5 AM E

N CONC CA RI

EN

T ASS

I OC

CONCRETE
INFORMATION

5 TE RE

NATIONAL AGGREGATES ASSOCIATION

NATIONAL READY MIXED CONCRETE ASSOCIATION

Diagnosis and Control of Alkali-Aggregate Reactions in Concrete

by James A. Farny and Steven H. Kosmatka*

Introduction
Aggregates containing certain constituents can react with alkali hydroxides in concrete. The reactivity is potentially harmful only when it produces significant expansion.(1)** This alkali-aggregate reactivity (AAR) has two formsalkali-silica reaction (ASR) and alkali-carbonate reaction (ACR, sometimes called alkali-carbonate rock reaction, or ACRR). ASR is of more concern than ACR because the occurrence of aggregates containing reactive silica minerals is more common. Alkali-reactive carbonate aggregates have a specific composition that is not very common. Alkali-silica reactivity has been recognized as a potential source of distress in concrete since the late 1930s.(2,3) Even though potentially reactive aggregates exist throughout North America, ASR distress in structural concrete is not common. There are a number of reasons for this: Most aggregates are chemically stable in hydrauliccement concrete. Aggregates with good service records are abundant in many areas. The concrete in service is dry enough to inhibit ASR. The use of certain pozzolans or slags controls ASR. In many concrete mixtures, the alkali content of the concrete is low enough to control harmful ASR. Some forms of ASR do not produce significant deleterious expansion. To reduce ASR potential requires understanding the ASR mechanism; properly using tests to identify potentially reactive aggregates; and, if needed, taking steps to minimize the potential for expansion and related cracking. Alkali-carbonate reaction in concrete was not documented until 1957. Although ACR is much less common, this report also briefly reviews the mechanism, visual distress symptoms, identification tests, and control measures. interconnected microscopic pores through which water or ions in solution can migrate. The pore water in concrete is an alkaline solution; the measure of alkalinity is pH. The alkali-silica reaction forms a gel that swells as it draws water from the surrounding cement paste. Reaction products from ASR have a great affinity for moisture. In absorbing water, these gels can induce pressure, expansion, and cracking of the aggregate and surrounding paste. The reaction can be visualized as a two-step process: 1. Alkali + reactive silicagel reaction product 2. Gel reaction product + moistureexpansion Gel reaction product is the name given to the alkalisilica gel formed in the first part of the two-step process. The presence of gel doesnt always coincide with distress, and thus, gel presence does not necessarily indicate destructive ASR.
*Concrete Technologist and Director, respectively, Construction Information Services, Portland Cement Association. **Superscript numbers in parentheses designate references. pH is approximately calculated from the concentration of hydroxide ions (OH-) in the solution. The pH of concrete pore solution is at least 12.5 at which point the solution is in equilibrium with calcium hydroxide (Ca(OH)2); an abundant hydration product present in concrete. The pH of the concrete pore solution increases as the alkali content of the cement increases. Alkalies, sodium (Na) and potassium (K), in concrete are primarily derived from cementitious materials. The alkali salts dissolve in the pore solution and increase its alkalinity. Hydroxide, alkali, and calcium ions react with silica in the aggregate particle to form a gel. Reactive silica dissolves in the high-pH solution. The dissolved silica reacts with alkalies and calcium to form a calcium-alkali-silicate-hydrate gel. The gel has a tendency to swell by absorbing water from the surrounding paste. The swelling gel forms the initial cracks in the aggregate and cement paste. Microcracks form near the reaction sites which propagate and join to form large cracks and an overall expansion of the concrete. ASR gels of certain characteristic composition and viscosity have swelling properties. The characteristic composition has not been accurately established. However, initially gel containing less calcium will swell to a greater extent. As the gel moves through concrete, it picks up more calcium, which reduces its swelling potential.

Alkali-Silica Reaction
MECHANISM OF ASR
Concrete consists of aggregatesstone or gravel and sand, in a matrix of cement paste. The cement paste contains
ISBN 0-89312-146-0 Portland Cement Association 1997

TI

PAV

ON

1 IS413

Some gels expand very little or not at all. If a gel is lowswelling, it will not create problems. High-swelling gel may cause pressures exceeding the tensile strength of concrete, which results in cracking of the concrete. Rate of migration of pore fluids to the reaction site and temperature also influence swelling pressures.(4) Consequently, the presence of gel must be linked to destructive cracking for a positive identification of harmfully expansive ASR.

optically strained or microcrystalline quartz more than 5.0% (as found in granites, granite gneiss, graywackes, argillites, phyllites, siltstones, and some natural sands and gravels); natural volcanic glassesmore than 3.0%.

See also ASTM C 294 and C 295, and Table 1 of this document. High-alkali-content pore solution. Alkali hydroxides in solution will react readily with reactive forms of silica in aggregate. As the aggregate reactivity increases, gel reaction products can be formed with lesser concentrations of alkali. That is why use of low-alkali cements alone may not be sufficient to control ASR with highly reactive aggregates containing minerals such as opal. As the pH, or alkalinity, of the pore solution increases, potential for the alkali-silica reaction increases. At higher concentrations of alkali hydroxides, even the more stable forms of silica are susceptible to attack.(5) If the alkali concentration is great enough, the alkali hydroxides break stronger silicon bonds found in less reactive aggregates to form the gel reaction product. This explains why aggregates thought to be nonreactive sometimes exhibit ASR. Repeated cycles of wetting and drying can create high localized concentrations of alkalies. As moisture travels through concrete, dissolved alkalies move in solution, remaining when the moisture evaporates from the concrete surface. This process, known as alkali migration, can cause high alkali concentrations at an evaporative surface even when the overall concrete alkali content is low. Sufficient moisture. Moisture allows migration of alkali ions to reaction sites and the resulting gel absorbs moisture, leading to expansion. For this reason, deleterious ASR does not occur in concretes that are dry in service. Research has shown that expansive ASR can occur in concrete having a relative humidity above 80%.(9) It is possible for well*Reactive minerals: Silica, SiO2, can exist in a variety of textures and crystalline structures. Forms of silica are generally related to the rate at which volcanic magma cooled during formation of the rock. Forms of siliceous minerals in aggregates range from amorphous or glassy (non-crystalline) to cryptocrystalline, microcrystalline, and crystalline, listed in order of decreasing cooling rate. During the formation of quartz crystals, some strain may be introduced. This strain can be seen under a microscope under polarized light. Aggregate containing strained quartz tends to be reactive. Cristobalite and tridymite are crystal forms of silica that exist at higher temperatures and are frozen as such due to rapid cooling. These crystal forms are unstable (metastable) at normal temperatures and rocks containing them are reactive. Opal is an amorphous form of silica with a variable amount of water in its structure. It is a very reactive form of silica. ASTM C 294 contains a description of silica minerals. The reactivity of silica is related to the degree of order in the crystal structure. The forms vary from very reactive silica glass or opal to non-reactive unstrained quartz. One of the convenient ways of classifying silica minerals is as follows: metastable group: opal; tridymite; cristobalite; intermediate glass (52%-66% SiO2); acid glass (more than 66% SiO2) quartz: cryptocrystalline quartz; microcrystalline quartz; chalcedony group which includes chalcedony, agate, chert, flint, and jasper; and optically strained or fractured quartz. Aggregates containing silica minerals in the metastable group typically react more rapidly than those in the quartz group.

Factors Affecting ASR

For alkali-silica reaction to occur, three conditions must be present: reactive forms of silica in the aggregate, high-alkali (pH) pore solution, sufficient moisture.

The amount of gel formed in the concrete depends on the amount and type of silica, and alkali hydroxide concentration. Natural aggregates contain various forms of silica minerals, which have varying reactivitiesa measure of the readiness of the silica to react with alkali. Alkali (sodium and potassium) can come from the cement, pozzolans, aggregates, admixtures, and mix water. When the alkali and silica react, they form the gel reaction product. Exact composition will vary, but the gel always contains alkali, calcium, silica, and water.(5) Reactive silica in the aggregate. Reactivity is a function of the type and form of constituents comprising the aggregate.* Silica minerals in aggregates are generally stable if crystalline and generally reactive if amorphous, but there are exceptions. For instance, there are a few common crystalline forms of silicon dioxide: quartz, tridymite, and cristobalite. Quartz, unless it is microcrystalline or highly strained, is stable. Tridymite and cristobalite are crystalline also, but are low density, porous materials, and are susceptible to attack from alkali hydroxides. An aggregate that presents a large surface area for reactionpoorly crystalline, many lattice defects, amorphous, glassy, microporousis susceptible to reaction.(6) The constituent minerals of an aggregate are obtained from a petrographic analysis. The following list of rock types contain critical amounts of potentially reactive forms of silica: chert and flint containing chalcedony; acidic and intermediate volcanic rocks, such as rhyolite, dacite, latite, and andesite, and the associated porphyries and tuffs; shale and slate; sandstone, siltstone, and quartzite; siliceous carbonate rocks; graywackes; argillites; phyllites; granites and grano-diorites; granite and grano-diorite gneisses. The list is not all-inclusive and many aggregates listed will perform adequately in concrete which contains more than enough alkali to promote ASR. Fine and coarse aggregate containing more than the following quantities of constituents are considered potentially reactive (adapted from Reference 7): opalmore than 0.5% by mass; chert or chalcedonymore than 3.0%; tridymite or cristobalitemore than 1.0%;

0 20 40
Depth into pavement, mm
B

Pavement wearing surface

0
H J F

60 80
B

HB J F

3
California

Depth into pavement, in.

cured concrete in arid regions to have a relative humidity at or above 80% just beneath its surface, even after many decades (see Fig. 1). Any reduction in permeability, by using a low water-cement ratio, supplementary cementing materials or other means, reduces movement of moisture and alkalies into and within the concrete. Stark found that lower water-cement ratio (0.35) concretes expanded significantly less than higher water-cement ratio concretes at ages up to 19 months (see Fig. 2).(10) Concrete alkali content. The potential for ASR increases as the alkali content of concrete increases. A concept used in Europe and Canada addresses the total alkali loading in concrete. Using field performance as a guide, countries have established alkali limits in concrete to control ASR. The approach is most applicable with concretes using portland cement as the sole cementing material. However, it can also be applicable, with refinement, to concrete containing supplementary cementing materials. Ideally, the concept of total alkali loading should include the alkalies from all of the concrete ingredients. However, commonly only the alkalies from the cement or cementitious materials are included in the determination because alkali contribution from other ingredients is usually small. Therefore, total alkali content of concrete is calculated as follows: (kg cement/m 3 ) x (% Na 2O equivalent* in cement)/100 = kg alkali/m3 (lb cement per cu yd) x (% Na2O equivalent in cement)/100 = lb alkali per cu yd

100 120 140 160

Georgia HB F J

New Mexico South Dakota

6
BH F

0 10 20 30 40 50 60 70 80 90 100 Relative humidity, percent

Fig. 1. Relative humidity versus depth of concrete for pavement in various climates. Reference 24. (PCA Slide #K20295)

0.9 0.8 B 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.7 0.6 0.5 0.4 Water-cement ratio 0.3 B Cement alkali: 1.0% 75x75x225 mm (3x3x11-1/4 in.) concrete prisms B B

Note: If supplementary cementitious materials are present, a portion of their alkalies may be added to this equation. See the section on limiting concrete alkalies.
When potential for ASR exists, the accepted allowable limit of alkali content used in Canada and some European countries is 3 kg/m3 (5 lb per cu yd) of concrete.(16,48) Figs. 3A and 3B show the total alkali content of concrete for various cement alkali levels and cement contents, along with the European limit. In the United States, the method used to control the concrete alkali content is to specify a low-alkali cement. Use of low-alkali cement is an accepted and generally successful method for reducing expansion caused by ASR. Low-alkali cement is defined in ASTM C 150 as having an equivalent sodium oxide content of no more than 0.60%. However, concrete made with low-alkali cement can still exhibit expansive ASR if moisture movement concentrates the alkalies in one location,(17) if the aggregate is extremely reactive, if alkalies are provided by certain mineral and chemical admixtures, as well as from the aggregates and mixing water, or if total alkali content of concrete is high due to a high cement content.
*The total (acid soluble) alkali content of portland cement includes both sodium oxide (Na2O) and potassium oxide (K2O), but is conventionally expressed as equivalent sodium oxide as follows: Na2O eq = Na2O + 0.658(K2O). Sodium and potassium oxides are determined as part of chemical tests required by ASTM C 150.

