresh concrete workability just identical, the resistance of CRA compared to the resistance of CNA happens to yield satisfactory values (Brito, 2005). However, there are some additional constraints to keep in mind such as, for example, the cement consumption, the replacement ratio, the fine aggregate percentage included in the mixture and the kind of aggregate, the ratio w/c and the adjuvant utilization. As far as tensile resistance is concerned, Brito (2005), in is synthesis lesson, considers that value difference is still present when adding RA to concrete, however it is less prominent compared to compression resistance.

Aggregate roughness is normally higher in RA than in NA, contributing to a higher abrasion resistance, although CRA and CNA abrasion behavior may also depend on each effective w/c ratio adopted, on the porosity, on irregular aggregate surface and also on the content of binder. Bibliography shows a trend of decreasing the elastic modulus and increasing CNA shrinkage by the replacement rate of NA by RA (Brito, 2005).

Concrete permeability is one of the facts that affects its durability, allowing damage and aggravating of some degradation causes such as carbonation, chloride attacks and ASR development. The porosity on CRA, contributes to a higher exposure to foreign agents influence of this concrete, as indicated by (Levy 2001) on his CRA durability study.

3. The alkali-silica reactions

Constraints on the durability of concrete may have chemical or physical origin. Chemical facts for concrete degradation are (Santos Silva, 2007):

•  Seawater attack;

•  Sulfate attack;

•  Fresh water and acid environment attacks;

•  Biochemical attack;

•  Reinforcement corrosion;

•  Internal expansion reactions.

ASR, which is included in concrete internal expansion reactions, scope of this study are defined as the reaction between an alkaline solution and some silica minerals that in the presence of water generate an expansive gel. For ASR to develop, the simultaneous presence of appropriate content of moisture, alkalis and reactive aggregate will be required in the concrete (Fernandes, 2005).

One of the main alkali supply sources is cement although, any sodium or potassium source can also contribute to trigger the reaction, as indicated in the technical report TR3 (2003). Thus, the alkalis amount contained in the concrete must also register the different external and internal potential alkalis sources, specifically alkalis from aggregates.

The equivalent sodium oxide (Na₂O_eq = Na₂0 + 0.658 x K₂0) is conventionally used to indicate the alkalis content in the Portland cement, being normally limited in order to mitigate ASR to values lower than 0.6% (ASTM C150-02, 2003), even though some authors recommend lower limits (Stievenard-Gireaud, 1987; Prince and Perami, 1993). The most recent recommendations suggest that the control of Na₂O_eq content in concrete must be equal to the addition of Na₂O_eq, content and components, limited to 3 kg/m³, except for vulnerable structures whose value is even lower.

Reaction mechanism in ASR has been studied by several researchers as Dent Glasser and Kataoka (1981a, b), Chatterji (1989), Hobbs (1988) among others. At the present time, different scientific contributions have led to understand the reaction mechanism in ASR for two different models, the topochemical and the dissolution/precipitation model (Santos Silva, 2006). In the topochemical model, the reaction is described as being developed on the reactive aggregates surface without necessarily a movement of the aggregate reactive species to the solution. In the dissolution/precipitation model the reaction is developed in the interstitial solution after different reactive species have changed into ionic state. The formation theories for alkali-silica gel and its expansive properties are presently inserted in these two models.

The expansibility of gel produced by the ASR development, causes concrete degradation because of diverse mechanical effects to the material and to the structure. Normally, concrete degradation by ASR may take long before revealing itself. However, cracking, bleeding, efflorescence, pop-outs, delamination and structure expansion are the evidences of degradation. Subsequently an in-situ diagnosis is confirmed by means of microscopic examination in the laboratory.

Figure 1 illustrates some structures affected by ASR, showing the typical cracking of the reaction.

Figure 1. ASR macroscopic revelation. Source: figure a) (Pecchio et al., 2006); figures b) and c) (Santos Silva 2007)

There is no standard solution to repair the structures affected by ASR yet, so (Santos Silva, 2007) in order to guarantee the structures durability faced with this noxious reaction, it is necessary to take preventive measures and follow the recommendations in construction materials selection, by using non-reactive alkali aggregates, reducing interstitial solution alkalization of concrete and controlling moisture.

