Skip to contentUnited States Department of Transportation - Federal Highway AdministrationSearch FHWAFeedback

Pavements

CHAPTER 2 ALKALI-SILICA REACTION

2.1 INTRODUCTION

This chapter provides an overview of the fundamentals of ASR in concrete. The basic mechanisms of ASR are summarized, followed by discussions on ASR manifestations and symptoms in field structures, relevant test methods, methods of preventing ASR, and specifications. The use of lithium compounds to control ASR is covered only to a limited extent in this chapter but is discussed in detail in Chapter 3.

Alkali-aggregate reaction (AAR) was first identified as a cause of concrete deterioration more than 60 years ago (Stanton, 1940). Since this initial discovery, cases have been reported throughout the world. Of the two types of AAR, ASR and alkali-carbonate reaction (ACR), ASR is considerably more widespread and is of more significance in the United States. Incidences of ACR are relatively scarce and generally restricted to a few isolated regions; ACR will not be considered in this report. ASR-induced damage has become a widespread durability issue throughout the United States and has been implicated in virtually every State in the Nation.

ASR has been implicated in the deterioration of various types of concrete structures, including dams, pavements, bridges, and other structures. The manifestation of ASR in such structures is discussed later in this chapter. The impact of ASR on field structures is best understood in the context of its underlying mechanisms, which are presented below.

2.2 ALKALI-SILICA REACTION

2.2.1 Essential Components of ASR

The mechanisms governing ASR and expansion are quite complex; moreover, there are several schools of thought on which mechanisms are most important in field structures. This section starts with the well-documented and agreed-upon fundamentals of ASR and continues with a mechanistic discussion of the process.

It is widely accepted that the three essential components necessary for ASR-induced damage in concrete structures (as shown in Figure 1) are: (1) reactive silica (from aggregates); (2) sufficient alkalis (mainly from portland cement, but also from other constituent materials or external sources); and (3) sufficient moisture. Eliminating any one of the above components effectively will prevent damage resulting from ASR, as discussed next.

Figure 1.  Diagram. The Three Necessary Components for ASR-Induced Damage in Concrete. An equilateral triangle shows that the three necessary components are reactive silica, sufficient alkali, and sufficient moisture.

Figure 1. The Three Necessary Components for ASR-Induced Damage in Concrete.

Table 1. Rock Types and Reactive Minerals Susceptible to ASR (After CSA, 2000b).
Rocks Reactive Minerals
Arenite Crisobalite
Argillite Cryptocrystalline (or microcrystalline) quartz opal
Arkose Strained quartz tridymite
Chert Volcanic glass
Flint -
Gneiss -
Granite -
Greywacke -
Hornfels -
Quartz-arenite -
Quartzite -
Sandstone -
Shale -
Silicified carbonate -
Siltstone -
Reactive Silica

The presence of reactive aggregates or another reactive silica source in concrete is necessary for ASR to occur. The term reactive refers to aggregates that tend to breakdown under exposure to the highly alkaline pore solution in concrete and subsequently react with the alkalis (sodium and potassium) to form an expansive ASR gel. More detailed information on the specific mechanisms governing aggregate breakdown and subsequent gel formation is provided later in this section. Based on years of laboratory and field experience, a list of reactive minerals and typical rock types that are susceptible to ASR have been compiled, as summarized in Table 1. It is important to note that not all siliceous aggregates are prone to ASR. The inherent reactivity of aggregates depends on several factors, including aggregate mineralogy, degree of crystallinity, and solubility (of the silica in high-pH concrete pore solution). The rocks and minerals shown in Table 1 represent those that are most prone to ASR, but it does not suggest that these are always prone to ASR or that other rocks or minerals not listed in Table 1 are completely immune from ASR.

Sufficient Alkalis

The presence of sufficient alkalis is another required ingredient for ASR. While portland cement is considered the main contributor of alkalis, under certain conditions, other materials may provide additional alkalis that are available to the reaction. The source of alkalis can be from any of the following:

  • Portland cement.
  • Supplementary cementing materials (e.g., fly ash, slag, silica fume).
  • Aggregates.
  • Chemical admixtures.
  • External sources (e.g., seawater and deicing salts).
  • Wash water (if used).

The quantity of alkalis in portland cement typically is expressed as:

Sodium oxide equivalent in percent by mass Sodium oxide content in percent plus Potassium oxide content, in percent (1)

Where: Na2Oe = total sodium oxide equivalent (or equivalent soda), in percent by mass

Na2O = sodium oxide content, in percent

K2O = potassium oxide content, in percent

Although the percentage of alkalis in portland cement is relatively low (in the range of 0.2 to 1.1 percent) in comparison to other oxides or compounds, the bulk of the alkalis ultimately resides in the pore solution of concrete, and it is the associated hydroxyl (OH-) concentration (necessary to maintain charge balance) that produces the inherent high pH in the pore solution (i.e., 13.2 to 14.0). Based on Stanton's early work (1940), it was proposed for many years that expansion resulting from the ASR reaction is unlikely to occur when the alkali content of the cement is below 0.6 percent Na2Oe. This rule-of-thumb value has been cited in various specification limits and was adopted as part of American Society for Testing and Materials (ASTM) C 150. However, it is now recognized that limiting the alkali content of portland cement is not, by itself, an effective way of preventing ASR-induced damage, because this approach does not control the total alkali content of the concrete mixture. Therefore, limiting the maximum alkali content of concrete is the preferred approach when specifying alkali levels. Nixon and Sims (1992) reported that maximum permissible alkali contents between 2.5 and 4.5 kg/m3 Na2Oe have been specified by various countries and agencies, with the allowable alkali content sometimes varying depending on aggregate reactivity.