Fig 2. Nineteen-month expansions for different water-cement ratio concrete mixtures. Reference 10. (#K20296)

Expansion at 19 months, percent

7
Alkalies, kilograms per cubic meter of concrete

Alkalies, pounds per cubic yard of concrete

Cement Content
6 5 4 3 2 300 kg/cu m 1 250 kg/cu m 0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 3 kg/cu m 400 kg/cu m

12

Cement Content
700 lb/ cu yd

10

350 kg/cu m

600 lb/cu yd

6 5 lb/cu yd 4 500 lb/cu yd 2 400 lb/cu yd

Cement alkalies (Na2O equivalent), percent Fig. 3A. Alkalies (as Na2O equivalent) per cubic meter of concrete as a function of cement alkali and cement content. (#K20297) External alkalies. External alkalies may increase expansion caused by ASR, especially when concrete is cracked or is highly permeable.(11) Common sources of external alkalies are deicing salts, seawater, groundwater, and water from industrial processes. In particular, use of deicers has increased significantly in recent years. Sodium chloride deicing salt solutions and seawater can provide unlimited amounts of alkali.(12) Immersing concrete prisms containing reactive aggregates in a sodium chloride solution has demonstrated increases in expansion and deterioration of the concrete, especially at elevated temperatures.(13,14) There are ways to reduce the ingress of external alkalies. Use of supplementary cementitious materials and a low water-cementitious materials ratio will reduce concrete permeability, slow the entrance of external alkalies, and reduce potential ASR expansion. Protective coatings and sealers provide a barrier to seawater, deicing salts, and other alkali sources. In some cases, regular cleaning of the structure might be worthwhile so that unwanted salts are carried away before they have a chance to enter the concrete and contribute to reaction. Unfortunately some of these are not practical solutions for structures such as pavements. Wetting and drying. Dry exposures reduce potential for expansive cracking due to alkali-silica reactivity. Indoor concrete is usually drier than exterior concrete. However, concrete that has high initial water content may maintain a high internal relative humidity if not permitted to dry. High humidities can sustain the ASR reaction. Its unknown 4

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Cement alkalies (Na2O equivalent), percent

Fig. 3B. Alkalies (as Na2O equivalent) per cubic yard of concrete as a function of cement alkali and cement content. (#K20298) whether continuous saturation or cycles of wetting and drying cause more damage.(15) It is known, however, that alkali migration can occur with alternate wetting and drying, concentrating alkalies near the drying zone and inducing reaction there.(17) It is desirable to minimize both available moisture and wet-dry cycles by providing good drainage. Temperature. Structures in warmer exposures are more susceptible to ASR than those in colder exposures because the ASR rate usually increases with increasing temperature.(17) For the majority of aggregates, higher temperatures also mean larger ultimate expansions. However, there are studies showing that lower temperatures, 13C and 20C (55F and 68F) compared to 38C (100F), resulted in significantly larger ultimate expansions with certain aggregates.(46,47) The effect of high or low temperatures on ultimate expansion is aggregate-dependent, with most aggregates reacting more at higher temperatures.

VISUAL SYMPTOMS OF EXPANSIVE ASR

Harmful ASR expansion does not occur without reaction products. But reaction products can occur without harmful ASR expansion. Since ASR products have been observed in good quality undamaged concrete and the presence of reaction products does not necessarily indicate that destructive ASR is occurring, a cause-and-effect relationship must link the presence of reaction products to harmful expansion.

The British Cement Association outlines a procedure for diagnosing ASR and assessing its impact on a structure. The procedure includes the following steps: (1) site inspection and testing, (2) sampling, (3) laboratory investigation, (4) evaluation, and (5) risk assessment of future reaction.(15) Visual observation of the structure is a major component of this program.

Expansion
Typical indicators of ASR presence are map (random pattern) cracking, and in advanced cases, closed joints, spalled concrete surfaces, or relative displacements of different portions of a structure (see Fig. 4). Because ASR deterioration is slow, the risk of catastrophic failure is low. ASR can cause serviceability problems and can exacerbate other deterioration mechanisms such as those that occur in freeze-thaw, deicer, or sulfate exposures. For instance, a concrete pavement might experience map cracking, and with subsequent freeze-thaw damage, begin to break apart. Likewise, cracking from other mechanisms can allow an ingress of alkalies and exacerbate ASR. Some of the more serious concerns regarding in-service concrete expansion relate to hydro-electric dams. High-speed rotating equipment requires strict tolerances be maintained between the machinery and the concrete to which it is anchored.(18,49)

Fig. 4. This parapet wall has been severely affected by ASR. Notice the cracking, joint closing, spalling of concrete surfaces, and lateral offset that have developed. Reference 21. (#56586)

Cracking
Concrete deleteriously affected by expansive ASR is characterized by a network of cracks (see Fig. 5). A visual inspection should note the location, length, width, apparent depth, and continuity of cracks, and whether the cracks go through or around the aggregate. Any other associations with stress directions, reinforcement, restraint conditions, or discolorations should also be made in order to accurately describe the cracking. Reference 21 has numerous illustrations of ASR in highway structures. Wide cracks are easy to see. Fine cracks arent always visible, but they may be easier to see on a wet concrete surface that is beginning to dry. Drying occurs unevenly and provides a contrast that makes fine cracks more prominent (see Fig. 6). For this reason, some inspectors prefer to examine a concrete structure as it is drying, such as after a rainfall. Expansive ASR begins with the formation of gel either in or on a reactive aggregate particle within the concrete. As the gel absorbs water, it can exert a fairly uniform pressure up to 10 MPa (1450 psi) or more in all directions.(19) This pressure exceeds the tensile strength of conventional concretes, which is generally about 10% of compressive strength. The concrete cracks in a 3- or 4-pronged star pattern. This cracking is usually enough to relieve the pressure and accommodate the resulting volume increase (see Fig. 7).(19) As more particles react, cracks radiating from these stars join with others to form a pattern resembling a map (see Fig. 8).(6) ASR is not the only mechanism to cause map cracking; cycles of freeze-thaw or other mechanisms causing shrinking or swelling of the concrete mass can cause similar patterns.

Fig. 5. View of well-developed map-cracking in a pavement, a typical pattern associated with ASR. The prominent cracks are in a longitudinal direction as there is less restraint to lateral expansion. Reference 21. (#58352) Pavement and slabs on grade. In pavements and slabs on grade, cracking from expansive ASR often begins near free edges and joints where moisture is abundant.(19) The ASR cracks are usually perpendicular to transverse joints, and parallel to free edges along the roadside, and against asphalt pavement, where there is less restraint. These cracks often progress to a map pattern. Continuously reinforced pavements will typically have ASR cracks parallel to the reinforcement. Traffic loads aggravate crack formation. Fig. 9 illustrates the progression of ASR cracking in concrete pavements.(12) Though this model is developed specifically for pavements, the sequence of events is simi5

Fig. 8. When three cracks are caused by a reacted particle, they will join other cracks of nearby reacted particles. This leads eventually to a pattern of cracks resembling lines on a map. Reference 21. (#58342) lar in other structures. These cracks should be differentiated from D-cracks (typically associated with freeze-thaw distress). Freeze-thaw damage usually results in cracks parallel to transverse joints and free edges. Reference 21 illustrates the difference between cracking caused by ASR versus freeze-thaw damage in pavements. In the first stage, moisture is lost at the top surface of the slab. This results in a slight drying shrinkage and some very fine cracking at the concrete surface. At this point, there is no appreciable alkali-silica reaction or expansion. Concrete can remain in this stage indefinitely. Stage two might begin only months after concrete placement or many years later. Reactive aggregates and highpH pore solutions speed progress to stage two, the onset of which is marked by gel formation and subsequent swelling. Usually, gel forms in cracks inside the aggregate particle, but sometimes it forms on the rim of the particle. Formation of the gel may represent an initial reduction in volume, but as the gel absorbs moisture, it swells and exerts a force on the surrounding concrete. The unrestrained concrete surface has no way of resisting the swelling, so the surface cracks begin separating. Widening cracks are an indication that ASR is occurring. The widened cracks then allow easier moisture access to the concrete interior, sustaining additional gel formation and swelling. During this phase, gel may exude from the widening cracks. In stage three, continued drying slows reaction in the zone near the surface. But reaction continues in the moist interior concrete, contributing to formation and expansion of gel, exerting pressure on the surface, and widening the surface cracks. ASR continues until the silica is depleted, until the alkali ion concentration or pH is sufficiently reduced, or until sufficient drying occurs to stem the formation and swelling of gel. The process represented by this three-step model can be temporarily or indefinitely interrupted, for instance, during periods of dry weather. However, if conditions again become conducive to ASR, the reaction will resume.

Fig. 6. Cracks can be more easily seen when the concrete has been wetted and is starting to dry. Cracks are highlighted with wetness as drying progresses. Reference 21. (#58356)

Fig. 7. ASR often induces three or more cracks at each reacted aggregate particle location to relieve pressure caused by expansive forces. Reference 19. (#K20299) 6

Relative Humidity, % 100 50

Stage one
Drying surface Slight drying shrinkage

100

50

Stage two

Expanding moist interior

100

50

Stage three

Expanding moist interior

Fig. 9. Model for ASR cracking in unrestrained slabs on ground. Drying shrinkage occurs in Stage One, reaction and expansion of the interior concrete in Stage Two, and continued drying at the surface accompanied by continued reaction in the interior during Stage Three. Relative humidity vs. depth is shown also. Reference 12. (#K20300) Other structures. Observed cracking is usually most strongly developed in areas of structures where the concrete has a constantly renewable supply of moisture, such as close to the waterline in piers, from the ground behind retaining walls, beneath pavement slabs, or by wick action in piers or columns.(20) In the absence of directional restraint, concrete cracks in a random pattern when ASR damage occurs. If the concrete is restrained, as most concrete structures are, cracks will be oriented along the stress direction:(6) for example, vertical cracks form in piers (Fig. 10). Since reinforcement parallels the major stress direction, linear ASR cracks occur roughly parallel to the steel bars. Unlike steel corrosion cracks that appear directly over the bars, linear ASR cracks commonly appear between the bars. A roughly rectangular cracking pattern results when the reinforcement is configured fairly evenly in two directions.(19) In unreinforced concrete, external restraint influences the orientation of major cracks. In gravity dams, expansion is less restrained in an upward direction and cracking is thus predominantly horizontal.

Fig. 10. Cracking associated with stress directions. Predominant cracks are oriented longitudinally in this column. Reference 21. (#58337)

Surface Deposits (Efflorescence)

Deposits of ASR gel or calcium carbonate can be found along cracks in concrete, leaving a deposit on the surface ranging in color from white to dark gray. These deposits are sometimes called efflorescence or exudations. The material exuding from the cracks can be white, yellowish, or colorless, and viscous, fluid, waxy, rubbery, or hard. Surface deposits may or may not accompany expansive ASR. However, their presence is not indicative of ASR as other mechanisms (such as frost action) or the transmission of water through concrete can also cause efflorescence (without the presence of ASR gel). It is good investigative practice during a site survey, however, to record the extent and location of surface deposits along with their color, texture, dampness, and hardness. A chemical analysis is also helpful to determine if ASR gel is present in the deposit.

