Alkali-Aggregate Reactivity (Aar)

ALKALI-AGGREGATE REACTIVITY (AAR)
FACTS BOOK
1. Report No.
2. Government Accession No.
3. Recipients Catalog No.
FHWA-HIF-13-019
5. Report Date
March 2013
4. Title and Subtitle

Alkali-Aggregate Reactivity (AAR) Facts Book
6. Performing Organization Code

7. Author(s)
Thomas, M.D.A., Fournier, B., Folliard, K.J.
8. Performing Organization Report No.
9. Performing Organization Name and Address

The Transtec Group, Inc.
6111 Balcones Drive
Austin, TX 78731
10. Work Unit No.
12. Sponsoring Agency Name and Address

Office of Pavement Technology
Federal Highway Administration
1200 New Jersey Avenue, DE
Washington, DC 20590
13. Type of Report and Period Covered

Final Report March 2013
11. Contract or Grant No.

DTFH61-06-D-00035
14. Sponsoring Agency Code
15. Supplementary Notes

Contracting Officers Representative (COR): Gina Ahlstrom, HIAP-10
16. Abstract
This document provides detailed information on alkali-aggregate reactivity (AAR). It primarily discusses alkali-silica reaction (ASR), cover
ing the chemistry, symptoms, test methods, prevention, specifications, diagnosis and prognosis, and mitigation. Alkali-carbonate reaction
(ACR) is also addressed.
17. Key Words

Alkali-silica reactivity, alkali-aggregate reaction, reactive
aggregates, ASR, concrete durability, symptoms, identifica
tion, prevention, mitigation, lithium
9. Security Classif. (of this

report)
20. Security Classif. (of this

page)
18. Distribution Statement

No restrictions. This document is available to the public through the Na
tional Technical Information Service, Springfield, VA 22161.
21. No of Pages
211
22. Price
SI* (MODERN METRIC) CONVERSION FACTORS

APPROXIMATE CONVERSIONS TO SI UNITS
Symbol
When You Know
in
ft
yd
mi
inches
feet
yards
miles
Multiply By
LENGTH
25.4
0.305
0.914
1.61
To Find
Symbol
millimeters
meters
meters
kilometers
mm
m
m
km
square millimeters
square meters
square meters
hectares
square kilometers
mm
2
m
2
m
ha
2
km
AREA
2
in
2
ft
2
yd
ac
2
mi
square inches
square feet
square yard
acres
square miles
fl oz
gal
ft3
3
yd
fluid ounces
gallons
cubic feet
cubic yards
oz
lb
T
ounces
pounds
short tons (2000 lb)
Fahrenheit
645.2
0.093
0.836
0.405
2.59
VOLUME
29.57
milliliters
3.785
liters
0.028
cubic meters
0.765
cubic meters
NOTE: volumes greater than 1000 L shall be shown in m3
mL
L
m3
3
m
MASS
28.35
0.454
0.907
grams
kilograms
megagrams (or "metric ton")
g
kg
Mg (or "t")
TEMPERATURE (exact degrees)

F
5 (F-32)/9
or (F-32)/1.8
Celsius
lux
candela/m2
lx
cd/m2
ILLUMINATION
fc
fl
foot-candles
foot-Lamberts
10.76
3.426
lbf
lbf/in2
poundforce
poundforce per square inch
FORCE and PRESSURE or STRESS

4.45
6.89
newtons
kilopascals
N
kPa
APPROXIMATE CONVERSIONS FROM SI UNITS

Symbol
When You Know
mm
m
m
km
millimeters
meters
meters
kilometers
2
Multiply By
LENGTH
0.039
3.28
1.09
0.621
To Find
Symbol
inches
feet
yards
miles
in
ft
yd
mi
square inches
square feet
square yards
acres
square miles
in
ft2
2
yd
ac
mi2
fluid ounces
gallons
cubic feet
cubic yards
fl oz
gal
3
ft
3
yd
ounces
pounds
short tons (2000 lb)
oz
lb
T
AREA
mm
m2
2
m
ha
km2
square millimeters
square meters
square meters
hectares
square kilometers
mL
L
3
m
3
m
milliliters
liters
cubic meters
cubic meters
g
kg
Mg (or "t")
grams
kilograms
megagrams (or "metric ton")
Celsius
0.0016
10.764
1.195
2.47
0.386
VOLUME
0.034
0.264
35.314
1.307
MASS
0.035
2.202
1.103
TEMPERATURE (exact degrees)

C
1.8C+32
Fahrenheit
foot-candles
foot-Lamberts
fc
fl
ILLUMINATION
lx
2
cd/m
lux
2
candela/m
N
kPa
newtons
kilopascals
0.0929
0.2919
FORCE and PRESSURE or STRESS

0.225
0.145
poundforce
poundforce per square inch
lbf
2
lbf/in
*SI is the symbol for the International System of Units. Appropriate rounding should be made to comply with Section 4 of ASTM E380.
(Revised March 2003)
Table of Contents
1 Introduction .............................................................................................................................1
1.1 DEFINITIONS.....................................................................................................................1
1.2 HISTORY OF AAR.............................................................................................................2
1.3 AAR TODAY .......................................................................................................................4
REFERENCES ..........................................................................................................................8
2 Fundamentals of Alkali-Silica Reaction ................................................................................9
2.1 INTRODUCTION ...............................................................................................................9
2.2 CHEMISTRY OF THE REACTION ..............................................................................10
2.3 MECHANISM OF EXPANSION ....................................................................................13
2.4 REACTIVE SILICA .........................................................................................................15
2.4.1 Types of Reactive Silica ...............................................................................................15
2.4.2 Effect of Aggregate Size ...............................................................................................18
2.4.3 Pessimum Effect ........................................................................................................... 20
2.5 SOURCES OF ALKALI ...................................................................................................22
2.5.1 Pore Solution Composition ...........................................................................................22
2.5.2 Threshold Alkali Contents ............................................................................................25
2.5.3 Alkali Recycling ........................................................................................................... 27
2.6 ROLE OF MOISTURE.....................................................................................................29
REFERENCES ........................................................................................................................31
3 Symptoms of Alkali-Silica Reaction (ASR).........................................................................35
3.1 INTRODUCTION .............................................................................................................35
3.2 CRACKING .......................................................................................................................35
3.3 EXPANSION CAUSING MOVEMENTS AND DEFORMATIONS...........................40
3.4 POP-OUTS .........................................................................................................................43
3.5 SURFACE DEPOSITS (GEL EXUDATION VS. EFFLORESCENCE).....................45
REFERENCES ........................................................................................................................46
4 Test Methods..........................................................................................................................47
4.1 INTRODUCTION .............................................................................................................47
4.2 ASR TEST METHODS FOR EVALUATING AGGREGATE REACTIVTY ...........47
4.2.1 ASTM C 295: Standard Guide for Petrographic Examination of Aggregates for Con
crete........................................................................................................................................ 48
4.2.2 ASTM C 289: Standard Test Method for Potential Alkali-Silica Reactivity of Aggre
gates (Chemical Method) ....................................................................................................... 50
4.2.3 ASTM C 227: Standard Test Method for Potential Alkali Reactivity of Cement-
Aggregate Combinations (Mortar-Bar Method) .................................................................... 51
4.2.4 AASHTO T 303 (ASTM C 1260): Standard Test Method for Potential Alkali Reactiv
ity of Aggregates (Mortar-Bar Method) ................................................................................ 51
4.2.5 ASTM C 1293: Standard Test Method for Concrete Aggregates by Determination of
Length Change of Concrete Due to Alkali-Silica Reaction................................................... 54
4.3 ASR TEST METHODS FOR EVALUATING PREVENTIVE MEASURES ............ 58
4.3.1 ASTM C 441: Standard Test Method for Effectiveness of Mineral Admixtures or
Ground Blast-Furnace Slag in Preventing Excessive Expansion of Concrete Due to the Al