Figure 2 illustrates the effects of the alkali content on the expansion of concrete prisms stored over water at 38 °C (this and other tests are described later in this chapter). Using an expansion limit of 0.04 percent, the graph shows that laboratory concrete containing less than 3.0 kg/m3 Na2Oe generally resisted excess expansion, even after 2 years of testing. Although laboratory tests have shown that keeping the total alkali content below 3.0 kg/m3 Na2Oe is an effective method of limiting expansion, field structures have exhibited damage with even lower alkali loadings, especially when alkalis have also been contributed by the aggregates in the mixture or by external sources, such as deicing salts. Thus, when considering imposing a limit on the alkali content for a given concrete mixture, consideration should be given to the aggregate type and reactivity, exposure conditions, and nature of the structure (i.e., design life or relative importance).

Figure 2.  Chart. Effect of Alkali Content on Expansion of Prisms Stored Over Water at 38 degrees Celsius (after Thomas, 2002).  The X-axis is alkali content of concrete, in sodium oxide equivalent.  The Y-axis is expansion at 2 years, in percent.  For alkali content of concrete between 1 and 3 kilograms per cubic meter, expansion at 2 years is below 0.03 percent. As alkali content of concrete exceeds 3 kilograms per cubic meter, the expansion at 2 years exceeds the Canadian Standards Association limit of 0.04 percent.  The expansion continues to increase until leveling off at about 0.25 percent expansion with 5.5 kilograms per cubic meter alkali content.

Figure 2. Effects of Alkali Content on Expansion of Prisms Stored Over Water at 38 °C

(After Thomas, 2002).

There has been significant debate regarding the impact of alkalis from supplementary cementitious materials (SCMs) on ASR. Specifically, different countries and agencies vary in how they treat SCMs when calculating the total alkali content of a given concrete mixture. Although an advanced discussion of this issue is beyond the scope of this report, readers are directed to the comprehensive review of current practice compiled by Nixon and Sims (1992). Some countries ignore the contribution of alkalis from SCMs, whereas others use the available alkali content (ASTM C 311) or a percentage of the total alkali content in calculating the total alkali loading of the mixture. Perhaps the most recent advice regarding the contribution of alkalis from fly ash and slag is included in the latest Canadian Standards Association (CSA) guidelines, where it is assumed that SCMs do not contribute any alkalis to concrete when computing total alkali loading. However, limits are placed on the total alkali content of the SCMs, and replacement levels (by mass of cement) are specified, based on the chemistry of the SCMs. Additional information on the mitigation of ASR with SCMs is provided later in this chapter, including information on relevant test methods.

As previously mentioned, alkalis also can be released from certain aggregates within concrete, thereby increasing the alkali content of the mixture (Thomas, et al., 1992; Stark and Bhatty, 1986, Gillott and Rogers 1994) and probably contributing to the increased expansion resulting from ASR (Durand 2000a). Stark and Bhatty (1986) reported that certain aggregates can release alkalis equivalent to a 10 percent alkali contribution from the portland cement under extreme conditions.

The total alkali content within a given concrete mixture may also be increased by the penetration of alkalis from external sources, such as seawater, ground water (containing sulfates), deicing salts, brackish water and industrial wastes. In addition, Nixon, et al. (1987) demonstrated that seawater (used as part of the batch water) increased the OH- concentration in the pore solution and resulted in higher concrete expansion values.

Sufficient Moisture

Available moisture is important when considering the potential for ASR-induced damage in field structures. Concrete mixtures comprised of highly reactive aggregates and high-alkali cements have shown little or no expansion in certain very dry environments. Likewise, local differences in moisture availability within the same structure have resulted in vastly different performance within that structure. Specifically, portions of the structure exposed to a constant or steady source of moisture (e.g., as a result of poor drainage or poor detailing) have exhibited significant ASR-induced damage, while other portions of the structure that remain essentially dry have shown little or no damage. Therefore, the exposure conditions, in general, and the availability of moisture, specifically, play key roles in the durability of field structures.

Figure 3.  Chart. Effects of Relative Humidity on Expansion Using the ASTM C 1293 Storage Regime (Pedneault, 1996).  The X-axis is relative humidity, in percent, and the Y-axis is expansion at 2 years, in percent.  Using siliceous limestone, potsdam sandstone, spratt limestone, and rhyolitic tuff, all trends are similar, with spratt limestone having the greatest expansion for all relative humidity, and potsdam sandstone having the least. Expansion is relatively constant from 70 to 80 percent relative humidity, in no case exceeding 0.1 percent, and increases somewhat as relative humidity increases to 90 percent. It increases even more in all cases as relative humidity rises to 100 percent, for expansion values of 0.35 percent for potsdam sandstone to 0.55 percent for spratt limestone.

Figure 3. Effects of Relative Humidity on Expansion Using the ASTM C 1293 Storage Regime (Pedneault, 1996).

After the formation of the ASR gel, it is the subsequent imbibing of water that causes expansion within concrete, which ultimately can lead to tensile stresses and cracking. It is generally believed that a minimum relative humidity of 80 percent is required to cause significant expansion as a result of ASR. Data supporting the importance of moisture on expansion are shown in Figure 3, where four different reactive aggregates were stored under different moisture conditions, and the expansion of concrete prisms (similar to ASTM C 1293) was assessed (Pedneault, 1996). In this experiment, concrete that was maintained in an environment with less than 80 percent relative humidity had greatly reduced expansion (e.g., expansion in two of the four mixtures was less than 0.04 percent after 2 years).

Limiting the availability of external moisture in field structures is an effective way of reducing ASR-induced damage; however, it is often not feasible to reduce the moisture content below the critical threshold value (i.e., 80 percent relative humidity). However, any attempt at reducing available moisture, whether through proper detailing and design of drainage or through the use of low-permeability concrete, will improve the long-term durability of concrete.

2.2.2 Mechanisms of ASR

The previous sections described the major contributors to ASR in concrete, namely reactive silica, sufficient alkalis, and sufficient moisture. This section provides more detailed information on the specific mechanisms involved, including the chemical reaction that leads to the breakdown of reactive silica, the formation of the ASR gel, and the proposed mechanisms for subsequent expansion.