Popouts
A popout is caused by a fragment breaking out of the surface of the concrete, leaving a hole that may vary in size, but is usually 25 to 50 mm (1 to 2 in.) wide. Popouts caused 7

by sand-sized particles can be much smaller (see Fig. 11). A fractured aggregate particle can be found at the bottom of the hole. The number, size, and location of popouts provide valuable information about the quality of aggregates in a concrete. A popout caused by expansive ASR is formed to relieve pressure created by gel just beneath the concrete surface. Locating gel at the site of a popout is a strong indication of ASR. Floor coverings may play a role in development of ASR popouts at later ages. Especially on slabs placed over wet cohesive soils, condensation can occur under the covering to develop popouts. Examination of the aggregate at the bottom of the resulting pit can usually explain the cause of the popout. Popouts are a cosmetic nuisance and usually do not affect the serviceability or durability of the concrete. Still, there are ways to minimize their occurrence from ASR: Do not use hard steel troweled surfaces where not needed, such as on most exterior slabs. Choose wet curing methods, such as continuous sprinkling with water, fogging, or covering with wet burlap instead of poly films, curing papers, or curing compounds. Flush the surface with water before final drying. Use blended cement or a supplementary cementitious material such as fly ash that has been demonstrated to control ASR.

Fig. 11. Popouts caused by sand sized particles. Inset shows closeup of such a popout. (#51117, #51118)

The presence of ASR-induced popouts is not necessarily an indication that the concrete structure will expand and have map cracking or other signs of ASR distress. Popouts can also be caused without ASR by freezing and thawing of low density porous aggregate at or near the concrete surface. For example, porous chert can absorb water and upon freezing, expand and form a popout.

Color Change
Surface discoloration is common in conjunction with map cracking.(6) Darkened or blotchy areas are often associated with ASR. Areas along cracks may be bleached, pinkish, or brownish in color, extending several millimeters (tenths of an inch) from the crack. Fig. 12. Reacted particles are identified by dark areas that indicate ASR gel (either inside the particle, or along edges of particles), cracks extending from aggregate particles into the surrounding paste, and ASR gel reaction products. (#43090)

METHODS FOR IDENTIFYING ASR DISTRESS

It is important to distinguish between the ASR reaction and damage resulting from the reaction. In the diagnosis of concrete deterioration, it is most likely that a gel product will be identified. In some cases, significant amounts of gel are formed without causing damage to concrete. It is therefore important that in analysis of deteriorated concrete, signs of distress, such as microcracks and separation of aggregate from the paste, be accurately attributed to ASR gel formation if the damage is to be associated with ASR. Other causes of distress should not be precluded.

To pinpoint ASR as the cause of damage, the presence of ASR gel must be verified. A site of expansive reaction can be defined as an aggregate particle that is recognizably reactive or potentially reactive and is at least partially replaced by gel (see Fig. 12). ASR-affected coarse aggregates usually exhibit internal fracturing, with cracks extending into the surrounding concrete matrix. If only fine aggregate is reacting, cracks can form in the matrix without affecting coarse aggregate particles. Gel can be present in cracks and voids and may also be present in the area around aggregate particle edges. A network of internal cracks connecting reacted aggregate particles is a strong indication that ASR is responsible for cracking.

A petrographic examination (ASTM C 856) is the most positive method for identifying ASR distress in concrete. Prepared sections of concrete are examined under a microscope by an experienced petrographer to determine the presence and location of reactive aggregates and gel. Silica gel appears as a darkened area in the aggregate particle or around its edges. Petrography, when used to study a known reacted concrete, can confirm the presence of reaction products, and verify ASR as an underlying cause of deterioration. Another method for detecting alkali-silica gel in concrete is the uranyl-acetate treatment procedure, discussed in the Annex to ASTM C 856. A freshly exposed concrete surface is sprayed with a solution of uranyl acetate, rinsed with water, and then viewed under ultraviolet light.(21,22) Reacted particles and gel appear as bright yellow or green areas. The uranyl-acetate treatment procedure requires experienced technicians for correct interpretation. The test does not differentiate between a harmless presence of gel or reactivity, and that which is detrimental. Not all fluorescence is an indication of ASR gel. For instance, some aggregates fluoresce naturally. Also, uranyl ions can be absorbed on cement hydration products and appear as broad, faint areas of fluorescence. Neither of these conditions is an indication of ASR gel. Furthermore, a positive identification of gel by this technique doesnt necessarily mean that destructive ASR has occurred. The test is ancillary to more definitive petrographic examinations and physical tests for determining concrete expansion. The uranyl-acetate treatment procedure must not be used alone to diagnose ASR.

alkali content of the cement, and the water-cement ratio of the concrete the same or higher than proposed for future use, (2) is the field concrete at least 15 years old, (3) are the exposure conditions for the field concrete at least as severe as those proposed for future use, and (4) were pozzolans or slags used in the field concrete? In addition, the current aggregate supply should be examined petrographically to be sure that it is representative of that used in the field concrete. When field history is not available, laboratory tests can be used to evaluate the potential reactivity of aggregate. Fig. 13 illustrates the process of evaluating an aggregate. Several tests to identify potentially reactive cement-aggregate combinations were developed in the 1940s and 1950s, while newer tests have been developed in the 1990s. Each test has advantages and disadvantages as well as limitations that are outlined below. A description of each test follows. See Table 1 at the end of this document for details of test conditions, required samples, and measurement criteria.

Mortar-Bar Method (ASTM C 227)

The Test Method for Potential Alkali Reactivity of CementAggregate Combinations (Mortar-Bar Method), ASTM C 227, measures the expansion of mortars made with the test aggregate. The aggregate should conform to a standard grading, and may require crushing to meet the grading. Either a job cement or reference cement may be used. Alkali content of the reference cement should be at least 0.6% Na2O eq, and preferably, have the highest alkali content representative of the cement use intended. At least four mortar bars using this aggregate and cement are prepared to standard dimensions: 25x25x285 mm (1x1x111/4 in.). The bars are stored over water at 100% RH at 37.8C (100F). Length measurements are taken periodically beginning at 14 days and extending to 12 months or longer. According to ASTM C 33, the maximum allowable expansion for an aggregate to be considered potentially nonreactive is 0.10% at six months, or 0.05% at three months if longer testing periods arent feasible. Longer test periods are preferred for differentiating the reactivity of an aggregate. One positive aspect of this test is that it measures the ASR susceptibility of cement-aggregate combinations. Unless very reactive materials are being tested, though, obtaining meaningful results may require a year or more.(23) It is frequently impractical to wait a year for test results. Lack of time to run an adequate test is one of the major drawbacks associated with C 227. Even after long test periods, not all deleterious aggregates will exhibit expansive behavior. Sometimes, the C 227 test fails to distinguish between slowly reacting aggregates and innocuous ones because the test conditions arent severe enough or the test would have to be run for several years.(24) A petrographic analysis of the aggregate is helpful in determining if this test is suitable for testing particular slowly reactive aggregates. Problems with the test setup have also been identified. The test method requires the inside of the container to be lined with wicks to maintain a high relative humidity. The wicks cause water to condense on 9

CONTROL OF ASR IN NEW CONCRETE

The best way to avoid ASR is to take appropriate precautions before concrete is placed. Standard concrete specifications may require modification to address ASR. These modifications should be carefully tailored to avoid unnecessarily limiting specifiers options. This requires careful analysis of cementitious materials and aggregates and choosing a control strategy that optimizes effectiveness and economic selection of materials. Because different geographic regions have different needs and materials available, Reference 25 was developed as a guide specification for concrete subject to alkali-silica reactions. It is based on Reference 7 written for a region on the East Coast of the United States. Fig. 13 illustrates the guide specification process in determining if potential reactivity exists and in selecting materials to control ASR.(25)

IDENTIFICATION OF POTENTIALLY REACTIVE AGGREGATE

Field performance history is the best method of evaluating the susceptibility of an aggregate to ASR. When evaluating past field performance, the following should be determined: (1) are the cement content of the concrete, the

Is there an adequate field performance history for the aggregate? YES Based on field history, is the aggregate potentially reactive? NO Will new concrete contain materials similar to those used in the field? YES

NO

YES

NO

Test the aggregate Will the environment be more severe than that in the field history? NO NO Does testing indicate reactive aggregate? YES Petrography ASTM C 295 and Mortar-bar test ASTM C 1260

No special requirements

OPTIONAL YES Concrete-prism test ASTM C1293

NO

Does testing indicate reactive aggregate?

YES

Use special requirements

NO Limit concrete alkalies

Are pozzolans, slags, or blended cements available? YES Prove effectiveness

Option A ASTM C 1260 OR

Option B ASTM C 441

Fig. 13. Flow chart listing steps to be taken to control ASR. Reference 25. (#K20301) 10

the bars thus leaching alkalies from the mortar. Reduced expansions have been oberved in some containers using wicks. Although not a perfect test, it is a useful method for studying the ASR susceptibility of aggregates and cementitious materials.

Chemical Method (ASTM C 289)

The Test Method for Potential Reactivity of Aggregates (Chemical Method), ASTM C 289, commonly called the quick chemical test, estimates the potential reactivity of siliceous aggregate. An aggregate is crushed and sieved to yield three samples of 25 grams each. This material is then reacted with an alkaline solution (1 N sodium hydroxide) at 80C (176F). At 24 hours, the amount of dissolved silica from the aggregate and the reduction in alkalinity of the solution are measured. By plotting these data against known materials, its possible to estimate reaction potential. The aggregate falls into one of three ranges: innocuous, deleterious, or potentially deleterious. ASTM C 289 identifies highly reactive aggregates fairly reliably; however, it fails to identify slowly reactive aggregates. Also, certain aggregates have high amounts of soluble silica present, but produce only small expansions in service. Thus, the test does not always give reliable results.(23,24) The method also is usually not applicable to test carbonate aggregates. This test is a helpful research tool and may be useful for initial screening of aggregate; however, other tests should be relied upon to better define which aggregates are potentially reactive.

such as transmission electron microscope (TEM) and scanning electron microscope (SEM) can also be used to characterize the mineral composition of an aggregate sample. Analytical techniques, like X-ray diffraction and infrared spectroscopy, may be used to characterize the texture and crystalline form of silica in the aggregate particles. A petrographic analysis of aggregates can be time consuming. Since small samples are evaluated, precautions must be taken to ensure the sample is representative of the product source. Correlating the results of a petrographic analysis with service record of the aggregate in concrete is very useful. The results of a petrographic analysis will not reveal whether an aggregate will cause deleterious expansion in concrete; this needs to be evaluated using other test methods.

Rapid Mortar-Bar Test (ASTM C 1260)

Assessing aggregate reactivity by accelerated methods is increasingly common. The shortcomings of ASTM C 227 and C 289 prompted development of a rapid test method for assessing potential reactivity of aggregate. ASTM C 1260, Test Method for Potential Alkali Reactivity of Aggregates (Mortar-Bar Method), offers a relatively rapid test to supplement lengthier test methods. It involves measuring the length change of mortar bars made with the sample aggregate stored in a strongly alkaline solution at an elevated temperature. This test measures potential aggregate reactivity, not reactivity of specific cement-aggregate combinations. A nonmandatory appendix to C 1260 classifies an expansion of 0.10% or less at 14 days as innocuous behavior. Expansions of more than 0.20% are indicative of potentially deleterious expansion. Between 0.10% and 0.20%, aggregates may exhibit either innocuous or deleterious behavior in field performance. Various organizations use different criteria depending on local experience. The expansion level of the aggregate may determine what materials are used in concrete. ASTM C 1260 should be used as a screening test in conjunction with other tests to determine the potential reactivity of an aggregate. Due to the severe test conditions, aggregates with good field performance and no history of ASR can sometimes test to be reactive. This is because aggregates in field concrete are rarely exposed to the severe alkali and temperature conditions of the test method. Thus, aggregates identified as potentially reactive by C 1260 can perform well in practice when exposed to more reasonable alkali levels. ASTM C 1260s severity makes it useful for identifying slowly reacting aggregates, which may not be identified by C 227. Stark(24) has suggested refining the C 1260 test by varying the concentration of the NaOH storage solution to match the pore solution for a particular concrete. Test solution concentration can be correlated to cement or concrete alkali level. Concurrent testing in various concentrations of NaOH solution could then determine safe cement or concrete alkali levels for a particular aggregate. This would provide practical information to help define safe use of available materials. Another variation evaluates the effectiveness of mineral admixtures. This procedure has not been standardized. 11