kali-Silica Reaction ................................................................................................................ 59
4.3.2 ASTM C 1567: Standard Test Method for Determining the Potential Alkali-Silica Re
activity of Combinations of Cementitious Materials and Aggregate (Accelerated Mortar-Bar
Method).................................................................................................................................. 61
Length Change of Concrete Due to Alkali-Silica Reaction................................................... 63
4.3.4 Other ASR Test Methods.............................................................................................. 64
4.4 ACR TEST METHODS FOR EVALUATING AGGREGATE REACTIVITY......... 66
4.5 SUMMARY........................................................................................................................ 69
REFERENCES ........................................................................................................................ 70
5 Prevention of Alkali-Silica Reaction.................................................................................... 75
5.1 INTRODUCTION ............................................................................................................. 75
5.2 PREVENTIVE MEASURES OPTIONS...................................................................... 75
5.3 USE OF NON-REACTIVE AGGREGATES ................................................................. 76
5.4 LIMITING THE ALKALI CONTENT OF THE CONCRETE................................... 76
5.5 USE OF SUPPLEMENTARY CEMENTING MATERIALS ...................................... 81
5.5.1 Effect of SCM on the Availability of Alkalis ............................................................... 83
5.5.2 Effect of SCM on the Expansion of Concrete .............................................................. 91
5.5.3 Summary on Effect of SCM on ASR ......................................................................... 101
5.6 USE OF LITHIUM ......................................................................................................... 102
REFERENCES ...................................................................................................................... 104
6 Alkali-Aggregate Reactions: Specifications...................................................................... 111
6.1 INTRODUCTION ........................................................................................................... 111
6.2 EXISTING SPECIFICATIONS IN USA ...................................................................... 111
6.2.1 ASTM ......................................................................................................................... 111
6.2.2 ACI ............................................................................................................................. 113
6.2.3 State Specifications ..................................................................................................... 114
6.3 AASHTO PP65-11 ........................................................................................................... 114
6.3.1 Evaluating Aggregate Reactivity ................................................................................ 114
6.3.2 Selecting Preventive Measures ................................................................................... 117
REFERENCES ...................................................................................................................... 127
7 Diagnosis and Prognosis of Alkali-Silica Reaction in Concrete Structures................... 131
7.1 INTRODUCTION ........................................................................................................... 131
7.2 MANAGEMENT PROGRAM FOR ASR-AFFECTED STRUCTURES ................. 131
7.2.1 ASR Investigation Program Step 1 - Diagnosis ......................................................... 132
7.2.2 ASR Investigation Program Step 2 - Prognosis .......................................................... 153
7.3 CONCLUSION ................................................................................................................ 169
REFERENCES ...................................................................................................................... 171
8 Mitigation Methods for ASR-Affected Structures ........................................................... 177
8.1 INTRODUCTION ........................................................................................................... 177
8.2 OVERVIEW OF MITIGATION METHODS.............................................................. 177
8.3 REDUCING INTERNAL RELATIVE HUMIDITY ................................................... 179
8.4 APPLICATION OF LITHIUM COMPOUNDS .......................................................... 187
8.5 APPLICATION OF EXTERNAL RESTRAINT ......................................................... 192
8.6 STRESS RELIEF ............................................................................................................ 193
8.7 SUMMARY...................................................................................................................... 194
REFERENCES ...................................................................................................................... 195
9 Alkali-Carbonate Reaction................................................................................................. 197
9.1 INTRODUCTION ........................................................................................................... 197
9.2 CHARACTERISTICS OF ALKALI-CARBONATE REACTIVE ROCKS............. 198
9.3 CHEMISTRY OF ALKALI-CARBONATE REACTION.......................................... 200
9.4 MECHANISMS OF EXPANSION DUE TO ACR ...................................................... 201
9.5 FACTORS CONTRIBUTING TO EXPANSION DUE TO ALKALI-CARBONATE
REACTION............................................................................................................................ 202
9.6 DIFFERENCES BETWEEN ACR AND ASR ............................................................. 203
9.7 PREVENTION OF EXPANSION DUE TO ALKALI-CARBONATE REACTION204

9.8 SUMMARY OF ALKALI-CARBONATE REACTION ............................................. 206
REFERENCES ...................................................................................................................... 208
AAR Facts Book
Chapter 1 Introduction
1 Introduction
1.1 DEFINITIONS
Alkali-aggregate reaction (AAR) is a reaction in concrete between the alkali hydroxides, which
originate mainly from the portland cement, and certain types of aggregate. Two types of AAR
are currently recognized; these are alkali-silica reaction (ASR) and alkali-carbonate reaction
(ACR). As the names imply, these types of reaction differ in that they involve reactions with
either siliceous or carbonate phases in the aggregates. The following definitions can be found
on the website of the American Concrete Institute (ACI)1:
Alkali-aggregate
reaction, AAR
Chemical reaction in either concrete or mortar between hydroxyl

ions (OH-) of the alkalies (sodium and potassium) from hydraulic
cement (or other sources), and certain constituents of some
aggregates; under certain conditions deleterious expansion of
concrete or mortar may result.
Alkali-carbonate
reaction, ACR

cement (or other sources) and certain carbonate rocks,
particularly calcitic dolostone and dolomitic limestones, present
in some aggregates. The reaction is usually accompanied by
dedolomitization and expansion of the affected aggregate
particles, leading to abnormal expansion and cracking of concrete
in service.
Alkali-silica
reaction, ASR

cement (or other sources), and certain siliceous rocks and
minerals, such as opal, chert, microcrystalline quartz, and acidic
volcanic glass, present in some aggregates. This reaction and the
development of the alkali-silica gel reaction product can, under
certain circumstances, lead to abnormal expansion and cracking
of the concrete.
www.concrete.org
AAR Facts Book
ASR is far more widespread than ACR and is the main focus of this text. However, it should be
noted that some test methods used to detect alkali-silica reactive aggregates may fail to detect
alkali-carbonate reactive aggregates. In addition, measures used to prevent damaging ASR are
generally ineffective in preventing ACR expansion and, consequently, alkali-carbonate reactive
rocks should not be used in the production of concrete.
1.2 HISTORY OF AAR

Problems due to ASR were first identified in the State of California in the 1930s (see Figure
1.1) and reported by Thomas Stanton of the California State Division of Highways in 1940
(Stanton 1940). Stantons studies demonstrated that the expansion of mortar bars was
influenced by the alkali content of the cement, the type and amount of the reactive silica in the
aggregate, the availability of moisture, and temperature. He further showed that expansion was
negligible when the alkali content of the cement was below 0.60% Na2Oe and that expansion
could be reduced by pozzolans, thus setting the groundwork for preventive measures.
Subsequent to Stantons discovery, ASR was diagnosed as the cause of abnormal cracking in a
number of dams operated by the U.S. Bureau of Reclamation, such as the Parker Dam in
Arizona (Meissner 1941), and in the 1940s a number of agencies initiated studies on ASR in the
USA (Army Corps of Engineers, Bureau of Public Roads, Portland Cement Association) and
other countries (Denmark and Australia). ASR is now recognized as a major cause of concrete
deterioration in the USA, incidences having occurred in most, if not all, of the contiguous
states, and numerous countries worldwide.
AAR Facts Book
Figure 1.1. Thomas Stanton of the California State Division of Highways and a Bridge Parapet
Wall that is Showing Signs of Damage due to Alkali-Silica Reaction
Alkali carbonate reaction (ACR) was first discovered by Swenson (1957a) as the cause of
concrete deterioration in Canada at about the same time that ASR was first documented in the
same country (Swenson 1957b). ACR was subsequently implicated in cases of degradation of
concrete structures in the USA (Hadley 1961) and alleged cases of ACR have now occurred in
Virginia, West Virginia, Kentucky, Missouri, Tennessee, Iowa, Illinois, Indiana, and New
York, as well as England, Bahrain, Iraq, and China (Ozol 2006). However, unlike ASR,
problems with ACR are still restricted to a few isolated locations worldwide. Consequently,
there has been comparatively little research conducted on this topic.
A series of international conferences on alkali-aggregate reaction (ICAAR) in concrete began in
1974 (see Table 1.1). The first conference was held in Kge, Denmark in 1974 with 23
delegates presenting 13 papers and representing just 5 countries (Denmark, Germany, Iceland,
U.K., USA). Interest in AAR grew rapidly from this time and in 1992 over 300 delegates
representing 29 countries attended the 9th ICAAR in London, U.K., and 150 papers were
3
AAR Facts Book
published in the proceedings. Interest has remained at this level since that time with the most
recent conference being held in Austin, Texas, USA, in 2012 (131 papers from 27 countries).
Table 1.1. International Conferences on Alkali-Aggregate Reaction in Concrete
Year Host
Year Host
1974 Kge, Denmark
1989
Kyoto, Japan
1975 Reykjavik, Iceland
1992
London, U.K.
1976 London, U.K.
10
1996
Melbourne, Australia
1978 West Lafayette, USA
11
2000
Quebec City, Canada
1981 Cape Town, S. Africa
12
2004
Beijing, China
1983 Copenhagen, Denmark
13
2008
Trondheim, Norway
1986 Ottawa, Canada
14
2012
Austin, U.S.A.
Alkali-silica reaction is now widely recognized as one of the more prevalent deterioration
mechanisms affecting concrete worldwide2.
1.3 AAR TODAY