However, there are some materials that included in the concrete mixture, or in some cases, applied on a structure affected by ASR, may mitigate these reactions. Apart from adding type II minerals, adding chemicals based on lithium salts are the most mentioned materials in the bibliography.

4. Research on ASR evaluation methods for CRA

At the 11^th ICAAR, the researchers Gress and Kozikowski (2000) introduced a study referred to methodology changes on different trial methods in ASR for CRA and CNA, in order to accelerate reactions and avoid long waiting periods for final results. These authors justified the research on the need to have an accelerated ASR testing method, to evaluate RA and its incorporation in CRA, so that the concrete mixture may be easily tested. This deficiency rose during the ASR degradation sequence of several kilometers of pavement in the United States. Recycling possibility as an option for pavement restoration, led to study RA reactivity and ASR prolongation in CRA after alkali content increased.

Trials performed were based on the accelerated mortar and concrete tests in accordance with ASTM C 1260 and ASTM C 1293 standards respectively, although using concrete prisms with in different shape and preservation conditions. The tests were accelerated far beyond standards, by increasing temperature, using microwave energy, using ultra-sonic energy and increasing alkalis contents. For example, the ASTM C 1260 is an effective tool for evaluating NA using 25 x 25 x 280 mm little specimens, however this test was not used to evaluate RA due to maximum size restrictions of the aggregates in the for small specimens. Therefore, the authors suggested increasing specimen sizes of ASTM C1260 test up to 76.2 x 76.2 x 279 mm. ASTM C1293 test, similar to RILEM AAR-3, is considered by the authors as the most adequate test to relate in situ concrete behavior to lab concrete behavior. Since this test yields results only after a year, Gress and Kozikowski (2000) suggested some modifications such as increasing the maximum temperature, using different moisture conditions with concrete samples sealed in evacuated plastic bags containing water, in order to obtain correlated expansion results statistically comparable to ASTM C 1293 test, in a shorter periods.

For the experimental trial, the following aggregates were employed: coarse RA coming from concrete from a recycled pavement affected by ASR; NA used in the original concrete (OC) of the pavement; and innocuous fine NA. The modified ASTM C 1260 test results using modified specimens (solids, with holes, with sides cut, among others) and alkali content, figure 2, indicate that the RA shows a minor expansion than NA, when employing Na₂0_eq cement at 1.15%. thus the authors justify the RA minor reactivity per volume unit and the higher porosity capability to avoid alkali-silica gel without expansion. Therefore, this test confirms that CRA has a minor expansive reaction than CNA, although the contrary occurs upon alkalis increase.

Figure 2. Results from ASTM C1260 modified test using (A) NA, (B) RA at 1.15% of Na₂O_eq, (C) RA at 1.37% of Na₂O_eq (Gress and Kozikowski, 2000)

Gress and Kozikowski observed a good correlation between results obtained from modified ASTM C 1293 tests and the results obtained with microwave energy tests. They concluded that alterations applied to the ASTM C 1260 and ASTM C 1293 tests actually accelerated the ASR. The tests employing microwave energy also caused effects on the accelerated expansive phenomenon. The analysis from modified tests and the research on ASR in CRA were studied by Scott (2006). The author studied the ASR mitigation in CRA researching the CRA reactivity using concrete RA that presented degradation due to ASR. He also researched initial hydration moisture existing in CRA, and the results from modified ASTM C 1260 and ASTM C 1293 tests. In its bibliografhic research, he said that he has found little available information about ASR in CRA however, he found some recommendations to develop previous reactivity tests on ASR in RA. Regarding the RA origin and the ASR, Scott (2006) concluded that either the RA comes from a concrete with ASR distress, or from concrete where signs of this noxious reaction was not noticed. Thus in the case when the CRA contains RA that had been exposed to degradation by ASR, this reaction may not develop because it has been extinguished in the original CNA. On the other hand, the reaction could have never been developed because the ideal conditions to expand were not available, and may appear in the CRA if such conditions are present. It may also happen that if the employed RA comes from a partial reactive concrete, then degradation by ASR may develop immediately in the new CRA, leading to a reduction in structure lifetime. The same author refers to some other conditions causing degradation in CRA due to ASR, when using RA from concrete that reacted only partially, such as:

•  For instance, pH value variation. When the pH of CNA does not have compatible values with ASR, if CRA is incorporated in a mix having a high pH value, then ASR may develop;

•  The recycling process and environmet alterations for instance in the CNA where ASR has not developed due to lack of favorable conditions, may initiate expansive reactions;

•  Alkalis contained in the alkali-silica gel of the RA that had developed ASR may become harmful for CRA since they may contribute to a new ASR development;

•  Cement and RA alkalis may increase the pH in CRA thus becoming more aggressive than OC pH where the cement was the only alkalis source.