The term alkali-silica reaction is somewhat misleading-the reaction itself is actually between the hydroxil (OH-) ions in the pore solution and certain siliceous components of the aggregates. The alkalis, specifically sodium and potassium, do not actually attack the reactive silica. The importance of the alkalis is that their presence in high concentrations in the pore solution results in an equally high concentration of OH- ions (to maintain charge equilibrium). It is this high OH- concentration, and thus high pH, that leads to the initial breakdown of reactive silica components in the aggregates. The alkalis also ultimately contribute to the expansive ASR gel formation, as discussed later in this section.

There is general agreement by researchers regarding the specific chemical process governing the breakdown of reactive silica components by a highly alkaline solution, such as pore water inside concrete. When poorly crystalline hydrous silica is exposed to a highly alkaline solution, there is an acid-base reaction between the OH- ions in solution and the acidic silanol (Si-OH) groups (Dent Glasser and Kataoka, 1981), as shown in equation 2. As additional OH- ions penetrate into the structure, some of the stronger siloxane (Si-O-Si) linkages are also attacked, as shown in equation 3.

°Si-OH + OH- ® °Si-O- + H2O (2)

°Si-O-Si° + 2OH- ® 2 (°Si-O-) + H2O (3)

To maintain charge equilibrium, cations (Na+ and K+) diffuse into the structure to balance the negative charges present on the terminal oxygen atoms. The disruption of the siloxane linkages ultimately weakens the structure. Provided that sufficient amounts of alkali-hydroxides are available, this process continues, producing an alkaline silicate solution.

Figure 4.  Chart. Effects of PH on Dissolution of Amorphous Silica (Tang and Su-fen, 1980).  The X-axis is the PH of the solution, and the Y-axis is the amount of dissolved silica in millimoles per liter.  Dissolved silica is relatively low (less than 10 millimoles per liter) for PH ranging up to approximately 12.5, and then increases rapidly, up to 70 millimoles per liter at a PH of approximately 13.

Figure 4. Effects of pH on Dissolution of Amorphous Silica (Tang and Su-Fen, 1980).

The chemical process described above results in the dissolution of the reactive silica components; the alkalinity of the pore solution and the structure of the silica govern the rate and/or amount of dissolution. Poorly crystalline or amorphous silica (e.g., opal, volcanic glass, cristobalite) is much more prone to ASR than well-crystallized or dense forms of silica (e.g., macrocrystalline quartz). One reason for this difference in behavior is that the solubility of amorphous silica increases significantly with pH, as illustrated in figure 4. Thus, aggregates composed of amorphous or poorly crystalline silica will tend to dissolve more readily in the inherently high-pH pore solution in concrete. Well-crystallized or dense forms of silica, such as quartz, are relatively inaccessible to alkaline hydroxide solution, and dissolution only occurs at the surface, at a very slow rate.

As previously mentioned, there is general acceptance of the chemical reactions governing ASR; however, there are several schools of thought regarding the mechanisms of expansion of ASR gel. It is beyond the scope of this publication to thoroughly examine the various proposed mechanisms; however, a basic overview of these mechanisms is provided.

Hansen (1944) proposed an osmotic theory, in which the cement paste surrounding reactive grains behaves like a semi-permeable membrane through which water (or pore solution) may pass but not the larger complex silicate ions. The water is drawn into the reacting grain, where its chemical potential is lowest. An osmotic pressure cell is formed and increasing hydrostatic pressure is exerted on the cement paste, inevitably leading to cracking of the surrounding mortar.

McGowan and Vivian (1952) disputed the osmotic theory proposed by Hansen on the basis that cracking of the surrounding cement paste membrane because of ASR would relieve hydraulic pressure and prevent further expansion. They proposed an alternative mechanism, in which water is physically absorbed into the alkali-silica gel, resulting in swelling of the gel. Tang (1981) later concurred with this water imbibition and swelling theory.

Other researchers, including Powers and Steinour (1955), proposed a compromise, in which pressures resulting from both osmotic pressure and water absorption may be generated, depending on the nature of the alkali-silicate complex, specifically whether it is a fluid or solid. Despite these theories' differences, the fundamental cause of expansion is essentially the same-the entry of water into a region where the effects of absorption or a solute reduces its free energy (Diamond, 1989).

One aspect involving ASR and expansion that has received renewed interest in recent years is the important role of calcium. Although early proposed mechanisms (Hansen, 1944; McGowan and Vivian, 1952) did not recognize calcium's role in ASR, later studies have identified the presence of calcium in the reactive system as being essential to the expansion process. Diamond (1989) proposed that, in the absence of calcium, silica simply dissolves in alkali-hydroxide solution and does not form alkali-silicate gel.

Further support of the solubility of silica in the absence of calcium is found in the work of Kilgour (1988). She found low-calcium ashes to be partially soluble (losing 20 percent soluble mass in 6 months) when exposed to alkali-hydroxide solution (1 g ash in 100 mL of 0.7M NaOH-KOH) in the absence of calcium. The principal soluble phase was analyzed as silica, which remained in solution (no evidence of gel formation). Under the same conditions, except with addition of CaOH, an increase in mass was observed and attributed to the formation of a reaction product analogous to ASR gel. Thomas, et al. (1991) found that gels that are low in calcium and high in alkali are relatively fluid and readily dispersed into cement paste, whereas gels higher in calcium are more viscous and less able to dissipate when they swell on contact with water.

2.2.3 Symptoms of ASR

This section briefly describes the symptoms of ASR in concrete, beginning first with the impact on the microstructure of concrete and concluding with the manifestation in concrete structures. Figure 5 shows a thin-section cut from concrete affected by ASR, which is viewed under transmitted-light microscopy. The reaction product of ASR gel is shown, as is a crack forming through the aggregate and extending into the surrounding cement paste. The crack itself also is filled with ASR gel in some locations. This type of damage is typical of ASR-induced deterioration at the microstructural level of concrete.