Seventy years after ASR was first documented, much is now known about the chemistry of the
reaction, the factors that contribute to the reaction and expansion, methods for testing
aggregates, and strategies for preventing expansion. Proper application of the knowledge
available today to new concrete construction should result in a very low risk of damage due to
ASR occurring in the normal service life of the structure. A number of specifications or
practices have been developed in recent years to assist the practitioner in the selection of
materials and preventive measures for ensuring durable construction (with regards to AAR).
This text reviews one such practice, AASHTO PP65, in detail (see chapter 6).
It is well established that the ASR results from a reaction between the alkali hydroxides
provided (mainly) by the portland cement and certain types of reactive silica minerals present in
some aggregates, and that limiting the availability of one (or both) of these is an effective
2
A comprehensive review of the history of alkali-aggregate reaction from 1940 to 1996 is given by Idorn (1997).
AAR Facts Book
means of preventing deleterious expansion due to ASR. Thus, selecting non-deleteriously

reactive aggregates or using low-alkali cement have become common practices for
preventing ASR, although it is now considered that controlling the alkali content of the concrete
is more appropriate than merely limiting the alkali content of the cement.
The potential for using pozzolans to control damaging ASR was demonstrated by Stanton
(1940) in his landmark paper that first revealed the phenomenon of alkali-silica reaction to the
concrete community. The use of pozzolans for this purpose was first put into practice in the
same decade when calcined clay was used to prevent ASR in the Davis Dam (Gilliland and
Moran 1949), which was constructed between 1942 and 1950, the reaction having been
implicated as the cause of cracking in the Parker Dam (Meissner 1941), which was completed
shortly before construction began on the Davis Dam and is located 88 miles (141 km) upstream
on the Colorado River. Ten years after Stantons (1940) discovery of ASR the potential for
using fly ash and slag for controlling expansion was first documented, and it is now widely
accepted that supplementary cementing materials are an effective means for controlling ASR
expansion provided they are used at a sufficient level of replacement (see chapter 5). The longterm field performance of fly ash in the role of ASR prevention was recently documented
(Thomas et al. 2012) in the form of the excellent condition of the 50-year-old Nant-y-Moch
Dam in Wales (Figure 1.2) and the 40-year-old Lower Notch Dam in Canada, both structures
being built with the combination of fly ash and highly reactive aggregates.
AAR Facts Book
Figure 1.2. The Nant-y-Moch Dam in 2011 No Symptoms of ASR after 50 Years
Constructed with Reactive Aggregate and 25% Fly Ash
Test methods for correctly identifying reactive aggregates and evaluating the efficacy of
preventive measures have constantly evolved since Stantons (1940) mortar-bar test, which was
a precursor to the standard ASTM C 227 method. At the current time practitioners face the
dilemma of selecting between tests that are reliable but have (often unacceptably) a long test
duration or rapid tests that often have a poor correlation with field performance. The search is
still on for rapid and reliable test methods. Testing methods for ASR and ACR are discussed in
chapter 4.
There are few options available for mitigating ongoing AAR in existing structures. In other
words, once concrete has alkali-aggregate reaction, it is very difficult to stop the reaction. In
some cases, it may be possible reduce the availability of moisture and slow the reaction down.
In other cases, methods have been developed to either physically confine the expansion or
create space to allow for expansion and relieve stresses. Chapters 7 and 8 in this text discuss
AAR Facts Book
methodologies for evaluating existing concrete structures to (a) determine the extent of ASR
and its impact on the concrete, (b) predict the future growth of the concrete due to ASR, and (c)
select appropriate strategies for mitigating the effects of ASR.
There is comparatively little information on ACR, and consensus has yet to be reached on the
exact mechanisms of expansion. Although it is agreed that alkali-carbonate reactive dolomitic
limestones have a characteristic texture and composition and undergo a chemical reaction
resulting in dedolomitization (dolomite brucite + calcite), there is disagreement as to
whether the accompanying expansion results from this reaction or from reaction of
cryptocrystalline silica in the limestone (i.e., ACR expansion may be due to ASR). There does
appear to be consensus that, regardless of the true mechanism of expansion, there are features
of the alkali-carbonate reaction that set it apart from ASR with aggregates that are undisputedly
alkali-silica reaction. These features include (a) the relatively short timeframe before damage is
observed, (b) reaction (and expansion) at very low alkali levels, (c) the general ineffectiveness
of pozzolans and slag in controlling expansion, and (d) the inability of certain tests to identify
the reactive aggregates. ACR is discussed in chapter 9 and a case is made for treating these
reactive rocks as a separate category of AAR, irrespective of whether expansion is due to ACR
or ASR.
AAR Facts Book
REFERENCES
AASHTO. 2011. Standard Practice for Determining the Reactivity of Concrete Aggregates
and Selecting Appropriate Measures for Preventing Deleterious Expansion in New Concrete
Construction. PP65-11, American Association of State and Highway Transportation Officials,
Washington, DC, 24 p.
ASTM. 2010. Standard Test Method for Potential Alkali Reactivity of Cement-Aggregate
Combinations (Mortar-Bar Method). ASTM C 227, ASTM International, West Conshohocken,
PA, 6 p.
Gilliland, J.L. and Moran, W.T. 1949. Siliceous admixture specified for Davis Dam concrete.
Engineering News Record, February, 62-64.
Hadley, D.W. 1961. Alkali reactivity of carbonate rocks; expansion and dedolomitization.
Proceedings of the Highway Research Board, 40, 462-474.
Idorn, G. 1997. Concrete progress: from antiquity to the third millennium. Thomas Telford,
London, 359p.
Meissner, H.S. 1941. Cracking in concrete due to expansive reaction between aggregate and
high-alkali cement as evidenced in Parker Dam. Proceedings of the American Concrete
Institute, 57, 549-568.
Stanton, T.E. 1940. Expansion of concrete through reaction between cement and aggregate.
Proceedings of the American Society of Civil Engineers, 66(10): 1781-1811.
Swenson, E.G. 1957a. A reactive aggregate undetected by ASTM tests. Proceedings of
American Society for Testing and Materials, 57, 48-51.
Swenson, E.G. 1957b. Cement aggregate reaction in concrete of a Canadian bridge. ASTM
Proceedings, 57 10431056.
Thomas, M.D.A., Hooton, R.D., Rogers, C. and Fournier, B. 2012. 50 years and still going
strong: fly ash puts paid to ASR. Concrete International, January, 35-40.
AAR Facts Book
Chapter 2 Fundamentals of ASR
2 Fundamentals of Alkali-Silica Reaction
2.1 INTRODUCTION
As described in the introduction, alkali-silica reaction is a reaction between the alkali
hydroxides in the pore solution of concrete (or mortar) and certain types of silica minerals
present in some aggregates. The reaction product, an alkali-silica gel with varying amounts of
calcium, is hygroscopic having a tendency to absorb water and swell. Under certain conditions
the reaction causes expansion of the concrete and may eventually lead to cracking. A schematic
showing the sequence of events is shown in Figure 2.1.
It is clear from this brief description of ASR that there are three requirements for damaging
reaction to occur; these are:
A sufficient quantity of reactive silica (within aggregates)
A sufficient concentration of alkali (primarily from portland cement)
Sufficient moisture
Elimination of any one of these requirements will prevent the occurrence of damaging alkalisilica reaction; this is discussed further in the sections dealing with prevention and mitigation of
ASR.
This chapter describes the chemistry of the reaction and the mechanisms of expansion, and
discusses the role of various contributing factors including the composition and nature of the
reactive aggregate, sources of alkali, and the impact of the exposure conditions.
Alkali-carbonate reaction (ACR) is discussed in chapter 9.
AAR Facts Book
Figure 2.1. Alkali-Silica Reaction Sequence
2.2 CHEMISTRY OF THE REACTION