Petrographic Examination (ASTM C 295)

Mineral composition is a good indicator of potential aggregate reactivity. The Guide for Petrographic Examination of Aggregates for Concrete, ASTM C 295, is used to determine mineral composition and form (page 2 of this document discusses aggregate sources and amounts that are prone to ASR). Petrographic examination should be used as a screening method for aggregates. It gives quick results to help predict possible aggregate reactivity but cant give quantitative information about the aggregates actual behavior in concrete. A petrographic evaluation of an aggregate sample, by an experienced petrographer, will identify potentially reactive minerals in an aggregate sample. The ASTM practice provides guidance on megascopic and microscopic methods of analyzing an aggregate sample. Descriptive nomenclature for minerals is found in ASTM C 294, Constituents of Natural Mineral Aggregates. In a megascopic analysis, the aggregate sample is separated into groups of similar rock types. Potentially reactive rock types are quantified on the basis of the total aggregate sample. A microscopic analysis is a more involved and time-consuming procedure. Hence this analysis will typically use a smaller aggregate sample. A petrographic microscope is most commonly used and thin sections of aggregates are evaluated under polarized light. Refractive index and specific gravity of powdered aggregate samples are also evaluated. More exotic types of electron microscopes,

Concrete Prism Test (ASTM C 1293)

The Standard Test Method for Concrete Aggregates by Determination of Length Change of Concrete Due to AlkaliSilica Reaction, ASTM C 1293, measures the expansion of concrete prisms made with the fine or coarse aggregate in question and a Type I high-alkali (0.90%0.1% Na2O-equivalent) cement plus added NaOH to simulate a 1.25% Na2Oequivalent cement. Concrete proportions are strictly defined, including a water-cement ratio of 0.42 to 0.45. At these proportions, concrete has an alkali content of 5.25 kg/m3 (8.85 lb per cu yd). At least three concrete prisms are prepared to standard dimensions: 75x75x285 mm (3x3x11-1/4 in.). An additional prism can be made for petrographic examination. All prisms are stored in a sealed container over water at 100% RH at 38C (100F). Length change measurements are compared to the reading taken at one day and are calculated to the nearest 0.001%. Measurements are taken at 7, 28, and 56 days, and 3, 6, 9, and 12 months. Additional readings, if required, are at 6-month intervals. Appendix X1 of the standard offers the following interpretation of results: potentially deleterious behavior is indicated if the one-year expansion is greater than or equal to 0.04%. C 1293 is similar to C 227 in that both involve cementaggregate combinations, and both can require a year or more to obtain meaningful results. C 1293 should be used to supplement results obtained from C 227, C 295, C 289, or C 1260 and has the potential to resolve uncertainties about results obtained with C 227 and C 1260. In Canada, the equivalent to ASTM C 1293 is CSA A23.2-14A, Potential Expansivity of Aggregates (Procedure for Length Change Due to Alkali-Aggregate Reaction in Concrete Prisms).

F H J 0.50 B

0.45
Expansion after 14 days, percent

Rhyolitic tuff ASTM C 1260

B J H CSF CSF FA FA FA GBFS

0.40 0.35 0.30 0.25 0.20 0.15

J H F H F F H J H F

0.10 0.05 0 0 10
B J B

20 30 40 50 60 Percent mineral admixture

70

Fig. 14. ASR expansion versus dosage for various mineral admixtures including condensed silica fume (CSF), fly ash (FA), and slag (GBFS). Reference 43. (#K20302) for testing in accordance with ASTM C 1260. To determine expansion characteristics, bars are stored in an alkaline environment at an elevated temperature, 1 normal NaOH at a temperature of 80C 2.0C (176F 3.6F). An expansion of not greater than 0.10% at 14 days is considered acceptable to control ASR for a particular job aggregate. Fig. 14 illustrates the need for optimization of dosage of pozzolan or slag to control ASR. ASTM C 1260 should not be used to test the effectiveness of mixes with portland cement alone (no slag or pozzolan) as the method is insensitive to cement alkalies when using the 1 N sodium hydroxide solution. Reference 51 studies long-term field exposure demonstrating the applicability of this procedure. In Option B, control specimens containing low-alkali cement and test specimens containing the pozzolan, slag, or blended cement are mixed using Pyrex glass as aggregate. Proportions for the test mixtures are set at 900 g of aggregate and a total of 400 g of blended cement or cement plus pozzolan or slag. The same proportions are used in the control mixture with a cement having a Na2O equivalent of less than 0.60%. Specimens are mixed as specified in ASTM C 441. All specimens are stored over warm water in accordance with ASTM C 227. As long as the test mixture expansion does not exceed that of the control mixture, the blended cement or cement plus pozzolan or slag is considered effective in controlling ASR. ASTM C 227 and C 1293 (CSA A 23.2-14A, the concrete prism test) can be used to determine the effectiveness of a mineral admixture or blended cement. The Mid-Atlantic

DESIGNING MIXES TO CONTROL ASR

Most concrete is not affected by ASR and special requirements are not needed. However, if historical experience or the tests discussed above demonstrate that ASR is a potential concern, then concrete mixtures must be specifically designed to control ASR. The concrete mix should be designed with a safe and economic combination of locally available materials. Historical guidance or tests must be used to establish the ASR resistance of selected material combinations. Fig. 13 illustrates what steps should be taken to assist in the mix design. When pozzolans, slags, or blended cements are used to control ASR expansion, their effectiveness must be determined. The limits on expansion referenced in ASTM C 595, C 989, and C 1157 may not be applicable. The tests do not use job aggregate and the limits may be more restrictive than is necessary or achievable. An alternative comparative procedure is in C 311. Testing options outlined in the Guide Specification for Concrete Subject to Alkali-Silica Reactions, IS415 (Reference 25) are summarized here. Where possible, different amounts of pozzolan or slag should be tested to determine the optimum dosage. In Option A of Reference 25, 990 g of prepared job aggregate is mixed with 440 g of blended cement or cement plus pozzolan or slag. At least three mortar bars are made 12

0.075 ASTM C 227

0.7
G No pozzolan

0.6

Expansion, percent

ASTM C 33 90-day limit

0.5 0.4

Expansion, percent

0.050 Cement + fly ash

B B B B B B B B J J J B J J J J J

ASTM C 441 Cement: 0.92% alkali

0.3 G
G

0.025

0.2 20% Calcined shale 0.1

E EHH

Type IP cement (interground)

20% Calcined clay

H E

0H G E 0

200 300 400 500 600 Time, days Fig. 15. ASR mortar bar expansion comparing a Type IP cement against a cement with added fly ash. Dosage of materials is constant. Reference 26. (#K20303)
Specification (Reference 7) uses ASTM C 227, stating that the test should be run for at least one year or longer to demonstrate effectiveness (see Fig. 15). In Canada, CSA A 23.1-94 recommends a two-year test period for the concrete prism test to evaluate fly ash or slag, and longer periods for silica fume. Petrographic examination of the test specimens is also helpful. Although a limit of 0.04% is used by CSA to establish effectiveness, some researchers suggest a comparison against a low-alkali control.(43) This may also apply to C 227. Reference 43 provides comparison between test methods.

0B J 0

100

4 6 8 Age, months

10

12

Fig. 16. ASR mortar bar expansion versus calcined natural pozzolan additions used with Pyrex glass aggregate. Reference 27. (#K20304) reactivity. ASR expansion decreases as the ash content increases (see Fig. 14). Ashes require different dosages depending on their chemical composition, the alkali content of the concrete, and the reactivity of the aggregate. Laboratory results show that certain high-alkali/high-calcium fly ashes are not effective at normal dosages to control ASR (see Fig. 14). Some of these materials may even exacerbate ASR expansion when used at lower amounts.(24,26) Specifiers should invoke the optional physical and chemical requirements in C 618 when fly ash is used to control ASR. The optional chemical requirement calls for a maximum limit of 1.5% on available alkalies as determined by ASTM C 311, Method for Sampling and Testing Fly Ash or Natural Pozzolans for Use as a Mineral Admixture in Portland Cement Concrete. Blended hydraulic cement conforming to ASTM C 595 or ASTM C 1157 can be used to control ASR (see Fig. 15).(26) Blended cements are made with slag, fly ash, silica fume, or natural pozzolans. Raw or calcined natural pozzolans are sometimes used to control ASR (see Fig. 16).(27) The ASR optional requirement in ASTM C 618 should be specified. In some parts of the country it is common to control ASR by using ground granulated blast-furnace slag. Slag is usually used at higher dosages (25% to 50%) than fly ash to be effective (see Figs. 14 and 17).(28) Silica fume is also an effective mineral admixture for controlling ASR, as only small dosages of silica fume are needed (Figs. 14 and 18).(28) Silica fume should conform to ASTM C 1240, including the optional requirements concerning ASR. Low-alkali portland cement. Low-alkali portland cement (ASTM C 150) with an alkali content not exceeding 0.60% 13

Materials and Methods to Inhibit ASR

A variety of locally available materials can be used to control ASR. Using the above tests where applicable to demonstrate effectiveness in controlling ASR, the following materials and approaches should be considered in designing a concrete mixture. The following options are not listed in priority and, although usually not necessary, they can be used in combination with one another. Pozzolans, slag, and blended cement. Decades of laboratory and field research demonstrate that certain fly ashes can control ASR. The type of ash, its alkali content, chemical composition, and dosage all affect how effectively it will control expansion. In general, ASTM C 618 Class F fly ashes work better than Class C ashes. Class F ashes have higher silicon dioxide contents and lower lime contents, properties that increase the ability to control expansive

0.40 0.35 0.30

Expansion, percent

0.4 Rhyolitic tuff Silica fumes A and B LALow-alkali cement HAHigh-alkali cement
B H

Expansion, percent

ASTM C 441 Cement alkali: 1.19% 14-day data

0.3

B-5%

0.25 0.20 0.15 0.10

B

0.2
H B

HA-0% 0.1 0.04% limit 0B F H J

H B F J H B F J B H F J H F J H F J F J

A-5%
B

B-10%
F J

0.05 0.00 0 10
B B

LA-0%, A-10% -0.1

20 30 40 50 60 Slag content, percent

70

80

50

100

150 200 Days

250

300

350

Fig. 17. ASR mortar bar expansion versus slag content. Reference 28. (#K20306) (equivalent sodium oxide) can be used to reduce ASR. Its use has been successful with slightly reactive to moderately reactive aggregates. Higher alkali levels (between 0.65% and 0.80%) also have been safely used with certain moderately reactive aggregates. However, low-alkali cements are not available in all areas. Also, deleterious reactivity has been observed with certain highly reactive glassy volcanic aggregates, especially andesite and rhyolite rocks, even when low-alkali cements (alkali contents of 0.35% to 0.60%) were used.(30,31) Thus, the use of locally available cements in combination with pozzolans, slags, or blended cements is preferable for controlling ASR. Limiting concrete alkalies. European and Canadian experience indicates that deleterious expansion usually does not take place when reactive aggregates are used in concrete containing less than 3 kg of alkali per cubic meter (5 lb per cubic yard).(16) In some European countries, the effective alkali content is taken as the summation of the total sodium oxide equivalent for portland cement, natural pozzolan, and silica fume, and a percentage of the total sodium oxide equivalent for fly ash and slag.(33) Restrictions on pozzolans are also included in this approach. In Canada, CSA Standard A23.1-94-Appendix B, AlkaliAggregate Reaction dictates the requirements for supplementary cementing materials used to control ASR. A supplementary cementing material may be used to control ASR provided it meets two criteria: (1) it must conform to certain chemical requirements and (2) it must be used at or above the minimum dosage specified. If less than the required minimum amounts are to be used, the concrete prism expansion test (CSA Test Method A23.2-14A) or the mortar-bar accelerated expansion test (CSA Test Method A23.2-25A) is run to verify the materials effectiveness in controlling expansion. Per the chemical requirements, blast-furnace slag can have a maximum 1.0% total alkali content (as Na2O equivalent). Fly ash can have a 14

Fig. 18. ASR concrete prism expansion using two different silica fumes at 0%, 5%, and 10% dosages. Reference 43. (#20305) maximum total alkali content of 4.5% and a maximum water-soluble alkali content of 0.5%. Minimum dosages are 50% for blast-furnace slag and 20% to 30% for fly ash, depending on the alkali content of the ash or slag and combined silica plus alumina content. Appendix B of CSA A23.1-94 cautions that restricting concrete alkali content to 3 kg per cubic meter (5 lb per cubic yard) may not be an effective strategy, however, when: the concrete will be used in mass concrete structures that cant tolerate even slight expansions, the concrete will be exposed to external alkali sources, highly reactive aggregates are used.