During the experimental campaign, Scott (2006) used cement with high alkalis contents and, as coarse aggregate, a Blue Rock control aggregate and a RA concrete from pavement containing NA from Blue Rock, well known as reactive agents also used in studies by Li (2005). The author used non-reactive sand as fine NA and limestone coarse NA, that helped to control and checked if, besides reactive coarse aggregates, some other material contributed to ASR. For the evaluation of reactive material, he applied the ASTM C 1260 and ASTM C 1293 tests. He also studied alternative tests to accelerate the results by modifying specimens dimensions, applying electric current on the concrete specimens and changing moisture and temperature conditions in such tests. Such alterations were based on the studies made by Kozikowski (Gress and Kozikowski, 2000; Kozikowski, 2000).

The use of ASTM C 1260 test for RA is commented by Scott (2006), referring indicating that size reduction of RA particles to be included in the mortar mixture causes the pulverizing of matrix aggregates. The pulverizing process crushes the RA separating the NA from adhered mortar, not testing the RA as a whole. The author examined the NA reactivity content in RA, after separating adhered mortar, by using the ASTM C 1260 test. The alterations to mortar bar method ASTM C 1260 consisted of replacing the mortar bar by concrete prisms used for the ASTM C 1293 test, but maintaining the conditions of ASTM C 1260 test, except for the expansion limit value that was modified from 0.10 to 0.04% at 28 days. He also analyzed the influence of replacing a concrete bar by a cube with side holes, which facilitate the NaOH solution penetrability, thus increasing kinetic reactions. For the concrete prisms test specified encouraged by ASTM C 1293 standard, the author made some modifications as far as specimen dimensions are concerned and he also used plastic vacuum sealed cubes with side holes, standardizing moisture distribution.

Considering that cathodic protection of steel promotes the ASR, though protects the reinforcement, Scott (2006) also studied a method to accelerate the reaction and get expedite results by passing electric current - Figure 3.

Figure 3. Expansion test by passing electric current (Scott, 2006)

Scott (2006) observed from expansion tests that CRA produced had a high initial expansion, wich he attributed to the water absorption by the concrete mixture during the initial concrete hydration process. Thus, in accordance with Li (2005), the author pre-saturated the RA before adding them to the test mixes, by filling the pores and assuming that initial expansions would be avoided. The RA concrete was placed under water during 48 hours and water from pre-saturation was added to the mixture, thus avoiding loss of alkali. He also applied vacuum saturation in other RA set, in order to test expansions from ASR only. Using the methodologies described, the author observed the effect of some mineral additions, such as fly ash, blast furnace slag and silica fume, as well as the effect of using low alkali content cement, and the addition of lithium nitrate to mitigate ASR in CRA. It was verified that only low alkali content concrete, 55% of blast furnace slag, 25% of fly ash or 100% of lithium nitrate, were effective in ASR mitigation, even though initial preparation proceedings of aggregates also had to be considered. The author also confirmed the highest content of soluble alkalis existing in the concrete RA, in relation to the control NA, wich was attributed to the mortar fraction in RA.

According to his conclusions, this researcher considers that tests on ASR in CRA should be focused on the material characteristics, including absorption, mortar fraction and alkali-free rates. He also comments that results for modified tests showed satisfactory correlations, however, he suggested to continuing to follow up some cases.

High initial expansions in concrete were also studied by Li (2005) concerning ASR mitigation in CRA. However, this author indicates that expansions are clearer in the concrete tests than in mortar tests, which is explained by the small size of particles.