Despite the term, alkali-silica reaction is initiated by a reaction between the hydroxyl ions in the
pore solution and certain types of silica in the aggregate. Silica (SiO2) is composed primarily of
siloxane groups (Si-O-Si) but even crystalline silica is disordered at the surface and the
10
AAR Facts Book
surface oxygens are hydroxylated (even in pure water) forming silanol groups (Si-OH); see
Figure 2.2. In the presence of a high concentration of hydroxyl ions (OH-) silica tends towards
dissolution first by neutralization of the silanol groups and then by attack on the siloxane
groups; the reactions may be represented as follows3:
Si-OH + OH- Si-O- + H2O
Eqn. 2.1
Si-O-Si + 2OH- 2Si-O- + H2O
Eqn. 2.2
Figure 2.2. The Structure of Silica

The structure breaks down progressively as the siloxane bridges are broken (see Figure 2.3).
The negatively charged Si-O- ions attract positively charged species such as sodium (Na+) and
potassium (K+), which are abundant in concrete pore solution. The initial result is an alkalisilicate solution or gel depending on the moisture content. However, in the presence of calcium,
the silica precipitates from solution as an alkali-silicate gel (CaO-Na2O/K2O-SiO2-H2O),
primarily composed of sodium, potassium, and silica, with small amounts of calcium. The
hydroxyl ion concentration (and the pH) decreases as silica dissolves. In a system composed
solely of alkali hydroxide and silica, the silica will continue to dissolve until the concentration
in solution reaches the silica-pH equilibrium curve (see Figure 2.4). The data show that final
The reactions may as also be written as follows (modified from Dent Glasser and Kataoka 1981a):
Si-OH + OH- + Na+ Si-O-Na + H2O
Eqn. 2.1a
Si-O-Si + 2OH- + 2Na+ 2(Si-O-Na) + H2O
Eqn. 2.2a
11
AAR Facts Book
silica concentration depends on the initial SiO2/Na2O ratio as will be discussed later in the
section on pessimum behavior. The situation is more complex in concrete, probably due to
the presence of abundant calcium, which reduces silica concentrations in solution and provides
an additional source of hydroxyl ions. Consequently, equilibrium conditions are reached slowly
in concrete (and mortar).
Figure 2.3. Mechanism of Dissolution of Silica due to Attack by Hydroxyl Ions. Dotted Line
Represents the Interface between Silica and Water (Iler 1979; Urhan 1987)
Figure 2.4. Equilibrium Solubility Curve (SiO2-pH) and Change in Composition for Different
Solutions of SiO2-NaOH (from Dent Glasser and Kataoka 1981b)
The chemical composition and morphology of the reaction product (gel) vary widely as shown
by Moranville-Regourd (1989), and other researchers have shown that the physical properties of
12
AAR Facts Book
gels, in terms of the capacity for unconfined swelling or for exerting forces when confined, also
vary considerably.
2.3 MECHANISM OF EXPANSION

Figure 2.5 shows a petrographic thin section of ASR-affected concrete viewed under a
microscope with transmitted light. The image shows a flint sand particle (at left) embedded in
cement paste. The flint particle has reacted, expanded, and cracked, and the crack extends out
into the surrounding cement paste. The crack within the aggregate and in the paste is partially
filled with a reaction product, alkali-silica gel.
Figure 2.5. Photograph of a Thin Section of Concrete Undergoing ASR (field of view ~ 5 mm)
Despite general acceptance of the chemical reactions involved, a number of different
mechanisms of expansion have been proposed as follows:
Hansen (1944) suggested that the cement paste surrounding reactive grains acts as a
semi-permeable membrane through which water (or pore solution) may pass but not the
larger complex silicate ions. The water is drawn into the reacting grain where its
chemical potential is lowest. An osmotic pressure cell is formed and increasing
hydrostatic pressure is exerted on the cement paste, inevitably leading to cracking of the
surrounding mortar.
McGowan and Vivian (1952) disputed the classical osmotic theory on the basis that
cracking of the surrounding cement paste membrane due to ASR would relieve
hydraulic pressure and prevent further expansion. They proposed an alternative
13
AAR Facts Book
mechanism based on the physical absorption of water by the alkali silica-gel and
subsequent swelling of the gel.
Powers and Steinour (1955a, 1955b) proposed a compromise, suggesting that both
osmotic and imbibition pressures may be generated depending on whether the alkalisilicate complex is fluid or solid. In their hypothesis, the reaction product itself may act
as a semi-permeable membrane depending on its composition.
Regardless of the mechanism, the fundamental cause of swelling is thermodynamically the
same, i.e., the entry of water into a region where the effect of a solute or of adsorption reduces
its free energy.
A number of experimental studies have concluded that significant expansion only occurs when
an adequate supply of calcium is available as calcium hydroxide [Ca(OH)2]. In systems with
abundant alkali hydroxides and reactive silica, but no calcium hydroxide, silica dissolves and
remains in solution. Although the precise role calcium plays in gel expansion remains unclear, a
series of mechanisms have been proposed, as listed below (ACI 201):
Calcium promotes alkali recycling by replacing alkalis in the reaction product thereby
regenerating alkalis for further reaction (Thomas 2001; Hansen 1944). Alkali recycling
is discussed further in section 2.5.3 of this chapter.
Ca(OH)2 provides a reservoir of OH- ions to maintain a high level of OH- in solution
(Wang and Gillott 1991).
High calcium concentrations in the pore solution prevent the diffusion of silica away
from reacting aggregate particles (Chatterji 1979; Chatterji and Clausson-Kass 1984).
If calcium is not available, reactive silica dissolves in alkali hydroxide solution without
causing damage (Thomas 1998; Diamond 1989).
The formation of calcium-rich gels is necessary to cause expansion either directly or
indirectly through the formation of a semi-permeable membrane around reactive
aggregate particles (Thomas 1998; Thomas et al. 1991; Bleszynski and Thomas 1998).
Pozzolans are effective in controlling the expansion of concretes (and mortars)
immersed in alkaline salt (including alkali hydroxide) solutions (Chatterji et al. 1987;
Kawamura et al. 1988; Alasali and Malhotra 1991; Bleszynski and Thomas 1998) and
this is attributed to the consumption of Ca(OH)2 by the pozzolanic reaction which
reduces the availability of calcium for the alkali-silica reaction. Bleszynski and Thomas
(1998) showed considerable evidence of reaction in concrete containing 40% fly ash and
a reactive flint sand after 4 years immersion in 1 NaOH at 80C, but no expansion.
Alkali-silica gel was observed to be migrating into and saturating the cement paste
surrounding reactive flint particles with no signs of damage (see Figure 2.4). They
14
AAR Facts Book
attributed this to the very low viscosity of the reaction product formed in the absence of
calcium. By contrast, other workers (Tang et al. 1983; Wang and Gillott 1991) showed
that the addition of CaO can induce expansion on concretes containing fly ash or silica
fume.
Expansion of mortars or concretes containing reactive aggregate can be prevented by
removal of the Ca(OH)2 prior to immersion of the sample in alkaline solutions. Chatterji
(1979) removed the Ca(OH)2 from concrete by leaching, and Thomas (2000) carbonated
the Ca(OH)2 by exposing mortar bars to an enriched-CO2 environment.
Although the precise role(s) of Ca(OH)2 is equivocal, it is evident that calcium must be
available for damaging reaction to occur. Thus reducing the availability of calcium for
example, by consuming Ca(OH)2 through pozzolanic reactions should result in a reduction in
the expansion due to ASR.
Dependence on Ca(OH)2 for the promotion of damaging AAR is not a recent phenomenon.
Conrow (1952) suggested that the expansion of concretes containing a siliceous sand-gravel
may be related to the quantity of Ca(OH)2 produced by the cement and that the beneficial effect
of pozzolan is related to its ability to react with Ca(OH)2. In a discussion of this paper, Mather
stated an observation that concrete that has undergone ASR is characterized by materially
reduced quantities of crystalline calcium hydroxide and suggested that Ca(OH)2 may be
consumed by ASR and that ... the mere consumption of calcium hydroxide by reaction with a
pozzolan is sufficient to explain the beneficial effects of pozzolans in preventing abnormal
expansion.
2.4 REACTIVE SILICA

2.4.1 Types of Reactive Silica
Table 2.1 provides a list of deleteriously reactive rocks, minerals, and synthetic substances.
15
AAR Facts Book
Table 2.1. Deleteriously Reactive Rocks, Minerals and Synthetic Substances