As an alternative to the 3-kg limit, specifiers can determine the effective alkali burden from tests on properly performing field structures in a locality. Then they can specify the historically established limit for new structures built with similar materials in a similar environment. ASR-inhibiting compounds. Certain chemical admixtures or cement additives can control ASR.(24,34,35) For example, lithium compounds can effectively reduce ASR expansion in concrete. When lithium hydroxide is added to concrete, it reacts to form a lithium-bearing ASR gel. The lithium-bearing gels have a greatly reduced potential for expansion; some are even non-expansive. Also, less gel is formed in the presence of lithium. The degree to which lithium compounds suppress expansive ASR depends on two factors: aggregate reactivity and cement alkali level. Fig. 19 shows the reduction in expansion with two highly reactive aggregates and varying proportions of sodium to lithium.(24) The optimum level

1.0 1

0.9 B
14-day expansion, percent

1 Normal NaOH solution

terials should be tested in accordance with ASTM C 227 or ASTM C 1293 to verify control of deleterious expansion.

0.8 0.7 0.6 0.5 0.4 0.3 J 0.2 0.1 0 1:0

J B J B J B J B

Highly reactive volcanic aggregate

RECYCLED CONCRETE AS AGGREGATE

When using recycled concrete as coarse aggregate in new concrete, recycled aggregate should be evaluated in the same manner as virgin aggregate (see Fig. 13). Recycling concrete for use as coarse aggregate in new construction can be economical, saving disposal costs of old concrete and reducing the need for virgin aggregate. Increased environmental concerns and diminishing quarry resources may make recycling even more popular in the future.(44) There are two main uses for recycled aggregate in new pavement construction: as a granular subbase material and as coarse aggregate for the new concrete pavement. Research was undertaken to study expansion of new concrete made with recycled concrete aggregate.(36) It was determined that potential for ASR in the new concrete is affected by the old concretes original alkali level, extent of expansion, and the remaining potential reactivity of the aggregate. Also, the alkali content of new concrete had a significant effect on subsequent expansions due to ASR. The use of a low-lime ASTM Class F fly ash greatly reduced expansions due to ASR in new concrete. The research demonstrated that with appropriate selection of cementitious materials, even recycled concrete containing highly reactive aggregate can be used safely. The pavement engineer must know the ASR potential of the recycled aggregate. If information is not available about the cementitious materials and aggregates contained in the old concrete, even if the old concrete has not experienced ASR deterioration, its ASR potential should be petrographically evaluated prior to recycling. Laboratory expansion tests may also be helpful.

Slowly reactive granite gneiss

1:2/3 2/3:1 0:1 1:1 1:1/3 1/3:1 Molar sodium to lithium

Fig. 19. Fourteen-day expansion of mortar bars versus sodium-to-lithium ratio. The effective molar ratio was about 1 to 2/3 for these two reactive aggregates. Reference 24. (#K20307) of lithium hydroxide monohydrate to control expansive ASR is at a level equal to that of the equivalent alkali content of the cement.(35) ASTM C 1260 should not be used to test the effectiveness of lithium compounds. Aggregate selection and beneficiation. Using a nonreactive aggregate is ideal but not always practical. The surest way of predicting aggregate performance is a good service record, but this information is not always available. Instead, it is frequently necessary to investigate aggregate quality through the tests that measure potential reactivity. It may be possible to improve aggregate quality by selective processing known as beneficiation. Processes that have been used are heavy media separation, jigging, rising-current classification, and crushing. In some cases, one of these steps may be able to remove a large portion of the reactive aggregate, but also, may remove some of the good aggregate. It may be possible to blend reactive aggregate with nonreactive aggregate to mitigate the effects of ASR. This blending sometimes occurs as part of the quarrying process, when only small amounts of reactive aggregate are present. Nonreactive aggregates can be purposefully blended with known reactive aggregates. In some cases, this means of diluting aggregates can adequately control expansive ASR. An example of this method is limestone sweetening. Limestone sweetening, or replacing up to 30% of a reactive sand-gravel aggregate with crushed limestone, is effective in preventing deterioration in some sand-gravel aggregate concretes.(50) The resulting combination of ma-

Alkali-Carbonate Reaction
MECHANISM OF ACR
Reactions observed with certain dolomitic rocks are associated with alkali-carbonate reaction (ACR). Reactive rocks usually contain larger crystals of dolomite scattered in and surrounded by a fine-grained matrix of calcite and clay. Calcite is one of the mineral forms of calcium carbonate; dolomite is the common name for calcium-magnesium carbonate. ACR is relatively rare because aggregates susceptible to this reaction are usually unsuitable for use in concrete for other reasonsstrength potential, etc. Argillaceous dolomitic limestone contains calcite and dolomite with appreciable amounts of clay and can contain small amounts of reactive silica. Alkali reactivity of carbonate rocks is not usually dependent on clay mineral composition.(37) Aggregates have potential for expansive ACR if the following lithological characteristics exist:(38,39) clay content, or insoluble residue content, in the range of 5% to 25%; 15

Calcite grains Network of clay

Channel for migration of water and ions (K+ and OH+)

Mg(OH)2 K2CO3 CaCO3

Matrix

Dolomite Crystal

Ratio Ca/Mg reduced gradually

Reactive carbonate rock KOH solution Original sample Immersed in alkali solution Expansion and cracking

Fig. 20. Schematic diagram of the mechanism of alkalicarbonate reaction. A dolomite crystal combines with alkalies in solution to form brucite, and potassium and calcium carbonates. Reference 40. (#K20308) calcite-to-dolomite ratio of approximately 1:1; increase in the dolomite volume up to a point at which interlocking texture becomes a restraining factor; small size of the discrete dolomite crystals (rhombs) suspended in a clay matrix.

Fig. 21. Map cracking pattern caused by ACR (#13664). ticular ACR rock and high-alkali cement showed that both the rate and degree of expansion decreased with a decrease in nominal maximum aggregate size.(41) Concrete can contain a certain percentage of carbonate reactive aggregates without experiencing detrimental expansion. Alkali-carbonate reaction is also affected by pore solution alkalinity. ACR can occur in a solution with a relatively low pH. As the pH of the pore solution increases, potential for the alkali-carbonate reaction increases. During ACR, calcium hydroxide produced by portland cement hydration can combine with sodium carbonate to form more sodium hydroxide and calcium carbonate. This reaction not only regenerates alkali, but also reduces the concentration of carbonate ions and aggravates the dedolomitization reaction. Low-alkali cements and mineral admixtureseven at high levelsare thus ineffective at reducing ACR to acceptable limits.(8)

Dedolomitization
Dedolomitization, or the breaking down of dolomite, is normally associated with expansive ACR.(37) Concrete that contains dolomite and has expanded also contains brucite (magnesium hydroxide, Mg(OH)2), which is formed by dedolomitization. Dedolomitization proceeds according to the following equation:(38) CaMgCO3 (dolomite) + alkali hydroxide solution MgOH2 (brucite) + CaCO3 (calcium carbonate) + K2CO3 (potassium carbonate) + alkali hydroxide Fig. 20 illustrates this process. Expansion may be due to a combination of migration of alkali ions and water molecules into the restricted space of the fine-grained matrix surrounding the dolomite rhomb, migration of these materials into the rhomb, and the growth and rearrangement of the dedolomitization products, especially brucite, which exerts pressures as it crystallizes.(40) The dedolomitization reaction and subsequent crystallization of brucite may cause considerable expansion. Whether dedolomitization causes expansion directly or indirectly, its usually a prerequisite to other expansive processes.(8) In one investigation of pavement deterioration, dolomite did not appear to be altered as no brucite was found. Cracking was due to expansive forces arising from a chemical reaction and expansion of the clay minerals in the aggregate matrix.(45)

VISUAL SYMPTOMS OF EXPANSIVE ACR

ACR, like ASR, is a chemical process that can induce physical damageexpansion and cracking of concrete (Fig. 21). Information obtained from a site inspection should be evaluated along with testing to determine the destructive mechanism. ACR-affected concrete does not exhibit telltale features to distinctly identify alkali-carbonate reaction as the cause of cracking. The crack pattern will be influenced by restraint conditions and moisture availability. In slabs, decks, and footings, when a top-to-bottom or side-to-side moisture gradient exists, map cracking is likely to occur. This growth can cause closed joints, misalignment of adjoining members, crushed concrete, and pavement blowups.(38) Because low levels of alkali can initiate ACR and higher levels can exacerbate it, the use of deicer salts that contain alkalies can increase damage.

Other Factors
Maximum size of the reactive aggregate influences the amount and extent of reaction. Testing done with a par16

TEST METHODS FOR IDENTIFYING ACR DISTRESS

Brucite is usually formed during destructive ACR and is evident in expanded concrete. Its presence must be determined through petrographic or other analysis. The concrete and

carbonate rocks must be examined petrographically in accordance with ASTM C 856 and C 295 respectively. Physical expansion tests, such as ASTM C 586, performed in a laboratory are also helpful to verify the presence of ACR.

TEST METHODS FOR IDENTIFYING POTENTIALLY REACTIVE AGGREGATE

ACR is rather uncommon because the carbonate rocks containing proportions of dolomite, calcite, and insoluble material that will lead to dedolomitization do not often comprise a major portion of the aggregate supply.(1) The three tests commonly used to identify potentially alkali-carbonate reactive aggregate are described below and in Table 2 at the end of this document.

Petrographic Examination (ASTM C 295)

Petrographic examination, ASTM C 295, can be used to establish the character of rock. Potentially alkali-carbonate reactive aggregates have a characteristic lithology that makes them easy to identify. The rock is considered to be reactive if the physical make-up is a fine-grained matrix of calcite and clay surrounding rhombic crystals of dolomite. With no known exceptions, all rocks with the characteristic lithology have expanded in an alkaline environment, and all rocks in concrete that have expanded due to ACR exhibit the characteristic lithology.(38) Petrographic testing identifies ACR-susceptible rocks.

cated with the aggregate in question and the job cement. If the job cement is not known, a Type I or Type II cement conforming to ASTM C 150 should be used. The test should preferably run for one year, but 3 or 6 months can be used if longer test times are not feasible. An expansion equal to or more than 0.03% at one year, 0.025% at 6 months, or 0.015% at 3 months indicates potentially deleterious aggregate (ASTM C 1105, Appendix). This test makes it possible to measure expansive tendencies of cement-aggregate combinations rather than expansive characteristics of the rock alone, but the yearlong test duration is a drawback. If this test has been performed on a cement-aggregate combination involving an aggregate that has not been examined petrographically using C 295, or has not been measured by the rock prism expansion test (C 586), one or both of those tests should also be performed. The Canadian Standards Association CSA A23.2-14A, Potential Expansivity of Aggregates (Procedure for Length Change Due to Alkali-Aggregate Reaction in Concrete Prisms), is similar to C 1105, as both tests require similar specimens and storage conditions. Cement alkali levels, aggregate gradation, and limits that indicate expansive behavior differ slightly for the two methods. The Canadian version of the test stipulates the cement alkali level to be used, making the test more standardized and allowing comparison among test results. The higher cement alkali content in A23.2-14A makes it a more aggressive test than C 1105, so it should identify potentially reactive rocks that might pass C 1105.