The methodology for the execution of the accelerated mortar bar method was modified by some researchers as Li (2005), Scott (2006), Etxeberria (2004) or Shehata et al. (2008), among others, in order to test the RA reactivity. Such researchers consider that in order to test the AR reactivity, the mortar bar method must be applied separately to the aggregate and to the adhered mortar contained in the RA. The adhered mortar pulverizing, changes the characteristics of the RA during aggregate reduction, before pouring it into the mixture.

Shehata et al. (2008) also consider that the accelerated bar mortar method should also include the fine RA resulting from secondary crushing, that is from coarse RA. Differences were found in relation to mortar expansion values for fine RA resulting from original concrete primary crushing and resulting from coarse RA secondary crushing process. The research also confirmed the presence of reactivity in CRA produced with RA coming from CRA affected by ASR. These authors believe that reactivity found in CRA was strengthen by CNA crushing, thus allowing new contact fresh phases between reactive NA and the new cement in the mixture, causing expansions in the new concrete.

5. Conclusions

It has been found out that some mortar tests for the evaluation of ASR, even though faster, may result in changes in the results due to the fragmentation of RA for testing that leads to pulverizing the adhered mortar and creating fine RA with different characteristics from those of the original coarse RA. Some studies presented recommend developing separate tests for adhered mortar and for NA, contained in RA, otherwise only RA, resulting from a secondary crushing process should be used. There are also proposals to modify the accelerated ASTM C 1260 test for RA by using concrete specimens that would avoid alteration of coarse RA characteristics.

Due to water absorption by RA, some articles recommend pre-saturation to avoid significant variations for expansion results on concrete specimens at early ages.

6. References

Alaejos R, Sánches M., Dapena E., Vásquez E., Barra M, Etxeberria M., Mari A., Agullo L., Marínez F., Fonteboa B., Polanco A., Francisco G., Aleza F., Parra J., Buron M. (2004), "Draft of Spanish regulations for the use of recycled aggregate in the production of structural concrete", International RILEM Conference on the use of recycled materials in buildings and structures, Barcelona, pp. 511 525

Alaejos P., Sánchez M. (2004), "Utilization of recycled concrete aggregate for structural concrete", International RILEM Conference on the use of recycled materials in buildings and structures, Barcelona, pp. 693 712

ASTM C 1293 (2004), "Standard test method for determination of length change due to alkali silica reaction", The American Society for Testing and Materials, Philadelphia, 6 p.

ASTM C 150 02 (2003), "Standard specification for portland cement", The American Society for Testing and Materials, Philadelphia, 7 p.

ASTM C1260 (1994), "Standard test method for potential alkali reactivity of aggregates (mortar bar method)", American Society for Testing and Materials

Brito J. de (2005), "Agregados reciclados e sua influencia ñas propriedades dos betóes", Licáo de Síntese de Agregacáo em Engenharia Civil, Instituto Superior Técnico, Universidade Técnica de Lisboa, Lisboa, Marco

Carrijo P. M. (2005), "Análise da influencia da massa específica de agregados graúdos provenientes de residuos de construcáo e demolicáo no desempenho mecánico do concreto", Dissertacáo de Mestrado em Engenharia Civil, Escola Politécnica, Universidade de Sao Paulo, Sao Paulo

Chatterji S. (1989), "Mechanisms of Alkali Silica reaction and expansion", Proceedings of 8th International Conference on Alkali Aggregate Reaction, Kyoto, Japan, pp. 101 105.

Dent Glasser L. S., Kataoka N. (1981), "Some observations on the rapid chemical test for potentially reactive aggregate", Cement and Concrete Research, Vol. 11, n.° 2, pp. 191 196

Dent Glasser L. S., Kataoka N. (1981), "The chemistry of "alkali aggregate" reaction", Cement and Concrete Research, Vol. 11, n.° 1, pp. 1-9

Etxeberria M. (2004), "Experimental study on microstructure and structural behaviour of recycled aggregate concrete", Tese de Doctorado en Ingeniería Civil, Universidad Politécnica da Cataluña, Barcelona

Evangelista L. (2007), "Desempenho de betóes executados com agregados finos reciclados de betáo", Dissertacáo de Mestrado em Construcáo, Instituto Superior Técnico, Universidade Técnica de Lisboa, Lisboa