(from ACI Committee 201, 1991)
Reactive substance (mineral) Chemical composition
Physical character
Opal
SiO2nH2O
Amorphous
Chalcedony
SiO2
Microcrystalline to
cryptocrystalline; commonly
fibrous
Certain forms of quartz
SiO2
Microcrystalline to
cryptocrystalline; crystalline,
but intensely fractured,
strained, and/or inclusionfilled
Cristobalite
SiO2
Crystalline
Tridymite
SiO2
Crystalline
Rhyolitic,dacitic, latitic, or
andesite glass or
cryptocrystalline
devitrification products
Siliceous with lesser

proportions of Al2O3, Fe2O3,
alkaline earths and alkalis
Glass or cryptocrystalline
material as the matrix of
volcanic rocks or fragments in
tuffs
Synthetic siliceous glass
Siliceous, with lesser

proportions of alkalis, Al2O3,
and/or other substances
Glass
The most important deleteriously alkali-reactive rocks (that is, rocks containing excessive
amounts of one or more of the substances listed above) are as follows:
Opaline cherts
Rhyolites and tuffs
Opaline concretions
Chalcedonic cherts
Dacites and tuffs
Quartzose cherts
Andesites and tuffs
Fractured, strained, and

limestone-filled quartz and
quartzites
Siliceous limestones
Siliceous shales
Siliceous dolomites
Phylites
NOTE: A rock may be classified as, for example siliceous limestone, and be innocuous if its
siliceous constituents are other than those indicated above.
16
AAR Facts Book
Silica, SiO2, is a component of many rocks; however, not all forms of silica react significantly
with the pore solution of concrete and, thus, not all siliceous aggregates produce damaging
ASR. For example, the mineral quartz is stable whereas the mineral opal is highly reactive,
although both are silica minerals with similar chemical composition, being primarily composed
of SiO2. Figure 2.6 shows the differences in the solubility of the silica and the crystalline
structure of these two minerals. Opal has a highly disordered (amorphous) structure which
renders it unstable at high pH and, as such, aggregates containing significant quantities of the
mineral opal may be expected to react and result in expansion when used in concrete, provided
there is sufficient alkali present. On the other hand, quartz will not react deleteriously regardless
of the alkali content of the concrete4.
Figure 2.6. Solubility and Structure of Opal and Quartz

The following silica minerals are considered to be alkali-silica reactive: opal, tridymite,
cristobalite, volcanic glass, chert, cryptocrystalline (or microcrystalline) quartz, and strained
quartz. These minerals may be found in the following rock types: shale, sandstone, silicified
carbonate rocks, chert, flint, quartzite, quartz-arenite, gneiss, argillite, granite, greywacke,
4
The potential for reaction exists even with well-crystallized silica (e.g. quartz), but the reaction is slow and may
be considered to be negligible with regards to the design life of a concrete structure. However, temperature
dramatically increases the rate of reaction and, in hydrothermal conditions, finely ground quartz is extremely
reactive.
17
AAR Facts Book
siltstone, arenite, arkose, and hornfels. However, this does not mean that all sources of such
rocks will produce deleterious reaction when used in concrete. For example, granitic aggregate
is widely used in concrete and only certain sources produce damaging ASR. The reactivity of a
rock depends on the type and quantity of reactive minerals present, if any.
Although, the presence of reactive minerals can usually be detected by an experienced
petrographer, appropriate performance testing of specific aggregate sources is recommended to
confirm alkali-silica reactivity. Test methods are discussed in chapter 4.
Poorly crystalline or amorphous silica minerals such as opal, cristobalite, volcanic and artificial
glasses react rapidly and may cause damaging reaction in a few years when present in amounts
as little as 1%. Varieties of quartz such as cryptocrystalline, microcrystalline, or strained quartz
react more slowly, take longer to produce damage, and are generally required to be present in
greater quantities than poorly crystalline and amorphous forms of silica. However, it is difficult
to classify aggregate reactivity based solely on mineralogy as aggregates may contain various
types of reactive minerals in different quantities, and the extent to which reactive minerals
cause damage in concrete depends on other factors such as particle size.
2.4.2 Effect of Aggregate Size
Stanton (1940) showed in his formative work on ASR that the particle size of the reactive
aggregate could have a profound effect on the magnitude of expansion of mortar. Figure 2.7(a)
shows the expansion of mortar bars containing a reactive siliceous magnesian limestone
partitioned into various size ranges. As expected, the expansion at a given age was greater for
mortars produced with the 180-600 micron fraction than for the larger fractions (0.6-2 mm and
2-6.7 mm), especially at earlier ages, and this is attributed to the greater surface area and
increased accessibility of the reactive silica. What was surprising was that the mortar bars
produced with reactive aggregate passing 180 micron did not exhibit any deleterious reaction.
Stanton (1940) hypothesized that the absence of expansion when the aggregate is in a finelydivided form is due either to (a) the reaction being dissipated throughout the mass in such a
way as to cause no high expansive forces or (b) the reaction being essentially complete before
the mortar sets. Other workers have since shown similar data although the particle size to which
the reactive aggregate has to be reduced to prevent expansion has not been established
unequivocally. For example, Vivian (1951) using the same reactive aggregate as Stanton (1940)
showed that expansion was delayed but not prevented for aggregate passing the 70-m but
retained 50-m sieve, whereas no expansion was observed for aggregate passing the 50-m
sieve. Vivian (1951) also showed similar trends in behavior with opal as the reactive aggregate.
18
AAR Facts Book
Figure 2.7. Effect of Aggregate Size (a) and Proportion (b) on
Expansion due to ASR (Stanton 1940)
Numerous laboratory studies have shown that the expansion of mortar or concrete containing
highly-reactive recycled glass as an aggregate can be eliminated provided the glass is crushed to
a certain fineness (Carpenter and Cramer 1999; Shayan and Xu 2004; Jin et al. 2000). In fact
glass crushed to high fineness may be effective in preventing damaging alkali-silica reaction
when larger particles of the same glass are used as (reactive) aggregates (Shayan and Xu 2004),
although such an approach may not be effective if the alkali content of the glass is high. Jin et
al. (2000) suggested that the pessimum size, that is the size fraction that causes the greatest
expansion when tested in mortar, reduces as the reactivity of the glass increases.
The strategy of using ground reactive aggregate as a preventive measure for ASR has been
employed in Iceland where up to 9% ground rhyolite has been interground with portland
cement clinker to produce a blended cement for controlling expansion with reactive rhyolitic
aggregates (Asgeirsson and Gudmundsson 1979). The use of finely-divided, alkali-silica
reactive materials (e.g., pozzolans) to control ASR is discussed in detail in chapter 5.
It should be noted that there is some evidence that reactive silica may cause damaging reaction
even when it is finer than 30 m. Diamond and Thaulow (1974) observed expansion of mortar
bars containing a high-alkali cement and reactive opal in the size range 2030 m (the fraction
was separated by sedimentation). However, to the authors knowledge, this is the only reported
case of damaging ASR with reactive silica particles in this size range (e.g., < 30 m) and such
behavior has not been confirmed by subsequent studies.
19
AAR Facts Book
2.4.3 Pessimum Effect