Rock Cylinder Method (ASTM C 586)

ASTM C 586, Test Method for Potential Alkali Reactivity of Carbonate Rocks for Concrete Aggregates (Rock Cylinder Method), can determine the expansive characteristics of carbonate rocks. A small rock cylinder 35 mm long by 9 mm dia. (1.38 x 0.35 in.) is immersed in an alkaline (1 N NaOH) solution at room temperature. Length change of the specimen may be monitored for up to or over a year, but expansive tendencies are usually evident after about 1 month. A 28-day expansion of 0.10% or more indicates potential for deleterious expansion in a service environment. This test procedure has a few disadvantages. Obtaining a representative sample may be difficult and test duration may be long. Furthermore, this test is intended only as a supplement to other test procedures. Expansive behavior of aggregate can be predicted by the rock cylinder method. But since expansion of concrete depends on other variables that include water-cement ratio, water-soluble alkalies, and the paste-to-aggregate ratio, ASTM C 586 alone should not be used as an acceptance test for aggregate.(38)

CONTROL OF ALKALI-CARBONATE REACTION

ACR-susceptible aggregate has a specific composition that is readily identified by petrographic testing. If a rock indicates ACR-susceptibility, preventive measures must be taken. Its difficult to arrest alkali-carbonate reaction once expansion has begun.

Selective Quarrying
Quarries are routinely mapped and tested to define the rock groupings and stratification. Reactive rock may be naturally isolated in layers or by other physical features. In this case, selective quarrying can be used to completely avoid the reactive strata. Small portions of reactive aggregate may become naturally diluted when the rock is processed.

Blended Aggregate Concrete Prism Tests (ASTM C 1105 and CSA A23.2-14A)
ASTM C 1105, Length Change of Concrete Due to AlkaliCarbonate Rock Reaction, is best used to verify potential reactivity of an aggregate that is known to contain ACRsusceptible rock. Six concrete prism specimens are fabriAnother option is blending reactive aggregate with aggregate proven by testingand service record, if possibleto be nonreactive. This can mitigate the reactive aggregates detrimental effects on concrete. Accordingly, the Appendix in ASTM C 1105 limits the safe proportion of reactive aggregate to 20% for either the coarse or fine sizes when used alone, or a limit of 15% total when used 17

in combination. Test the resulting diluted or blended aggregate using C 1105 to be certain it is acceptable for use in portland cement concrete.

Q: If an aggregate passes ASTM C 1260 and other ASR requirements, should there be restrictions on cementitious materials such as limits on cement alkalies?
A: No. Nonreactive aggregate can be used without undue restrictions on cementitious materials.

Aggregate Size
Limiting the nominal maximum size of aggregate is another way to control ACR. Use the smallest practical nominal maximum size of aggregate to minimize detrimental expansion.(41)

Q: Should an aggregate be eliminated from use if it fails ASTM C 1260 or other ASR tests?
A: No. First, deleterious behavior must be confirmed by field history or multiple test methods before an aggregate is classified as potentially reactive. Even then, aggregates that are classified as potentially reactive can often be safely used with pozzolans, slags, or ASR-inhibiting compounds, blended cements, or by limiting the total alkali content of the concrete. Another option would be to blend the aggregate with a nonreactive aggregate to a level that is not harmfully reactive.

Low-Alkali Cement
Low-alkali cement is not generally effective in controlling expansive ACR. Even cements with alkali levels as low as 0.40% can produce pore solutions with sufficient alkalinity to initiate reaction. Due to the chemistry of the alkali-carbonate reactions, alkalies are regenerated rather than consumed. They undergo a series of reactions with aggregates and are then available to react again with more aggregate.(37) The cement alkali level is of less concern than the alkali content of concrete, which should be kept as low as possible when using alkali-carbonate reactive aggregate.

Q: Is it possible for ASTM C 1260 to give a false negative? A false positive?

A: Yes to each. False negatives are not common with ASTM C 1260. Most reactive aggregates will be identified by this test because it is so severe. False positives, on the other hand, are more common, again because the test method is so severe. The alkali hydroxide concentration of most field concretes is much lower than that of the test method and is at a level that does not induce harmful reactivity for many aggregates that fail C 1260. It cannot be stressed enough that supplemental information should be developed to evaluate an aggregate.

Pozzolans
Mineral admixtures are not very effective in controlling ACR. Pozzolans tie up alkalies by reacting with them, but ACR can be initiated with low levels of alkali, and pozzolans may not consume enough alkalies to limit the reaction. Pozzolans do reduce paste porosity, which helps limit the moisture available for reaction. The reduced concrete porosity is beneficial, but concrete in most environments contains enough moisture to sustain alkali-carbonate reaction. Larger amounts of pozzolans could possibly control ACR, but these amounts are generally too high to be practical. ACR is more effectively controlled by processing the aggregate to minimize particle size or the amount of reactive aggregate in the concrete.

Q: If a pozzolan is required in a given concrete mixture, should the aggregate be tested alone to determine its potential for ASR? Should all of ASTM tests C 227, C 289, C 295, C 1260, and C 1293 be run?
A: If the concrete is required to contain a pozzolan anyway, consider using a modified ASTM C 1260 to investigate expansive tendencies of the combination of aggregates and cementitious materials. Note that some pozzolans at some dosages can increase ASR potential, so the mix design should be strictly defined and adhered to in order for the combination of materials to work effectively. It is not necessary to require every test to prove effectiveness and its not very efficient. Choose one test, such as a modified C 1260 or C 441, that uses the proposed cementitious materials to evaluate their effectiveness at controlling expansion. If one or two years is available, C 227 or C 1293 can be performed using the pozzolan.

ACR-Inhibiting Compounds
Lithium compounds appear to control ACR, but more research is needed.(35)

Common Questions on AAR

Q: If an aggregate has good field performance, why test the aggregate for deleterious behavior?
A: This question requires a two-part answer. Field history is a valuable source of information that indicates an aggregates behavior in service. If conditions of service and the ingredients, including total alkali content of the concrete, are the same in the existing and proposed concrete structures, there is no need to test an aggregate that has a good record of service. If however, the aggregate is varying significantly within the quarry, or if exposure conditions of the proposed structure will be more severe than in structures exhibiting good performance, or if the concrete ingredients are significantly different, the aggregate should be tested. 18

Q: What is a modified ASTM C 1260?

A: C 1260 is defined as a method of assessing potential aggregate reactivity. There are at least two nonstandard modified versions of this test, neither of which has been officially recognized by ASTM. One version assesses expansive tendencies of specimens made with the aggregate in question and blended cement or cement plus pozzolan or slag under the same (C 1260) accelerated test conditions. It tests the effectiveness of the cement-pozzolan-aggregate combination and not just the aggregate. In another suggested modified version of C 1260, the test samples remain the same as in C 1260, but the alkali con-

centrations of the exposure solutions vary to help determine safe portland cement alkali levels that may be used with the aggregate being tested. See Reference 24.

Related Publications
The publications cited in this text as well as other related pulbications are available from Portland Cement Association. The following are particularly useful:

Q: Isnt the easiest way to control ASR to avoid using potentially reactive aggregate and require use of a low-alkali cement?
A: Though this can effectively control ASR, it would not be efficient. Not all geographical areas have nonreactive aggregates or low-alkali cements. Requiring these materials would probably result in extra costs and lost time. Concrete mixtures should be designed to be economical in addition to being safe from ASR potential. Local materials usually provide the greatest availability and best economy.

Q: Are all fly ashes equivalent at controlling expansion from ASR, and is there one single dosage that will always be effective?
A: All fly ashes are not equivalent at controlling expansion due to ASR. Class F ashes (typically low CaO) as a group are usually more effective than Class C ashes. And within a group, one ash can be more effective than another. Also, aggregates of different degrees of reactivity and different cement alkali levels require different dosages of ash. Since no one dosage is effective with all fly ashes, it is necessary to investigate any fly ashs ability to control ASR by testinga 15% to 20% dosage is a good starting place. Concurrent testing of various dosages helps determine the appropriate fly ash dosage. Use a modified ASTM C 1260 or C 441.

Q: Can I use ASTM C 1260 to test the effectiveness of a portland cement to control ASR?
A: No. The standard test as written is not designed to test the effect of different portland cements on ASR. The test specimen alkali content rises to approach that of the solution in which it is submerged, regardless of the original alkali content of the cement.

Guide Specification for Concrete Subject to Alkali-Silica Reactions (IS415) Alkali-Silica Reactivity: An Overview of Research (LT177) Diagnosis of Alkali-Silica Reaction (LT166) Eliminating or Minimizing Alkali-Silica Reactivity (LT178) Handbook for Identifying Alkali-Silica Reactivity Highway Structures (LT165) Detecting and Avoiding Alkali-Aggregate Reactivity, Concrete Technology Today (PL913) Cornells Gel Fluorescence Test Identifies ASR Products in Concrete, Concrete Technology Today (PL922) Lithium-Based AdmixturesAn Alternative for Preventing Expansive Alkali-Silica Reaction, Concrete Technology Today (PL931) Design and Control of Concrete Mixtures (EB001) Alkali-Silica Reactivity: Some Reconsiderations (RD076) Lithium Salt AdmixtureAn Alternative Method to Prevent ASR (RP307) An Interpretation of Published Researches on the AlkaliAggregate Reaction: Part 1The Chemical Reactions and Mechanism of Expansion; Part 2A Hypothesis Concerning Safe and Unsafe Reactions with Reactive Silica in Concrete (RX055) Osmotic Studies and Hypothesis Concerning AlkaliAggregate Reaction (RX057) Field and Laboratory Studies on the Reactivity of SandGravel Aggregates (RX221) The Use of Recycled-Concrete Aggregate from Concrete Exhibiting Alkali-Silica Reactivity (RD114) Development of a Cement to Inhibit Alkali-Silica Reactivity (RD115) Recycling Concrete Pavement (TB014)
To order, write, phone or fax: Order Processing Portland Cement Association P. O. Box 726 Skokie, IL 60076-0726 847-966-6200 (phone); 847-966-9666 (fax) A complete list of PCA materials is available in the Associations free catalog of publications, computer software, and audiovisual materials (MS254).

Q: A pavement engineer wants to replace some of the cement in a standard mix with fly ash to control ASR but is concerned about poor strength gain in cold weather. Any solutions?
A: Yes. Use the fly ash as an addition to the standard mix instead of replacing the cement. The fly ash will still control ASR and the mix will gain strength in cold weather similar to that of the cement-only mix. Chemical admixtures or Type III cement can also be used to accelerate strength gain.