Fernandes M. I. G. (2005), "Caracterizacáo petrográfica, química e física de agregados graníticos em betóes. Estudo de casos de obra", Tese de Doutoramento em Geología, Faculdade de Ciencias da Universidade do Porto, Porto

Gress D. L., Kozikowski R. L. (2000), "Accelerated ASR testing of concrete prisms incor porating recycled concrete aggregate", Proceedings of the 11th International Conference on Alkali Aggregate Reaction in Concrete, Quebec City, QC, Canada, June, pp. 1139 1148

HansenT C. (1992), "Recycling of demolished concrete and masonry", RILEM Report 6, London & FNSPON imprint of Chapman & Hall

Hoobs D. W. (1988), "Alkali silica reaction in concrete", Thomas Telford Ltd., London, 183 p.

Kozikowski jr. R.L. (2000), "Accelerated ASR testing of concrete using recycled aggregates", Master's Thesis, University of New Hampshire

Levy S. M. (2001), "Contribuicáo ao estudo da durabilidade de concretos, produzidos com residuos de concreto e alvenaria", Tese de Doutoramento em Engenharia Civil, Escola Politécnica da Universidade de Sao Paulo, Sao Paulo, Dezembro

Li Xinghe (2005), "Mitigating alkali silica reactions in recycled concrete", Doctoral Thesis in Civil Engineering, University of New Hampshire, December

LNEC E 461 (2006), "Betóes. Metodologías para prevenir reaccóes expansivas internas", Laboratorio Nacional de Engenharia Civil, Lisboa, 6 p.

Matías D., Brito J. de (2005), "Influencia dos adjuvantes no desempenho de betoes com agregados grossos reciclados de betáo", Relatório ICIST DTC n.° 3/05, Instituto Superior Técnico, Lisboa

Miranda L.F.R. (2000), "Estudo de factores que influem na fissuracáo de revestimentos de argamassas com entulho reciclado", Dissertacáo de Mestrado, Escola Politécnica, Universidade de Sao Paulo, Sao Paulo

Pecchio M., KiharaY., Battagin A. F. yAndradeT. (2006), "Produtos da reaccáo álcalis silicato em concretos de edificacóes da regiáo do Grande Recife - PE", II Simposio sobre Reaccáo Álcalis agregado em Estruturas de Concreto, RAA 2006, IBRANCON, 10 p.

Prince W., Perami R. (1993), "Mise en evidence du role essential des ions OH dans les ré actions álcalis sílice", Cement and Concrete Research, Vol. 23(5), pp. 1121 1129

Santos J. R., Brito J. de, Goncalves A. P. y Branco E A. (2002),"Campanha experimental efectuada no 1ST", Seminario "Betáo para a Construcáo Sustentável", APEB, Lisboa

Santos Silva A. (2006), "Degradacáo do betáo por reaccóes álcalis silica Utilizacáo de cinzas volantes e metacaulino para a sua prevencáo", Tese de Doutoramento em Engenharia Civil pela Escola de Engenharia da Universidade do Minho, Teses e Programas de Investigacáo, Laboratorio Nacional de Engenharia Civil, 339 p., Lisboa

Santos Silva A. (2007), "Reaccóes expansivas no betáo", Seminario "Em ambiente marinho", Funchal

Scott H. C. (2006), "Mitigating alkali silicate reaction in recycled concrete", Master Thesis in Civil Engineering, University of New Hampshire, March

Shehata M., Christidis C, Mikhaiel W., Rogers C. y Lachemi M. (2008), "Re activity of reclaimed concrete aggregate produced from concrete affected by alkali silica reaction", Proceedings of 13th International Conference on Alkali Aggregate Reaction in Concrete, Trond heim, Norway, pp. 1230 1239

Stievenard Gireaud D. (1987), "Etude de la reaction álcalis sílice dans des bétons", Labo ratoire Central des Ponts et Chaussées, Rapport de recherche LPC n.° 144, Paris, 104 p.

TR 3 Technical Report 3 (2003), "Alkali silica reaction Minimizing the risk of damage to con crete, guidance notes and recommended practice", Cement and Concrete Association of New Zea land, 84 p.