Another phenomenon observed by Stanton (1940) in his landmark paper was that expansion did
not necessarily increase with an increase in the reactive aggregate content. Figure 2.7(b) shows
the expansion of mortar bars stored over water as a function of the proportion of reactive
aggregate, in this case a siliceous magnesian limestone, in the mix. The maximum expansion
occurred when 20% of the aggregate was comprised of the reactive sand, and the expansion
decreases with greater amounts of reactive aggregate such that mixes containing 60% or more
of reactive sand showed no significant expansion. This phenomenon is known as the pessimum
effect, and such behavior is associated with some, but not all, highly reactive aggregates such as
Beltane opal (an aggregate used in many earlier research studies on ASR).
Figure 2.8(a) shows pessimum behavior for six aggregates (five different types) containing
opaline material, and Figure 2.8(b) shows the four regions of behavior described by Hobbs
(1988), explained in Table 2.2.
Figure 2.8. (a) Pessimum Behavior for Five Aggregates Containing Opaline Material
and (b) Four Regions of Pessimum Behavior (from Hobbs 1988)
20
AAR Facts Book
Table 2.2. Regions of Pessimum Behavior
Region Effect
Explanation
Reaction but no cracking
Insufficient gel forms to cause damage.
Reaction, cracking
excess of alkalis
Expansion continues until all the reactive silica is depleted.
Reaction, cracking
excess of reactive silica
Expansion continues until the alkali level is insufficient to

sustain ASR.
Composition of gel is probably independent of the alkali/

reactive silica ratio.
Alkali/silica ratio and swelling capacity of gel decreases

with increasing silica content.
D
Reaction but no cracking
Reactive silica content very high and reaction so rapid that

the gel forms before concrete has hardened sufficiently and
there is no damage. Copious quantities of gel can be
observed.
Hobbs theory is based on two assumptions. Firstly, it is assumed that, as the reactive-silica-to
alkali ratio (SiO2/Na2Oe) in the system increases beyond the pessimum, the reaction product
(gel) formed has a lower alkali-to-silica ratio (Na2Oe/SiO2) which reduces its affinity for water
uptake and thus reduces its swelling capacity. Secondly, it is assumed that when the reactive
silica content is very high, gel forms rapidly before the concrete has gained sufficient rigidity to
suffer damage. The rapid gel formation consumes the alkalis and insufficient reaction occurs
after the concrete has hardened to result in damage.
Dent Glasser and Kataoka (1981b) showed that quantity of silica dissolved depends on the ratio
of silica to alkali (SiO2/Na2Oe). As discussed previously, in a system of reactive silica and
alkali hydroxides, the concentration of silica in solution increases until the equilibrium
solubility curve (SiO2-pH) is reached as shown in Figure 2.3. If the silica content is low, the
silica concentration increases until all of the silica is dissolved, and the solubility curve is not
reached. If the silica content is very high, the concentration increases until the solubility curve
is reached but then the excess silica in the mix removes hydroxyl ions (OH-) from solution and
reduces the pH. This leads to a reduction of the quantity of silica in solution. Consequently,
there is an optimum silica to alkali ratio (SiO2/Na2Oe) where the maximum silica concentration
is reached, and this occurs when the content of silica is such that it all dissolves when the
solubility curve is reached. This is the pessimum reactive silica content with regards to
expansion of mortar or concrete.
It is somewhat surprising that the pessimum effect is still observed when aggregates are tested
using the accelerated mortar bar test (e.g., Shayan et al. 1988). In this test mortar bars are
21
AAR Facts Book
immersed in 1 M NaOH at 80C (176F). Despite this, expansion may not be observed with
some aggregates in the test if the amount of reactive silica is above the pessimum level. It is
difficult to explain such behavior on the basis of there being an excess of silica (that is, SiO2/
Na2Oe too high) since there is an inexhaustible supply of alkali.
2.5 SOURCES OF ALKALI

While portland cement is considered the main contributor of alkalis, under certain conditions
other materials may provide additional alkalis that are available to the reaction. The source of
alkalis can be from any of the following:
Portland cement
Supplementary cementing materials (e.g., fly ash, slag, silica fume, natural pozzolans)
Aggregates
Chemical admixtures
External sources (e.g., seawater and deicing salts)
Wash water (if used)
The quantity of alkalis in portland cement is typically expressed in terms of equivalent sodium
(written either Na2Oe or Na2Oeq) which may be calculated using the following formula:
Na2Oe = Na2O + 0.658 x K2O
Eqn. 2.3
where: Na2O and K2O are the mass percentages of sodium oxide and potassium oxide in the
portland cement as reported on the cement mill test report. The percentage of alkalis in portland
cement is in the range of 0.2 to 1.3% Na2Oe for most North American sources, but may be as
high as 1.65% Na2Oe or more worldwide.
2.5.1 Pore Solution Composition
Although the percentage of alkalis in portland cement is relatively low in comparison to other
oxides, the alkalis are highly soluble and dominate the pore solution of the concrete. Figure 2.9
shows the composition of the pore solution extracted from portland cement pastes and mortars
at various ages. During the first few hours the pore solution is mainly composed of alkali
sulfates (Na+, K+, and SO42-), but beyond this time the formation of solid sulfate phases
(calcium mono-sulfo-aluminate and ettringite) results in a reduction of the SO42- concentration
in solution, and hydroxyl ions (OH-) come into solution to maintain balance with the positively
charged alkali cations. Beyond the first day, the pore solution is composed almost entirely of
22
AAR Facts Book
Na+, K+, and OH-. The pore solution is saturated with calcium (Ca2+) but the concentration is
very low, 0.6 to 2.5 mmol/L according to measurements made by Struble (1987), at the high pH
resulting from the high concentration of alkali hydroxides.
Figure 2.9. Pore Solution Composition in Portland Cement Pastes and Mortars (Diamond 1983)
Figure 2.10 shows that the alkali concentration of the pore solution is strongly influenced by the
alkali content of the portland cement, the concentration being approximately 0.7 mol/L per 1%
Na2Oe in the cement. Note in Figure 2.10 that the OH- concentration ranges from about 0.15 M
for the lowest alkali cement to about 1.00 M for the highest; this represents a pH range from
13.2 to 14.0 (calculated from the hydroxyl ion concentration assuming an activity coefficient of
unity). The data in Figure 2.10 are for paste samples produced at a water-to-cement ratio of
approximately w/c = 0.50.
23
AAR Facts Book
OH Concentration (mol/L) .
1.0
y = 0.7231x - 0.0011
R = 0.9148
0.8
0.6
0.4
Diamond & Penko, 1988
0.2
Nixon & Page, 1987
0.0
0.0
0.2
0.4
0.6
0.8
1.0
Cement Alkalis (% Na2Oe)
1.2
1.4
Figure 2.10. Relationship between the Alkalinity of the Pore Solution and the Alkali Content of
the Portland Cement (Diamond and Penko 1992; Nixon and Page 1987)
A more general formula has been developed by consideration of the data in Figure 2.10 together
with other published data; the expression is (Helmuth et al. 1993):
[OH-] = 0.339Na2Oe%/(w/c) + 0.022 0.06 mol/L
Eqn. 2.4
If reactive aggregate is present in the system the concentration of hydroxyl and alkali ions will
decrease due to the alkali-silica reaction. Figure 2.11(a) shows the effect of reactive Beltane
opal on the pore solution alkalinity of mortars (Diamond et al. 1981). After a brief induction
period, alkalis are rapidly consumed by reaction with the opal until a steady-state is reached
after approximately 1 month at 20C; the hydroxide concentration is approximately 0.28M OHat this time. Tests at 40C showed a rapid acceleration of the reaction, which is essentially
complete in mortar containing Beltane opal after only 3 days (Diamond et al. 1981).
24
AAR Facts Book
Figure 2.11. Effect of Reactive Aggregate on Pore Solution Composition in (a) Mortars
(Diamond et al. 1981) and (b) Concrete (Thomas 1996)
Results from similar studies on concretes containing reactive flint sand (U.K.) are shown in
Figure 2.11(b) (Thomas 1996). The pore solution of the control concrete showed some
reduction in alkalinity with time, suggesting that there may be some interaction between the
alkali hydroxides and the "inert" aggregate. However, this reduction is minimal compared to
that observed in the concrete containing flint sand. The rate of consumption of alkalis due to the
presence of flint was considerably slower compared with Beltane opal; steady-state conditions
do not appear to be attained until after 1 year has elapsed. However, the hydroxyl ion
concentration at this time (approximately 0.27M OH-) is remarkably similar to the steady-state
concentration observed with opal (approximately 0.28M OH-). Diamond (1983) has postulated
that the threshold alkali concentration required to sustain alkali-silica reaction is likely to be
0.25M OH- or higher.
2.5.2 Threshold Alkali Contents
Based on Stantons early work (1940), it was proposed for many years that expansion resulting
from the alkali-silica reaction is unlikely to occur when the alkali content of the cement is
below 0.6% Na2Oe. Many specifications allow the use of potentially-reactive aggregates
provided that the cement alkali content does not exceed 0.6% Na2Oe. However, it is now
recognized that limiting the alkali content of portland cement is not, by itself, an effective way
of preventing ASR-induced damage because this approach does not control the total alkali
content of the concrete mixture.
25
AAR Facts Book
Limiting the maximum alkali content of concrete is the preferred approach when specifying
alkali levels. The alkali content of portland cement concrete is calculated by multiplying the
cement content by the alkali content of the cement. Figure 2.12 illustrates the effects of the
concrete alkali content on the expansion of concrete prisms stored over water at 38C (100F)
for one year. Concrete mixtures were produced with a reactive siliceous limestone and varying
cement contents (275 to 450 kg/m3, 458 to 750 pcy) using cements with a wide range of alkali
contents (0.67 to 1.40% Na2Oe). From the data it is evident that the expansion of concrete with
a given reactive aggregate is dependent on the alkali content of the concrete. Figure 2.12
indicates that deleterious expansion of the concrete prisms containing this reactive aggregate
may be prevented if the alkali content of the concrete is kept below 3.0 kg/m3 (5.0 lb/yd3)
Na2Oe. It should be noted that expansion has been found to occur in the field at lower alkali
contents than that found necessary to cause expansion in concrete specimens stored over water
in the laboratory. The reason for this is that a portion of the alkalis may be lost through leaching
under the conditions of the concrete prism test (Thomas et al 2006). For example, the aggregate
for which expansion data are presented in Figure 2.12 caused expansion and cracking of fieldexposed concrete blocks (approximately 0.6 x 0.6 x 2.0 m, 2 x 2 x 6 ft.) with an alkali content
of just 1.9 kg/m3 (3.2 lb/yd3) Na2Oe (MacDonald et al. 2012).
Figure 2.12. Effect of Concrete Alkali Content on Expansion