Figure 5.  Enlarged Photo. Thin-Section Cut of ASR-Damaged Concrete, Showing ASR Gel and Typical Crack Pattern (Through Aggregate and into Surrounding Matrix). This photo shows cement paste, a particle of reactive aggregate (in this case, flint), and a ribbon of reaction product (alkali-silica reaction gel) running through both of them.  A crack runs from a point where the gel meets the cement paste and continues parallel to the gel.

Figure 5. Thin-Section Cut of ASR-Damaged Concrete, Showing ASR Gel and Typical Crack Pattern (Through Aggregate and Into Surrounding Matrix).

Figure 6. Photo. ASR-Induced Damage in Unrestrained Concrete Element.  Uniform Expansion in All Directions Results in Classic Map-Cracking.  This photo shows a series of interlocking, jagged-shaped cracks on the surface of a concrete barrier.

Figure 6. ASR-Induced Damage in Unrestrained Concrete Element. Uniform Expansion in All Directions Results in Classic Map-Cracking.

The outward manifestation of ASR distress in actual concrete structures varies, depending on the severity of the attack, exposure conditions, type of structure, amount and direction of restraint (internal or external), and other factors. Perhaps the most important of these factors in determining the physical manifestation of ASR-induced damage in field structures is the role of restraint on subsequent crack patterns. Restraint may originate either from external sources, such as adjacent structural elements or applied loads, or internal sources, such as reinforcing steel (conventional, prestressed, or post-tensioned). Figure 6 shows typical ASR-induced damage in unrestrained concrete, resulting in classic map-cracking. Figure 7 shows similar damage in restrained concrete structures, where cracking tends to align itself in the direction of the main reinforcement (i.e., principal stress direction).

Figure 7.  Photos. ASR-Induced Damage in Restrained Concrete Elements, Including A) Reinforced Concrete Column and B) Prestressed Concrete Girder. Photo A shows a concrete column with several vertical cracks, with smaller cracks branching off diagonally.
Figure 7.  Photos. ASR-Induced Damage in Restrained Concrete Elements, Including A) Reinforced Concrete Column and B) Prestressed Concrete Girder. Photo B shows several horizontal cracks along the length of the girder, with smaller cracks branching off vertically.

Figure 7. ASR-Induced Damage in Restrained Concrete Elements, Including (a) Reinforced Concrete Column, and (b) Prestressed Concrete Girder.

When field structures suffer from excessive expansion because of ASR, significant misalignment (with respect to adjacent elements) may result, as shown in Figure 8. For pavements suffering from ASR, the subsequent expansion can lead to extrusion of joint-sealing material or even joint failure, as shown in Figure 9.

Figure 8.  Photo. Misalignment of Adjacent Sections of a Parapet Wall on a Highway Bridge Due to ASR-Induced Expansion (SHRP-315, 1994). This figure shows the misalignment of adjacent sections of a parapet wall on a highway bridge.  Although there is no scale, the lateral misalignment could be on the order of several centimeters. Both sections contain horizontal and vertical cracks.

Figure 8. Misalignment of Adjacent Sections of a Parapet Wall on a Highway Bridge Due to ASR-Induced Expansion (Strategic Highway Research Program (SHRP)-315, 1991).

Figure 9.  Photo. Extrusion of Joint-Sealing Material Triggered by Excessive Expansion Due to ASR. This photo shows patches of joint-sealing material along either side of the joint. One side of the joint also contains numerous interlocking cracks.

Figure 9. Extrusion of Joint-Sealing Material Triggered by Excessive Expansion Due to ASR.

2.3 LABORATORY TEST METHODS FOR ASSESSING ASR

This section provides an overview of available laboratory test methods to assess ASR of aggregates and to measure the effectiveness of various methods of mitigating ASR in new concrete, including SCMs. Limited information about testing lithium compounds in laboratory mixtures is also provided, but more information is provided in Chapter 3.

Table 2 lists several of the most commonly used standard test methods to assess ASR. The table also provides some basic information on each test, including pros and cons of using the test to predict field performance. Of the laboratory tests described, only two are recommended as suitable tests for assessing ASR: (1) ASTM C 1260 (Accelerated Detection of Potentially Deleterious Expansion of Mortar Bars Due to Alkali-Silica Reaction), and (2) ASTM C 1293 (Test Method for Concrete Aggregates by Determination of Length Change of Concrete Due to Alkali-Silica Reaction).

ASTM C 1260 is based on the method developed by Oberholster and Davies (1986) at the National Building Research Institute in South Africa. The test, often referred to as the accelerated mortar bar test, has been adopted by various countries and agencies, including the United States (ASTM and the American Association of State Highway and Transportation Officials (AASHTO)) and Canada. The test entails casting mortar bars containing the subject aggregate (either coarse or fine), which is processed to a standard gradation. The mortar bars are then removed from their molds after 24 hours and placed in water at room temperature. The temperature of the water is then raised to 80 °C in an oven, and the mortar bars are stored in this condition for the next 24 hours. After removing the bars from the water, they are measured for initial length and then submersed in a 1N NaOH solution at 80 °C, where they are then stored for 14 days. Length-change measurements are made periodically during this storage period. Total expansion at the end of the 14-day soaking period typically is used in specifications, although the expansion limits specified by different agencies vary. For example, the expansion criteria established by ASTM and CSA are:

ASTM C 1260 expansion criteria:
  • < 0.10 percent is considered innocuous.
  • 0.10 to 0.20 percent is considered potentially reactive.
  • > 0.20 percent is considered reactive.
CSA A23.2-25A expansion criteria:
  • > 0.15 percent (0.10 percent for limestones) is considered potentially reactive.
  • < 0.15 percent (0.10 percent for limestones) is considered nonreactive.
Table 2. Available Standard Tests for Assessing Alkali-Silica Reactivity.
Test Method Comments
ASTM C 227: Standard Test Method for Potential Alkali Reactivity of Cement-Aggregate Combinations (Mortar Bar Method)
  • Mortar bar test (aggregate/cement = 2.25), intended to study cement-aggregate combinations.
  • Specimens stored in high-humidity containers at 38 °C.
  • Several reported problems with test, including excessive leaching of alkalis from specimens and its unreliable correlation to field performance.
ASTM C 289: Standard Test Method for Potential Alkali-Silica Reactivity of Aggregates (Chemical Method)
  • Aggregate test in which crushed aggregate is immersed in 1M NaOH solution for 24 hours-solution is then analyzed for amount of dissolved silica and alkalinity.
  • Poor reliability.
  • Problems with test include:
    • Other phases present in aggregate may affect dissolution of silica (Bérubé and Fournier, 1992).
    • Test is overly severe, leading aggregates with good field performance to fail the test.
    • Some reactive phases may be lost during pretest processing.
ASTM C 295: Standard Guide for Petrographic Examination of Aggregates for Concrete
  • Useful evaluation to identify many (but not all) potentially reactive components in aggregates.
  • Reliability of examination depends on experience and skill of individual petrographer.
  • Results should not be used exclusively to accept or reject aggregate source-findings best used in conjunction with other laboratory tests (e.g., ASTM C 1260 and/or ASTM C 1293).
ASTM C 856: Practice for Petrographic Analysis of Hardened Concrete
  • Useful for analyzing concrete (from laboratory or field) and for identifying presence of reactive aggregates or reaction products.
  • Reliability of examination depends on experience and skill of individual petrographer.
  • Essential for relating aggregate reactivity to field performance.
Table 2. Available Standard Tests for Assessing Alkali-Silica Reactivity (cont.).
Test Method Comments
ASTM C441: Standard Test Method for Effectiveness of Mineral Admixtures or Ground Blast-Furnace Slag in Preventing Excessive Expansion of Concrete Due to the Alkali-Silica Reaction
  • Mortar bar test, intended to assess effectiveness of SCMs in reducing ASR expansion.
  • Test uses high-alkali cement and PyrexTM glass.
  • Test not very reliable because of the use of Pyrex glass, which is sensitive to test conditions and contains alkalis that may be released during the test. Test does not correlate well with data from concrete mixtures containing natural aggregates (Bérubé and Duchesne, 1992).
ASTM C 1260: Standard Test Method for Potential Alkali Reactivity of Aggregates (Mortar-Bar Method)
  • Recommended Test
  • Mortar bar test, originally designed to assess aggregate reactivity.
  • Bars are soaked in 1N NaOH solution for 14 days.
  • Accelerated test suitable as screening test, but because of the severity of the test, it should not be used, by itself, to reject a given aggregate. If aggregate is tested using both ASTM C 1260 and ASTM C 1293, the results of ASTM C 1293 should govern.
  • Modified method (ASTM C 1567) has been used with success to evaluate effectiveness of SCMs to control ASR expansion (Thomas, et al., 2005).
  • Effectiveness of the method for testing lithium compounds has also been evaluated by adding lithium to the bar and/or soak solution; limited success to date (chapter 3).
ASTM C 1293: Standard Test Method for Concrete Aggregates by Determination of Length Change of Concrete Due to Alkali-Silica Reaction
  • Recommended Test
  • Concrete prism test, generally regarded as best indicator of field performance, is conducted at high humidity (close to 100 percent) at 38 °C.
  • Uses high-alkali cement (raised to 1.25 percent Na20e), with a cement content of 420 kg/m3.
  • Originally intended as aggregate test (using nonreactive fine aggregate to test reactivity of coarse aggregate, and vice-versa); test can also be used to evaluate potential alkali-reactivity of job combinations of fine and coarse aggregates.
  • Test requires 1 year for completion.
  • Also can be used to test effectiveness of SCMs and lithium compounds, but test is then typically run for 2 years.
  • Widely accepted test method; however, long duration of test is major drawback.

In ASTM C 1260, these expansion limits are proposed in an appendix to the standard test method and are not a mandatory part of the standard, whereas in CSA, the limits form part of a standard practice for assessing reactivity and evaluating the effectiveness of preventive measures (e.g., the use of SCMs) (CSA A23.2-27A). Draft guidelines developed by the AASHTO Lead States Program recommend an expansion limit of 0.10 percent for all aggregates, except metamorphic aggregates, for which an expansion limit of 0.08 percent is recommended (AASHTO, 2000).

ASTM C 1260 is recognized as a very severe test method because of the extreme test conditions, specifically the use of a highly alkaline storage solution and high temperature. Because of this inherent severity, the test has been shown to identify some aggregates as being reactive, even though they have performed well in concrete prism testing (ASTM C 1293) and in field applications (Bérubé and Fournier, 1992). Therefore, an aggregate should not be rejected solely based on ASTM C 1260 results unless the reactivity is confirmed using ASTM C 1293. Whenever data are available from both ASTM C 1260 and ASTM C 1293, the ASTM C 1293 results should govern.

Although ASTM C 1260 was developed initially only to test aggregate reactivity, the test has been found to be a suitable method for assessing the efficacy of SCMs in reducing ASR expansion (Thomas and Innis, 1999; Thomas, et al., 2005), where an expansion limit of 0.10 percent at 14 days is typically used. However, the test is not suitable for assessing cement alkalinity because the highly alkaline soak solution masks any effect of cement alkalinity. In addition, ASTM C 1260 is not suitable, in its present form, for testing lithium compounds because the sodium hydroxide soak solution dominates the beneficial effects of adding lithium to the mortar bars. Researchers have attempted to modify ASTM C 1260 to allow for testing lithium compounds; however, these attempts, described in chapter 3, have not yet yielded a rapid mortar bar test that accurately estimates lithium dosage requirements needed to suppress ASR in concrete.