19

20
Type of test Type of sample Duration of test Measurement Criteria Comments
33, maximum 0.10% 6 months, or if not a 6-month period, 0.05% at 3 months Per ASTM C expansion at available for maximum of Varies: first Length change measurement at 14 days, then 1,2,3,4,6,9, and 12 months and every 6 months after that as necesssary 16 days Length change If greater than 0.10%, go to supplementary test procedures; if greater than 0.20%, indicative of potentially deleterious expansion Point plotted on graph falls in deleterious or potentially deleterious area At least 4 mortar bars; standard dimensions: 25x25x285 mm (1x1x11-1/4 inches) Mortar bars stored over water at 37.8 C (100 F) and high relative humidity Immersion of mortar At least 3 mortar bars bars in alkaline solution at 80 C (176 F) Sample reacted with 3-25 gram samples of alkaline solution at 80 C crushed and sieved (176 F) aggregate 24 hours Drop in alkalinity and amount of silica solubilized Concrete prisms stored over water at 38C (100.4 F) 3 prisms per cementaggregate combination standard dimensions: 75x75x285 mm (3x3x11-1/4 inches) Varies: first measureLength change ment at 7 days, then 28 and 56 days, then,3,6,9, and 12 months and every 6 months after that as necesssary 52 weeks Length change Mortar bars stored in water at 23 C (73.4 F) Per ASTM C 33, unsatisfactory aggregate if expansion equals or exceeds 0.200% at 1 year 3 mortar bars per cement-aggregate combination standard dimensions: 25x25x285 mm (1x1x11-1/4 inches) Mortar barsusing Pyrex At least 3 mortar bars glass as aggregate and also 3 mortar bars of stored over water at control mixture 37.8 C (100 F) and high relative humidity Varies: first Length change measurement at 14 days, then 1,2,3,4,6,9, and 12 months and every 6 months after that as necessary Short durationas long as it takes to visually examine the sample Description of type and proportion of minerals in aggregate Per ASTM C 989, minimum 75% reduction in expansion or 0.020% maximum expansion or per C618, comparison against low-alkali control Highly reactive artificial aggregate may not represent real aggregate conditions Pyrex contains alkalies Visual identification Varies, but should be representative of entire source Not applicable Visual and microscopic examination of prepared samplessieve analysis, microscopy, scratch or acid tests Visual (unmagnified) and microscopic examination of prepared samples Staining of a freshlyexposed concrete surface and viewing under UV light Varies with knowledge of quarry: cores 53 to 100 mm in diameter (2-1/8 to 4 in.), 45 kg (100 lb.) or 300 pieces, or 2 kg (4 lb.) At least one core 150 mm diameter by 300 mm long (6 in. diameter by 12 in. long) Varies: core, lapped surface, broken surface Short durationvisual Particle characteristics, Not applicable examination does not like shape, size, texture, involve long test periods color, mineral composition, physical condition Short durationincludes preparation of samples and visual and microscopic examination Immediate results Is the aggregate known to be reactive? Orientation and geometry of cracks Is there any gel present? Intensity of fluorescence

Table 1. Test Methods for Alkali-Silica Reactivity

Test name

Purpose

To test the susceptibility of cement-aggregate combinations to expansive reactions involving alkalies

To test the potential for deleterious alkali-silica reaction of aggregate in mortar bars

Test may not produce significant expansion, especially for carbonate aggregate Long test duration Expansions may not be from AAR Very fast alternative to C 227 Useful for slowly reacting aggregates or ones that produce expansion late in the reaction

Potential AlkaliReactivity of CementAggregate Combinations (Mortar-Bar Method) (ASTM C 227) Potential Alkali Reactivity of Aggregates (MortarBar Method) (ASTM C 1260) Potential AlkaliSilica Reactivity of Aggregates (Chemical Method) (ASTM C 289)

To determine potential reactivity of siliceous aggregates

To determine the potential ASR expansion of cementaggregate combinations

Quick results Some aggregates give low expansions even though they have high silica content Not reliable Per Appendix X1, potentially Long test duration for deleteriously reactive if expansion meaningful results equals or exceeds 0.04% at one year Use as a supplement to C 227, C 295, C 289, and C 1260 Similar to CSA A23.2-14A Primarily used for aggregates from Oklahoma, Kansas, Nebraska, and Iowa

To determine the potential ASR expansion of cementaggregate combinations

Determination of Length Change of Concrete Due to Alkali-Silica Reaction (Concrete Prism Test) (ASTM C 1293) Potential Volume Change of CementAggregate Combinations (ASTM C 342) Effectiveness of Mineral Admixtures or GBFS in Preventing Excessive Expansion of Concrete Due to the Alkali-Silica Reaction (ASTM C 441) Constituents of Natural Mineral Aggregates (ASTM C 294)

To determine effectiveness of mineral admixtures or slag in controlling expansion from ASR

These descriptions are used to characterize naturallyoccurring minerals that make up common aggregate sources

Petrographic Examination of Aggregates for Concrete (ASTM C 295)

To give descriptive nomenclature for the more common or important natural mineralsa good starting point to predict behavior To outline petrographic examination procedures of aggregatesan aid in determining their performance

Petrographic Examination of Hardened Concrete (ASTM C 856)

Uranyl-Acetate Treatment Procedure (ASTM C 856 Annex)

To outline petrographic examination procedures of hardened concreteuseful in determining condition or performance To identify products of ASR in hardened concrete

See measurementthis examination determines if ASR reactions have taken place and their effects upon the concrete Used in conjunction with other tests Lack of fluorescence Identifies small amounts of ASR gel whether they cause expansion or not Opal, a natural aggregate, and carbonated paste can glow interpret results accordingly Test must be supplemented by petrographic examination and physical tests for determining concrete expansion

Usually includes optical microscopy; may include XRD analysis, differential thermal analysis, or infrared spectroscopysee C 294 for descriptive nomenclature Specimens can be examined with stereomicroscopes, polarizing microscopes, metallographic microscopes, and SEM

Table 2. Test Methods for Alkali-Carbonate Reactivity

Type of test Type of sample Duration of test Measurement Criteria Comments
Securing representative sample may be difficult Can be a long test period Intended as supplement to other test procedures Not applicable Up to or over 1 year, but Length change expansive tendencies are usually evident after 28 days Immersion of rock One rock cylinder, 35mm cylinders in alkaline long and 9 mm in solution at 23 C (73.4 F) diameter (1.38 inches long and 0.35 inches dia)

Test name

Purpose

To determine the expansive characteristics of carbonate rocks

Potential Alkali Reactivity of Carbonate Rocks for Concrete Aggregates (Rock Cylinder Method) (ASTM C 586) Length Change of Concrete Due to Alkali-Carbonate Rock Reaction (ASTM C 1105) Concrete prisms stored at 23 C (73.4 F) and high relative humidity Length change Aggregate is most likely reactive if length change at 1 year (preferred measurement) is greater than or equal to 0.030% (0.025% at 6 months or 0.015% at 3 months) Aggregate is most likely reactive if length change is greater than 0.04% at 1 year Concrete prisms stored over water at 38C (100.4 F) Length change 6 concrete prisms 12 months 75x75x285 mm (3x3x111/4 in.)or 100-mm (4in.) square for 50-mm (2 in.) maximum size aggregate 3 prisms 75x75x285 mm 1 year (3x3x11-1.4 in.) and an additional (fourth) prism may be cast for petrographic examination Varies, but should be representative of entire source Short durationas long as it takes to visually examine the sample Description of type and proportion of minerals in aggregate Not applicable

To determine susceptibility of cement-aggregate combinations to expansive ACR

To evaluate the potential expansivity of both coarse and fine aggregates

Long test period Best used as a verifier of potential reactivity for a rock known to contain a potentially deleterious alkali-carbonate rocksee Appendix Long test period Use in conjunction with other information such as petrographic examination Similar to ASTM C 1293

Potential Expansivity of Aggregates (Procedure for Length Change Due to AlkaliAggregate Reaction in Concrete Prisms) (CSA A23.2-14A) Constituents of Natural Mineral Aggregates (ASTM C 294) Visual identification

These descriptions are used to characterize naturallyoccurring minerals that make up common aggregate sources

Petrographic Examination of Aggregates for Concrete (ASTM C 295) Visual and microscopic examination of prepared samplessieve analysis, microscopy, scratch or acid tests Visual (unmagnified) and microscopic examination of prepared samples, acid etching Varies with knowledge of quarry: cores 53 to 100 mm in diameter (2-1/8 to 4 in.), 45 kg (100 lb.) or 300 pieces, or 2 kg (4 lb.) At least one core 150 mm diameter by 300 mm long (6 in. diameter by 12 in. long) Short durationincludes preparation of samples and visual and microscopic examination

To give descriptive nomenclature for the more common or important natural mineralsa good starting point to predict behavior To outline petrographic examination procedures of aggregatesan aid in determining their performance CaOH crystals are smaller than in undamaged concrete Are rims or partial rims depleted in CaOH? Pale color of rims may indicate gel

Particle characteristics, Not applicable Short durationvisual like shape, size, texture, examination does not involve long test periods color, mineral composition, physical condition See measurementthis examination determines if ACR reactions have taken place and their effects upon the concrete Used in conjunction with other tests

Petrographic Examination of Hardened Concrete (ASTM C 856)

To outline petrographic examination procedures of hardened concreteuseful in determining condition or performance

Usually includes optical microscopy; may include XRD analysis, differential thermal analysis, or infrared spectroscopysee C 294 for descriptive nomenclature Specimens can be examined with stereomicroscopes, polarizing microscopes, metallographic microscopes, and SEM

21

REFERENCES
1. Mather, Bryant, New Concern over Alkali-Aggregate Reaction , Joint Technical Paper by National Aggregates Association and National Ready Mixed Concrete Association, NAA Circular No. 122 and NRMCA Publication No. 149, Silver Spring, Maryland, 1975. 2. Stanton, Thomas E., Expansion of Concrete Through Reaction Between Cement and Aggregate, Proceedings, American Society of Civil Engineers, Vol. 66, New York, 1940, pages 1781-1811. 3. Tests of Concrete Road Materials from California, Major Series 285, Research Reports, Portland Cement Association, Skokie, Illinois, April 1940. 4. Diamond, S., Barneyback, R. S. Jr., and Struble, L. J., On the Physics and Chemistry of Alkali-Silica Reactions, Proceedings of the Fifth Conference on Alkali-Aggregate Reaction in Concrete , National Building Research Institute, Pretoria, South Africa, 1981, pages 1-11. 5. Xu, Huayong, On the Alkali Content of Cement in AAR, in Concrete Alkali-Aggregate Reactions, Proceedings of the 7th International Conference, edited by Grattan-Bellew, Patrick E., Noyes Publications, Park Ridge, New Jersey, 1987, pages 451-455. 6. Poole, A. B., Introduction to Alkali-Aggregate Reaction in Concrete, in The Alkali-Silica Reaction in Concrete, edited by Swamy, R.N., Van Nostrand Reinhold, New York, 1992. 7. Guide Specifications for Concrete Subject to AlkaliSilica Reactions, Mid-Atlantic Regional Technical Committee, [Available through NRMCA, Silver Spring, Maryland], June 1993. 8. Tang, Mingshu, and Deng, Min, and Lon, Xianghui, and Han, Sufen, Studies on Alkali-Carbonate Reaction, ACI Materials Journal, American Concrete Institute, Detroit, January-February 1994, Pages 26-29. 9. Stark, David, The Moisture Condition of Field Concrete Exhibiting Alkali-Silica Reactivity, CANMET/ACI Second International Conference on Durability of Concrete , SP-126, American Concrete Institute, Detroit, 1991, pages 973-987. 10. Stark, David, Effects of Water-Cement Ratio on Expansion Due to ASR, Concrete Technology Today, PL951, Portland Cement Association, Skokie, Illinois, March 1995. 11. Grattan-Bellew, P. E., Alkali-Silica Reaction Canadian Experience, in The Alkali-Silica Reaction in Concrete, edited by Swamy, R.N., Van Nostrand Reinhold, New York, 1992, pages 223-248. 12. Helmuth, Richard, Alkali-Silica Reactivity: An Overview of Research , SHRP-C-342, Strategic Highway Research Program, Washington, D. C., 1993. Also PCA Publication LT177. 22

13. Swamy, R.N., and Al-Asali, M.M., New Test Methods for Alkali-Silica Reaction, in Concrete Alkali-Aggregate Reactions, Proceedings of the 7th International Conference, edited by Grattan-Bellew, Patrick E., Noyes Publications, Park Ridge, New Jersey, 1987, pages 324-329. 14. Stark, David, Effects of Chloride Solutions on Expansion Due to ASR, Concrete Technology Today, PL953, Portland Cement Association, Skokie, Illinois, November 1995. 15. Palmer, D. (chairman), The Diagnosis of Alkali-Silica Reaction, Report of a Working Party, 2nd edition, British Cement Association, Slough, Great Britain, 1992. Also PCA Publication LT166. 16. Hawkins, M. R., Alkali-Aggregate Reaction: Minimising the Risk of Alkali-Silica Reaction, Guidance Notes, Report of a Working Party, Ref. 97-304, Cement and Concrete Association, Wexham Springs, Slough, UK, 1983. 17. Perenchio, William F., Kaufman, Irwin, and Krause, Robert J., Concrete Repair in a Desert Environment, Concrete International, American Concrete Institute, Detroit, February 1991, pages 23-26. 18. Danay, Ari, Structural Mechanics Methodology in Diagnosing and Assessing Long-Term Effects of AlkaliAggregate Reactivity in Reinforced Concrete Structures, ACI Materials Journal, American Concrete Institute, Detroit, January-February 1994, pages 54-62. 19. Figg, John, ASRInside Phenomena and Outside Effects (Crack Origin and Pattern), in Concrete AlkaliAggregate Reactions, Proceedings of the 7th International Conference, edited by Grattan-Bellew, Patrick E., Noyes Publications, Park Ridge, New Jersey, 1987, pages 152-156. 20. Liu, T. C., Causes of Deterioration, Lesson 201, Maintenance and Repair of Concrete Structures, U.S. Army Corps of Engineers Waterways Experiment Station, Vicksburg, Miss., 1981. 21. Stark, David, Handbook for the Identification of AlkaliSilica Reactivity in Highway Structures, SHRP-C/FR-91101, Strategic Highway Research Program, Washington, D.C., 1991. Also PCA Publication LT165. 22. Natesaiyer, Kumar, and Hover, Kenneth, Cornells Gel Fluorescence Test Identifies ASR Products in Concrete, Concrete Technology Today , PL922, Portland Cement Association, Skokie, Illinois, July 1992. 23. Stark, David, Alkali-Silica Reactions in Concrete, Significance of Tests and Properties of Concrete and Concrete-Making Materials, ASTM STP 169C, edited by Klieger, Paul and Lamond, Joseph F., American Society for Testing and Materials, Philadelphia, 1994, pages 365-371. Also PCA Publication LT205. 24. Stark, David, Eliminating or Minimizing Alkali-Silica Reactivity, SHRP-C-343, Strategic Highway Research Program, Washington, D. C., 1993. Also PCA Publication LT178.