The alkali content of concrete may increase during service due to (i) alkali migration caused by
moisture movements or electrical currents, (ii) penetration of alkalis from external sources (e.g.,
deicing salts), or (iii) long-term release of alkali from aggregates. These factors should be taken
into consideration when setting alkali limits for concrete containing reactive aggregates.
26
AAR Facts Book
2.5.3 Alkali Recycling

As discussed in previous sections of this chapter, alkali-silica reaction results in a reduction of
the concentration of alkali ions (Na+ and K+) and hydroxyl ions (OH-) in the concrete pore
solution due to the formation of an alkali-silica gel (CaO-Na2O/K2O-SiO2-H2O) containing
small amounts of calcium. There is evidence that a portion of the alkalis in the gel exchange for
calcium with time, and alkalis are released back into the pore solution presumably fueling
further reaction with the aggregate providing reactive silica is still available. This phenomenon
was first proposed in Hansen in 1944 (Hansen 1944). Knudsen and Thaulow (1976) showed,
using electron probe microanalysis of concrete, that gel formed within or close to aggregates
was low in calcium and high in alkali, but that as the distance from the aggregate increased the
calcium content of the gel increased, presumably due to close association with the calcium-rich
cement paste.
Urhan (1987), summarizing observations from a number of studies, proposed that there is a
progressive passage from low viscosity alkali-silica gel to C-S-H and that this is accompanied
by a change in the physical and mechanical properties of the gel. Figure 2.13 illustrates changes
in composition, physical and mechanical properties, and structure of the reaction product. Gel
with a low calcium content may have a high swelling capacity, but the viscosity is very low. On
the other hand, gel high in calcium is more rigid, but does not swell. Although not explicitly
stated by Urhan (1987), at some point the composition of the gel is such that the gel has
sufficient swelling capacity and viscosity to cause damage to the surrounding cement paste.
Figure 2.13. Changes in the Nature and Properties of the Alkali-Silica Reaction Product
(Urhan 1987)
27
AAR Facts Book
Thomas (2001) compared the composition of alkali-silica gels found in 7-year-old laboratory
concrete and a 55-year-old concrete dam with the composition of calcium-silicate hydrate (C-S
H) in the concrete from the dam; the data expressed as alkali/silica and calcium/silica atomic
ratios are presented in Figure 2.14. The composition covers a wide range, but it appears that
there is a reasonable relationship between the alkali and calcium contents; i.e., as the calcium
content decreases the potassium content decreases. This supports the concept of a cation
exchange with the calcium replacing the alkali in the initially alkali-rich reaction product. This
exchange occurs as the gel migrates away from the aggregate particle and comes into contact
with the calcium-rich cement paste. The data in the figure indicate that the process continues
slowly as the concrete ages and that the composition of the reaction product may ultimately
approach something similar to C-S-H. This "final product" appears to retain very little alkali
indicating that almost all of the alkali that originally participated in the reaction process may
eventually be recycled to participate in further reaction.
Figure 2.14. Composition of Alkali-Silica Gels in Concrete

In the same paper, Thomas (2001) compared the evolution of the pore solution extracted from
concrete cubes containing a reactive flint aggregate with the expansion of concrete prisms from
the same mixture. The data in Figure 2.15 show that the pore solution alkalinity drops steadily
to an age of 12 weeks, after which it appears to remain stable with a hydroxyl ion concentration
between 0.26 and 0.27 Mol/liter, whereas the expansion of the concrete continues long beyond
the point at which the alkali concentration in the pore solution reaches a steady concentration.
28
AAR Facts Book
This can be explained by alkali recycling, the alkalis being released from the gel being
consumed by reaction, resulting in long-term expansion. Thomas (2001) suggested that alkali
recycling might partially explain why in many large concrete dams the expansion continues for
decades and does not appear to reach a maximum value as is the case for laboratory specimens
(stored over water) where the expansion-time relationship is typically represented by an Sshaped curve reaching a maximum after a few years or less. In large structures, alkali recycling
may continue to fuel ASR (until all the reactive silica is consumed), whereas in laboratory
specimens leaching of the alkalis eventually reduces the alkali concentration to a level below
that necessary to sustain ASR.
Figure 2.15. Evolution of the Pore Solution and Expansion in Concrete
Containing Reactive Flint Sand (Thomas 2001)
2.6 ROLE OF MOISTURE