ASTM C 1293, commonly referred to as the concrete prism test (CPT), is generally considered the most accurate and effective test in predicting the field performance of aggregates. In this test, concrete is cast with a cement content of 420 kg/m3. The cement is required to have an equivalent alkali content between 0.8 percent and 1.0 percent, and additional alkalis (NaOH) are then added to the mixing water to obtain a total alkali content of 1.25 percent (by mass of cement), which equates to a total alkali content in the concrete mixture of 5.25 kg/m3. Concrete prisms are cast, cured for 24 hours at 23 °C, and then stored enclosed over water at 38 °C. Expansion measurements are taken at regular intervals, and when testing plain concrete (without SCMs or chemical admixtures), the test typically is run for 1 year. When testing SCMs or lithium compounds, the test typically is carried out for 2 years. This relatively long period for conducting ASTM C 1293, either 1 or 2 years, has been the major drawback for the test and has somewhat limited its use.

In recent years, more countries and agencies have adopted the CPT as a standard method. An expansion limit of 0.04 percent (at the end of the 1- or 2-year test) is typically specified because this value has been reported to correlate well with the cracking of test prisms. This expansion limit (0.04 percent) is referenced in the appendix to ASTM C 1293. In a few rare cases, ASR has been evident as cracks, extruding gel, and spalls (or popouts) in test prisms that have expansions greater than 0.04 percent (Whiting 1999). Noting the appearance of the test prism is part of the test procedure recommended by ASTM C 1293. As part of the most recent guidance provided by CSA (2000a), the expansion limits for the CPT (CSA A23.2-14A) were delineated further to assess aggregate reactivity as follows:

CSA A23.2-14A expansion criteria:

  • < 0.04 percent is considered nonreactive.
  • 0.04 percent to 0.12 percent is considered moderately reactive.
  • 0.12 percent is considered highly reactive.

ASTM C 1260 and ASTM C 1293 are the two recommended tests for ASR (ASTM C 1293 is generally considered to be more representative of field performance). To supplement these tests, a petrographic evaluation of a given aggregate (ASTM C 295) is also suggested, but not required. A thorough petrographic evaluation, performed by a skilled petrographer, provides useful information about the types and amounts of minerals present in an aggregate, and can be used to identify a wide range of reactive components. However, because of inherent difficulties in identifying all potentially reactive phases within an aggregate, petrographic analysis findings alone should not be used to accept or reject a given aggregate, but rather to supplement the findings of other laboratory evaluations. Field performance histories of aggregates, supplemented with petrographic analysis of field concrete (ASTM C 856) containing the subject aggregate, also can provide useful information when considering future use of selected aggregates in new construction. However, as in the case of petrographic analysis of aggregates (ASTM C 295), field performance evaluations should not be used solely to accept or reject a given aggregate for use in new structures.

2.4 METHODS OF MITIGATING ASR

This section briefly describes common methods of mitigating or preventing ASR in new and existing concrete structures. The main focus is on minimizing ASR expansion in new concrete, with less emphasis on methods of extending the service life of structures already affected by ASR.

2.4.1 Minimizing or Preventing ASR in New Concrete

The most common methods of minimizing the risk of expansion resulting from ASR are discussed next, including:

  • Using nonreactive aggregates.
  • Limiting the alkali content of concrete.
  • Using supplementary cementing materials.
  • Using lithium compounds.
Using Nonreactive Aggregates

Using nonreactive aggregates is certainly a viable method of preventing ASR-induced damage. However, to use this approach, one must have a very high level of confidence that the subject aggregates to be used are, in fact, nonreactive. To confirm nonreactivity, the aggregates must be tested strictly (e.g., using ASTM C 1260 and ASTM C 1293), good quality control ensured, and, preferably, field performance well-documented. If the above conditions are met, such aggregates may be used without special precautions. However, given that these conditions often are not met, and given that some aggregates that were believed to be nonreactive (based on testing methodologies available at the time of construction) have caused damaging ASR expansion in field structures, further precautions should be taken in some situations. Instances that warrant such extra caution, even when using aggregates believed to be nonreactive, include the design of critical structures (e.g., prestigious structures or those with an extended design life) and the construction of structural elements exposed to a very aggressive environment (e.g., structures exposed to seawater or deicing salts, which may provide an external source of alkalis). The use of a suitable SCM is an example of taking special precautions with aggregates presumed to be nonreactive.

Limiting the Alkali Content of Concrete

Limiting the alkali content of concrete mixtures below some threshold value is generally effective in preventing ASR-induced damage, but this approach is not always effective by itself. For example, aggregates that are durable at relatively low alkali contents may become more reactive when exposed to higher alkali contents under field conditions, where unanticipated high concentrations of alkalis may result from exposure to deicing salts, alkali release from aggregates, drying gradients (resulting in alkali migration), and other field effects. For example, Stark (1978) reported increases in soluble alkalis from 1.1 to 3.6 kg/m3 Na2Oe close to the surface of some highway structures. Additional information on limiting the alkali content of concrete is provided in section 2.2.1.

Using Supplementary Cementing Materials

The use of SCMs to control ASR in concrete is the most common mitigation measure used in concrete construction. The benefits of properly using SCMs include not only ASR mitigation, but also improved resistance to other durability problems, including sulfate attack, corrosion of reinforcing steel, and freezing and thawing. The benefits related to ASR mitigation are both physical in nature, specifically by reducing permeability, and chemical, where SCMs affect pore solution alkalinity, alkali binding, and other parameters. This section briefly discusses minimizing the risk for ASR-induced damage by the prudent use of SCMs, including ASTM C 618 class C or F fly ash, ASTM C 989 ground-granulated blast-furnace slag (GGBFS), ASTM C 1240 silica fume, and combinations of SCMs (ternary blends). Other industrial byproducts or natural pozzolans (e.g., rice husk ash, calcined clay, metakaolin, etc.) have also shown effectiveness in controlling expansion caused by ASR. However, their effectiveness should be evaluated through an appropriate testing program. CSA standard practice A23.2-28A is the only current standardized approach for evaluating the effectiveness of such materials.