25. Guide Specification for Concrete Subject to AlkaliSilica Reactions, IS415, Portland Cement Association/ American Concrete Pavement Association, Skokie, Illinois, 1995. 26. Carrasquillo, Ramon L., and Farbiarz, Josef, AlkaliAggregate Reaction in Concrete Containing Fly Ash: Final Report, Research Report 450-3F, Center for Transportation Research, University of Texas at Austin, Austin, Texas, October 1989. 27. Lerch, William, Studies of Some Methods of Avoiding the Expansion and Pattern Cracking Associated with the Alkali-Aggregate Reaction, Research Department Bulletin RX031, Portland Cement Association, Skokie, Illinois, February, 1950. 28. Buck, A., Use of Cementitious Materials Other Than Portland Cement, Concrete Durability, Katharine and Bryant Mather International Conference , SP100, American Concrete Institute, Detroit, 1987, pages 1863-1881. 29. Guide for the Use of Silica Fume in Concrete, ACI Committee 234 Report, American Concrete Institute, Detroit, 1996. 30. Stark, D. C., Alkali-Silica Reactivity: Some Reconsiderations, Research Department Bulletin RD076, Portland Cement Association, Skokie, Illinois, 1981. 31. Kosmatka, Steven H., and Fiorato, Anthony E., Detecting and Avoiding Alkali-Aggregate Reactivity, Concrete Technology Today, PL913, Portland Cement Association, Skokie, Illinois, November 1991. 32. Johnston, Colin D., Alkali-Silica Reactivity in ConcreteImportance of Cement Content and Alkali Equivalent, in Concrete Alkali-Aggregate Reactions, Proceedings of the 7th International Conference , edited by Grattan-Bellew, Patrick E., Noyes Publications, Park Ridge, New Jersey, 1987, pages 477-482. 33. Hobbs, D. W., Effect of Mineral and Chemical Admixtures on Alkali-Aggregate Reaction, Eighth International Conference on Alkali-Aggregate Reaction, British Cement Association, Wexham Springs, Slough, UK, 1989. 34. Stark, David C., Lithium Salt AdmixtureAn Alternative Method to Prevent Expansive Alkali-Silica Reactivity, Proceedings of the 9th International Conference on Alkali-Aggregate Reaction in Concrete, The Concrete Society, London, July, 1992. Also PCA Publication RP307. 35. Gajda, John, Development of a Cement to Inhibit Alkali-Silica Reactivity, Research and Development Bulletin RD115, Portland Cement Association, Skokie, Illinois, 1996. 36. Stark, David, The Use of Recycled-Concrete Aggregate from Concrete Exhibiting Alkali-Silica Reactivity, Research and Development Bulletin RD113, Portland Cement Association, Skokie, Illinois, 1996.

37. Hadley, D.W., Alkali Reactivity of Carbonate Rocks Expansion and Dedolomitization , Research Department Bulletin RX139, Portland Cement Association, Skokie, Illinois, 1961. 38. Ozol, Michael A., Alkali-Carbonate Rock Reaction, Significance of Tests and Properties of Concrete and Concrete-Making Materials, ASTM STP 169C, edited by Klieger, Paul and Lamond, Joseph F., American Society for Testing and Materials, Philadelphia, 1994, pages 372-387. 39. Swenson, E. G., and Gillott, J. E., Alkali Reactivity of Dolomitic Limestone Aggregate, Magazine of Concrete Research, Vol. 19 No. 59, Cement and Concrete Association, London, June 1967, pages 95-104. 40. Tang, Mingshu, Liu, Zheng, and Han, Sufeng, Mechanism of Alkali-Carbonate Reaction, in Concrete Alkali-Aggregate Reactions, Proceedings of the 7th International Conference, edited by Grattan-Bellew, Patrick E., Noyes Publications, Park Ridge, New Jersey, 1987, pages 275-279. 41. Swenson, E. G., and Gillot, J. E., Characteristics of Kingston Carbonate Rock Reaction, Concrete Quality Control, Aggregate Characteristics, and the CementAggregate Reaction, Bulletin No. 275, Highway Research Board, Washington, D. C., 1960, pages 18-31. 42. Newlon, H. H. Jr., Sherwood, W. C., and Ozol, M. A., Potentially Reactive Carbonate Rocks, Progress Report No. 8, A Strategy for Use and Control of Carbonate Rocks Including an Annotated Bibliography of Virginia Research, VHRC 71-R41, Virginia Highway Research Council, Charlottesville, VA, June 1972. 43. Brub, M. A., and Duchesne, J., Evaluation of Testing Methods Used for Assessing the Effectiveness of Mineral Admixtures in Suppressing Expansion Due to Alkali-Aggregate Reaction, Fly Ash, Silica Fume, Slag, and Natural Pozzolans in Concrete, Proceedings of the Fourth International Conference, SP132, edited by Malhotra, V. M., American Concrete Institute, Detroit, 1992, pages 549-575. 44. Recycling Concrete Pavement , TB014, American Concrete Pavement Association, Skokie, Illinois, 1993. 45. Wong, Sam, Investigation of Concrete from I-20 Near Monroe, Louisiana, Army Corps of Engineers Report to Louisiana Transportation Research Center, [PCA Serial No. 2059], 1996. 46. Wood, Jonathan G. M., Young, Jeff S., and Ward, David E., The Structural Effects of Alkali-Aggregate Reaction on Reinforced Concrete, in Concrete Alkali-Aggregate Reactions, Proceedings of the 7th International Conference, edited by Grattan-Bellew, Patrick E., Noyes Publications, Park Ridge, New Jersey, 1987, pages 157-162.

23

REFERENCES (continued)
47. Jones, Trevor N., and Poole, Alan B., Alkali-Silica Reaction in Several U. K. Concretes: The Effect of Temperature and Humidity on Expansion, and the Significance of Ettringite Development, in Concrete Alkali-Aggregate Reactions, Proceedings of the 7th International Conference, edited by Grattan-Bellew, Patrick E., Noyes Publications, Park Ridge, New Jersey, 1987, pages 446-450. 48. Concrete Materials and Methods of Concrete Construction, Appendix B, Alkali-Aggregate Reaction, CAN/CSA-A23.1, Canadian Standards Association, Toronto, 1994. 49. Second International Conference on Alkali-Aggregate Reactions in Hydroelectric Plants and Dams, U. S. Committee on Large Dams, Denver, Colorado, 1995, page 8. 50. Hadley, David W., Field and Laboratory Studies on the Reactivity of Sand-Gravel Aggregates, Research Department Bulletin RX221, Portland Cement Association, Skokie, Illinois, 1968. 51. Fournier, B., and Malhotra, V. M., Reducing Expansion Due to Alkali-Silica Reactivity, Concrete International, American Concrete Institute, Farmington Hills, Michigan, March 1996, pages 55-59.
This publication is intended SOLELY for use by PROFESSIONAL PERSONNEL who are competent to evaluate the significance and limitations of the information provided herein, and who will accept total responsibility for the application of this information. The Por tland Cement Association DISCLAIMS any and all RESPONSIBILITY and LIABILITY for the accuracy of and the application of the information contained in this publication to the full extent permitted by law.

ACKNOWLEDGMENTS
Many individuals and organizations within the cement and concrete industries have contributed to this first edition of Diagnosis and Control of Alkali-Aggregate Reactions in Concrete. We are particularly grateful for reviews by the following individuals: Rachel Detwiler (Construction Technology Laboratories, Inc.), Wendell Dubberke (Iowa Department of Transportation), Bernard Erlin (The Erlin Company), Anthony E. Fiorato (Portland Cement Association), Richard D. Gaynor (National Ready Mixed Concrete Association), Colin Lobo (National Ready Mixed Concrete Association), Ward Malisch (The Aberdeen Group), Claudio Manissero (FMC Corporation), Bryant Mather (Army Corps of Engineers), Richard C. Meininger (National Aggregates Association), Jon Mullarky (National Ready Mixed Concrete Association), Matt Ross (Ash Grove Cement Company), Oscar Tavares (Holnam, Inc., also Chair of the PCA Subcommittee on ASR/Pavement Durability, and Chair of the ACPA Pavement Durability Subcommittee), and Gerald Voigt (American Concrete Pavement Association). This acknowledgment does not necessarily imply approval of the text by these individuals since final editorial prerogatives have necessarily been exercised by the authors.
CAUTION: Contact with wet (unhardened) concrete, mortar, cement, or cement mixtures can cause SKIN IRRITATION, SEVERE CHEMICAL BURNS, or SERIOUS EYE DAMAGE. Wear waterproof gloves, a longsleeved shirt, full-length trousers, and proper eye protection when working with these materials. If you have to stand in wet concrete, use waterproof boots that are high enough to keep concrete from flowing into them. Wash wet concrete, mortar, cement, or cement mixtures from your skin immediately. Flush eyes with clean water immediately after contact. Indirect contact through clothing can be as serious as direct contact, so promptly rinse out wet concrete, mortar, cement, or cement mixtures from clothing. Seek immediate medical attention if you have persistent or severe discomfort.

PCA R&D Serial No. 2071 American Concrete Pavement Association 5420 Old Orchard Road, Suite A100, Skokie, Illinois 60077-1059 (847) 966-2272
5 AM E
N CONC CA RI

EN

T ASSOC

National Aggregates Association 900 Spring Street, Silver Spring, Maryland 20910 (301) 587-1400
NATIONAL AGGREGATES ASSOCIATION
The National Aggregates Association is an organization of producers of aggregates, meeting the needs of the industry, members and their employees, customers and communities, by providing leadership in technology and education, providing a forum for information exchange and serving as the voice for the industry.

National Ready Mixed Concrete Association 900 Spring Street, Silver Spring, Maryland 20910 (301) 587-1400
NATIONAL READY MIXED CONCRETE ASSOCIATION
NRMCA provides exceptional value for its members by responsibly representing and serving the entire ready mixed concrete industry and all who benefit from the use of our product through leadership, promotion, education and partnering to ensure ready mixed concrete is the building material of choice.

Portland Cement Association

AT

IO N

A national organization with the mission to increase the use of concrete pavement in construction and rehabilitation of transportation facilities in North America, by continually providing a quality product that is safe, cost effective and environmentally sound.

5 TE RE

PAV

Printed in U.S.A. 24

5420 Old Orchard Road, Skokie, Illinois 60077-1083, (847) 966-6200, Fax (847) 966-9781

An organization of cement manufacturers to improve and extend the uses of portland cement and concrete through market development, engineering, research, education and public affairs work.

IS413.01T