Fly ash is one of the most commonly used SCMs in the world for several reasons, including the economic and technical benefits. To control ASR, the following issues affect the efficacy of a given fly ash:

  • Dosage of fly ash (typically as a mass replacement of cement).
  • Chemical composition of fly ash (especially lime (CaO) content and Na2Oe content).
  • Reactivity of the aggregate.
  • Alkali content of the concrete (from portland cement).

Perhaps the most important parameter affecting the ability of fly ash to control ASR expansion is the CaO content of the ash (Shehata and Thomas, 2000). Generally, lower lime ashes are more effective than higher lime ashes in controlling ASR, mainly because of the higher alkali-binding capacity of concretes containing lower lime ashes (Shehata, et al., 1999). Specifically, fly ash lower in CaO contains more silica that contributes to a C-S-H structure with a lower calcium to silicon ratio, which imparts a negative surface charge, leading to the absorption of cations, especially alkalis (Glasser, 1992). In addition, low-calcium ashes are more effective in controlling ASR because the alkalis contained in the ash are generally not available to the pore solution (Diamond, 1981), whereas high-calcium ashes tend to have more readily available alkalis. Although higher lime ashes can still be used to combat ASR, significantly higher dosages may be needed, especially when using highly reactive aggregates.

GGBFS is also commonly used to mitigate ASR and is typically used in higher dosages than fly ash, typically 35 to 50 percent (by mass of cement) and, in some cases, in even higher dosages. The specific dosage needed to mitigate ASR in a given concrete mixture depends on the reactivity of the aggregate and the total alkali content of the concrete. Slag, which is high in silica, results in reduced formation of CH and produces C-S-H with a higher silica to calcium ratio (Odler, 2000), thereby increasing the alkali-binding capacity of concrete (similar to low-calcium fly ash); however, only limited research has been performed on this topic (Uchikawa, et al., 1989).

Silica fume has not been used as frequently as fly ash and slag to control ASR. Although the efficacy of silica fume in minimizing the risk of ASR-induced damage depends on the reactivity of the aggregate, it appears to depend more on the total amount of alkalis available within the concrete. Thus, dosing silica fume based on the total alkalis within a given concrete mixture has been proposed in recent years (Thomas and Bleszynski, 2000; CSA, 2004), with the required silica fume dosage ranging between the lower and upper limits shown below (based on required levels of prevention):

Lower limit: Minimum % SF = 2 x (alkali contributed by portland cement)

Upper limit: Minimum % SF = 3 x (alkali contributed by portland cement) Where: % SF is the percentage of silica fume (by total mass of cementitious material) the alkali contributed by portland cement is expressed as kilograms of Na2Oe per cubic meter of concrete

For highly reactive aggregates, the amount of silica fume required to control ASR (e.g., 10 percent) may be in excess of the typical dosage used in concrete construction, making it difficult for use in field applications (mainly because of workability concerns, high water and superplasticizer demand, and shrinkage problems). For such cases, it generally is more effective to use silica fume in conjunction with another SCM, such as fly ash or slag, to reduce the required amount of silica fume and to improve constructability attributes. The use of these combinations of SCMs, known as ternary blends, is discussed next.

Ternary blends of cementitious materials have gained in popularity in recent years, either through the use of a blended cement (type IP or IS) in conjunction with another SCM or through the use of plain cement with two SCMs. The benefits of such combinations may improve the economic situation, workability, early strength development, and durability properties. When considering ternary blends to control ASR, combining two or more SCMs may reduce the quantities that would otherwise be used individually as the combined effects are synergistic in their ability to control ASR. For example, concrete mixtures with relatively low quantities of silica fume (4 to 6 percent), combined with moderate levels of slag (20 to 35 percent) or fly ash (class F or class C), were found to be very effective in controlling the expansion of highly reactive aggregates (Bleszynski, et al., 2000; Shehata and Thomas, 2002), as illustrated in Figure 10.

Using Lithium Compounds

Using lithium compounds, especially LiNO3, is a viable approach to controlling ASR-induced damage. Because Chapter 3 provides a comprehensive review of research performed on using lithium compounds to mitigate ASR, no additional discussion is provided in this section. Chapter 4 provides an overview of selected case histories where lithium compounds have been used for both new and existing concrete structures.

Figure 10.  Chart. Synergistic Effects of Ternary Blends in Controlling ASR Expansion Using ASTM C 1260 (after Bleszynski and others, 2000).  The X-axis in this graph is the percent of blastfurnace slab replacement (replacement of cement), and the Y-axis is the percent of cement replacement with silica fume.  One straight line, sloping downward at approximately 45 degrees, indicates the theoretical effective combination, and a second bent line below it indicates the synergistic effect of ternary blends. Concrete mixtures with relatively low quantities of silica fume (4 to 6 percent) combined with moderate levels of slag (20 to 35 percent) were found to be very effective in controlling the expansion of highly reactive aggregates. All of the observations between the two lines show less than 10 percent expansion. Observations below the synergistic effect of ternary blends line show more than 10 percent expansion.

Figure 10. Synergistic Effects of Ternary Blends in Controlling ASR Expansion Using ASTM C 1260 (After Bleszynski, et al., 2000).

2.4.2 Mitigating ASR in Existing Concrete

When ASR-induced expansion and damage has already manifested itself in a field structure, there are some available techniques that can help extend the service life of the structure, as described in detail in CSA A684-00, Guide to the Evaluation and Management of Concrete Structures Affected by Alkali-Aggregate Reaction. To minimize future damage to such structures:

  • Provide adequate or improved drainage (to minimize availability of moisture).
  • Apply claddings or coatings to further limit ingress of moisture.
  • Treat existing cracks to minimize future expansion (and direct ingress of moisture, deicing salts, etc.).
  • Avoid future use of deicing salts that will increase alkali content within the structure.
  • Restrain or confine expansion of structural element.
  • Chemically suppress ASR using lithium compounds (see Chapter 4for case studies and Chapter 5 for guidelines for mitigation techniques).

2.5 SPECIFICATIONS

There are numerous specifications, recommendations, and guidelines for avoiding ASR in new construction or managing it in existing structures, including national specifications from Canada, Australia, the Republic of South Africa, France, the United Kingdom (UK), and many other European nations. Within the United States, there are guidelines from the American Concrete Institute, a guide specification from the Portland Cement Association, and individual specifications from State highway agencies and other government bodies. A summary of all the above specifications is beyond the scope of this report; however, the overall approach to specifying preventive measures to control ASR in new concrete structures can be placed in two categories: (1) performance-based specifications, and (2) prescriptive specifications.

Performance-based specifications dictate what tests should be conducted on a given aggregate to determine its potential reactivity and usually what tests should be run to assess methods of mitigating aggregate reactivity, if needed. An example is a specification that states that aggregates must pass ASTM C 1260, ASTM C 1293, or both, where the term "passing" would be linked to specific expansion limits (e.g., 0.10 percent in ASTM C 1260 or 0.04 percent in ASTM C 1293). Many State highway agencies and other agencies and organizations use these types of specifications for ASR, although all do not currently specify the use of ASTM C 1260 and ASTM C 1293.

Prescriptive specifications actually dictate what materials (and mixture proportions) must be used to control ASR. For example, specifications of this type limit the alkali content of the cement or concrete mixture and specify minimum required dosages of SCMs or lithium compounds.

Various agencies and organizations use different combinations of performance and prescriptive specifications. Perhaps the most progressive and comprehensive specifications for ASR are those recently adopted by CSA, where the aggregates are categorized first according to their reactivity based on either ASTM C 1293 or C 1260. The risk of deleterious ASR expansion develops in the structure and the prevention level thus required (i.e., W (mild) to Z (exceptional), Table 3) is then established as a function of the aggregate reactivity, exposure condition, and size of the element and required service life of the structure. The guidelines are shown in Table 3, which includes the minimum dosages of SCMs required for specific risk levels. This approach considers key parameters previously described in this chapter, such as the CaO content of fly ash and the total alkali content of the mixture when dosing silica fume. The approach also allows for the use of natural pozzolans and ternary blends. Although lithium compounds are not included in the current CSA guidelines, it is expected that the next version of the specifications will allow for the use of lithium salts and will provide specific guidelines on prescribed dosages.

Table 3. CSA Guidelines for Controlling ASR in New Concrete (CSA, 2000a).
Type of SCM Total Alkali Content of SCM (% Na2Oe)* Chemical Composition Requirement (% oxides) Cement Replacement Level (% by mass)a
Prevention Level W (mild) Prevention Level X (moderate) Prevention Levels Y and Z (strong-Y) (exceptional-Z)
Fly Ash < 3.0 CaO < 8% Greater than or equal to 15 Greater than or equal to 20 Greater than or equal to 25
CaO = 8%-20% Greater than or equal to 20 Greater than or equal to 25 Greater than or equal to 30
CaO > 20% See note b See note b See note b
3.0-4.5 CaO < 8% Greater than or equal to 20 Greater than or equal to 25 Greater than or equal to 30
CaO = 8% - 20% Greater than or equal to 25 Greater than or equal to 30 Greater than or equal to 35
CaO > 20% See note b See note b See note b
> 4.5 See note b
Blast Furnace Slag < 1.0b None Greater than or equal to 25 Greater than or equal to 35 Greater than or equal to 50
Silica Fume < 1.0b SiO2 2.0 x alkali contentc 2.5 x alkali contentc 3.0 x alkali contentc
Natural Pozzolans Natural pozzolans that meet the requirements of CSA A23.5 may be used provided that their effectiveness in controlling expansion due to ASR is demonstrated according to CSA Recommended Practice A23.2-28A.
Ternary Blends When two or more SCMs are used together to control ASR, the minimum replacement levels given in Table 5 of CSA, 2004 for the individual SCMs may be reduced partially, provided that the sum of the parts of each SCM is 1. For example, when silica fume and slag are combined, the silica fume level may be reduced to one-third of the minimum silica fume level given in Table 5, provided that the slag level is at least two-thirds of the minimum slag level given in Table 5.

* Na2Oe = sodium oxide content = Na2O + 0.658 * K2O

a To control the total alkali content of the concrete mixture, the maximum alkali content of the cement used in combination with any SCMs should be < 1.0 percent Na2Oe.

b In the presence of reactive or potentially reactive aggregates, blast furnace slag and silica fumes with alkali contents > 1.0 percent Na2Oe, and fly ash with alkali contents > 4.5 percent Na2Oe and/or with CaO contents > 20 percent may be used when their effectiveness in reducing expansion due to ASR is demonstrated in accordance with CSA Recommended Practice. In this respect, test results have indicated that higher alkali fly ashes (but not high CaO ashes), when used in large quantities (e.g., > 50 percent as cement replacement by mass), can significantly reduce expansion due to ASR.

c The minimum level of silica fume (as a percentage of material content) is calculated on the basis of the alkali content of the concrete (expressed as kg/m3 Na2Oe), but in cases where silica fume is the only SCM to be used, the silica fume content should be 7.0 percent by mass.

2.6 CONCLUSIONS

This chapter provided an overview of ASR in concrete, including basic mechanisms, relevant test methods, mitigation methods, and specifications. It provides the necessary background and technical basis for better understanding the role of lithium compounds in controlling ASR, as described in the remainder of this report.


<< PreviousContentsNext >>

More Information

Contact

Fred Faridazar
Office of Research, Development and Technology
202-493-3076
E-mail Fred

 
 
Updated: 04/07/2011
 

FHWA
United States Department of Transportation - Federal Highway Administration