Sufficient moisture is required to both sustain the chemical reaction and to provide for the
expansion of the gel. It is generally considered that the chemical reaction will cease if the
internal relative humidity inside the concrete falls below 80% (Figure 2.16). Local differences
in moisture availability within the same structure can result in very different levels of ASR
damage occurring within that structure. Specifically, portions of the structure exposed to a
constant or steady source of moisture (e.g., as a result of poor drainage or poor detailing) can
29
AAR Facts Book
exhibit significant ASR-induced damage, while other portions of the structure that remain
essentially dry may show little or no damage.
Figure 2.16. Effect of Relative Humidity on Expansion of Concrete Prisms (Pedneault 1996)
30
AAR Facts Book
REFERENCES
ACI Committee 201. 1991. Proposed revisions of: Guide to Durable Concrete. ACI Materials
Journal, 88(5): 544-582.
Alasali, M.M. and Malhotra, V.M. 1991. "Role of concrete incorporating high volumes of fly
ash in controlling expansion due to alkali aggregate reactions." ACI Materials Journal, 88(2):
159-163. Discussion in Vol. 89, 1992, 110-111.
Asgeirsson, H. and Gudmundsson, G. 1979. Pozzolanic activity of silica dust. Cement and
Concrete Research, 9: 249-252.
Bleszynski, R.F. and Thomas, M.D.A. 1998. Microstructural Studies of Alkali-Silica Reaction
in Fly Ash Concrete Immersed in Alkaline Solutions. Advanced Cement Based Materials, 7:
66-78.
Carpenter, A.J. and Cramer, S.M. 1999. Mitigation of ASR in pavement patch concrete that
incorporates highly reactive fine aggregate. Transportation Research Record 1668, Paper No.
99-1087: 60 67.
Chatterji, S. 1979. The Role of Ca(OH)2 in the Breakdown of Portland Cement Concrete due
to Alkali-Silica Reaction. Cement and Concrete Research, 9(2): 185-188.
Chatterji, S. and Clausson-Kass, N.F. 1984. Prevention of Alkali-Silica Expansion by Using
Slag-Portland Cement. Cement and Concrete Research, 14(6): 816-818.
Chatterji, S., Thaulow, N. and Jensen, A.D. 1987. Studies of alkali-silica reaction. Part 4.
Effect of different alkali salt solutions on expansion. Cement and Concrete Research, 17: 777
783.
Conrow, A.D. 1952. "Studies of abnormal expansion of portland cement concrete." Proceedings
of the American Society for Testing and Materials, 52: 1205-1227.
Dent Glasser, L.S. and Kataoka, N. 1981a. The chemistry of alkali-aggregate reactions.
Proceedings of the 5th International Conference on Alkali-Aggregate Reaction, Cape Town,
S252/23.
Dent Glasser, L.S. and Kataoka, N. 1981b. The chemistry of alkali-aggregate reactions.
Cement and Concrete Research, 11: 1-9.
Diamond, S. 1983. "Alkali reactions in concrete - Pore solution effects." Proceedings of the 6th
International Conference on Alkalis in Concrete, (Eds. G.M. Idorn and Steen Rostam), Danish
Concrete Association, Copenhagen, 155-166.
31
AAR Facts Book
Diamond, S. 1989. ASR - Another Look at Mechanisms. Proceedings of the 8th International
Conference on Alkali-Aggregate Reaction, (Eds. K. Okada, S. Nishibayashi, and M.
Kawamura), Kyoto, Japan, 8394.
Diamond, S. and Penko, M. 1992. Alkali silica reaction processes: the conversion of cement
alkalis to alkali hydroxide. Durability of Concrete G.M. Idorn International Symposium, (Ed.
Jens Holm), ACI SP-131, American Concrete Institute, Detroit, 153-168.
Diamond, S., Barneyback, R.S. and Struble, L.J. 1981. "On the Physics and Chemistry of Alkali
Silica Reactions." Proceedings of the 5th International Conference on Alkali-Aggregate
Reaction in Concrete, Cape Town, NBRI of the Council for Scientific and Industrial Research,
Pretoria, S.A., S252/22.
Fournier, B. and Berube, M.A. 2000. Alkali-aggregate reaction in concrete: a review of basic
concepts and engineering implications. Canadian Journal of Civil Engineering, 27: 167-191.
Glasser, L.S. and Kataoka, N. 1981. "The chemistry of alkali-aggregate reactions." Proceedings
of the 5th International Conference on Alkali-Aggregate Reaction, Cape Town, S252/23.
Proceedings of the Highway Research Board, 40: 462-474.
Hansen, W.C. 1944. Studies relating to the mechanism by which the alkali-aggregate reaction
proceeds in concrete. Journal of the American Concrete Institute, 15(3): 213-227.
Helmuth, R., Stark, D., Diamond, S. and Moranville-Regourd, M. 1993. Alkali-Silica
Reactivity: An Overview of Research. SHRP-C-342, Strategic Highway Research Program,
National Research Council, Washington, D.C., 105 p.
Hobbs, D.W. 1988. Alkali-Silica Reaction in Concrete. Thomas Telford, London, 183 p.
Iler, R.K. 1979. The Chemistry of Silica. Wiley-Interscience, New York, 886 p.
Jin, W., Meyer, C. and Baxter, S. 2000. 'Glascrete'Concrete with Glass Aggregate. ACI
Materials Journal, 97(2): 208-213.
Kawamura, M., Kayyali, O.A. and Haque, M.N. 1988. Effects of a fly ash on pore solution
composition in calcium and sodium chloride-bearing mortars. Cement and Concrete Research,
18: 763-773.
Knudsen, T. and Thaulow, N. 1979. Quantitative microanalyses of alkali-silica gel in
concrete. Cement and Concrete Research, 5: 443-454.
32
AAR Facts Book
MacDonald, C.A., Rogers, C. and Hooton, R.D. 2012. The relationship between laboratory
and field expansion observations at the Kingston outdoor exposure site for ASR after twenty
years. Proceedings of the 14th International Conference on Alkali-Aggregate Reaction in
Concrete, Austin, TX, USA, May.
McGowan, J.K. and Vivian, H.E. 1952. Studies in cement-aggregate reaction: correlation
between crack development and expansion of mortars. Australian Journal of Applied Science,
3: 228-232.
Moranville-Regourd, M. 1989. "Products of Reaction and Petrographic Examination.
Proceedings of the 8th International Conference on Alkali-Aggregate Reaction, (Eds. K. Okada,
S. Nishibayashi and M. Kawamura), Kyoto, The Soc. of Mat. Science, Japan, 445-456.
Newlon, H.H., Jr. and Sherwood, W.C. 1964. Methods for Reducing Expansion of Concrete
Caused by Alkali-Carbonate Rock Reactions. Highway Research Record No. 45, Highway
Research Board, Washington, D.C., 134-150.
Nixon, P.J. and Page, C.L. 1987. "Pore solution chemistry and alkali aggregate reaction."
Concrete Durability, Katherine and Bryant Mather International Conference, (Ed. J.M.
Scanlon), ACI SP-100, Vol. 2, American Concrete Institute, Detroit, 1833-1862.
Pedneault, A. 1996. Development of testing and analytical procedures for the evaluation of the
residual potential of reaction, expansion, and deterioration of concrete affected by ASR. M.Sc.
Memoir, Laval University, Qubec City, Canada, 133 p.
Powers, T.C. and Steinour, H.H. 1955a. An investigation of some published researches on
alkali-aggregate reaction. I. The chemical reactions and mechanism of expansion. Journal of
the American Concrete Institute, 26(6): 497-516.
Powers, T.C. and Steinour, H.H. 1955b. An interpretation of some published researches on the
alkali-aggregate reaction. Part 2: A hypothesis concerning safe and unsafe reactions with
reactive silica in concrete. Journal of the American Concrete Institute, 26(8): 785-811.
Shayan, A. and Xu, A. 2004. Value-added utilisation of waste glass in concrete. Cement and
Concrete Research, 34: 8189.
Swenson, E.G. 1957. A reactive aggregate undetected by ASTM tests. American Society for
Testing and Materials, 57: 48-51.
33
AAR Facts Book
Swenson, E.G. and Gillott, J.E. 1964. Alkali-carbonate rock reaction. Highway Research
Record, 45: 21-40.
Tang, M., Ye, Y.F., Yuan, M.Q. and Zheng, S.H. 1983. "The preventive effect of mineral
admixtures on alkali-silica reaction and its mechanism." Cement and Concrete Research, 13:
171-176.
Tang, M., Liu, Z. and Han, S. 1987. Mechanism of alkali-carbonate reaction. Proceedings of
the 7th International Conference on Concrete Alkali-Aggregate Reactions, (Ed. P.E. GrattanBellew), Noyes Publications, New Jersey, 275-279.
Thomas, M.D.A. 1998. The Role of Calcium in Alkali-Silica Reaction. Materials Science of
Concrete - The Sidney Diamond Symposium, (Eds. M. Cohen, S. Mindess and J.P. Skalny),
American Ceramic Society, Westerville, OH, 325-331.
Thomas, M.D.A. 2001. The role of calcium hydroxide in alkali recycling in concrete.
Materials Science of Concrete Special Volume on Calcium Hydroxide in Concrete, (Ed. J.
Skalny, J. Gebauer and I. Odler), American Ceramic Society, Westerville, OH, 269-280.
Thomas, M.D.A., Nixon, P.J. and Pettifer, K. 1991. The effect of pfa on alkali-silica reaction.
Second CANMET/ACI Conference on the Durability of Concrete, (Ed. VM Malhotra), ACI SP
126, II, American Concrete Institute, Detroit, 919-940.
Thomas, M.D.A., Fournier, B., Folliard, K., Ideker, J. and Shehata, M. 2006. Test methods for
evaluating preventive measures for controlling expansion due to alkali-silica reaction in
concrete. Cement and Concrete Research, 36(10): 1842-1856.
Urhan, S. 1987. Alkali silica and pozzolanic reactions in concrete. Part 1: Interpretation of
published results and an hypothesis concerning the mechanism. Cement and Concrete
Research, 17: 141-152.
Wang, H. and Gillott, J.E. 1991. Mechanism of Alkali-Silica Reaction and the Significance of
Calcium Hydroxide. Cement and Concrete Research, 21(4): 647-654.
34
AAR Facts Book
Chapter 3 Symptoms of ASR
3 Symptoms of Alkali-Silica Reaction (ASR)
3.1 INTRODUCTION
As mentioned before, three conditions are necessary to initiate and sustain AAR in concrete: (1)
reactive mineral forms must be present in the aggregate materials, (2) the concentration of
alkali hydroxides ([Na+,K+-OH-]) in the concrete pore fluid must be high, and (3) sufficient
moisture must be present. Concrete elements affected by AAR respond quite differently from
one another, reflecting wide variations in the above conditions.
Common visual symptoms of ASR consist of:
Cracking
Expansion causing deformation, relative movement, and displacement
Localized crushing of concrete
Extrusion of joint (sealant) material
Surface pop-outs
Surface discoloration and gel exudations
A brief description and photographs of these symptoms are given hereafter. Additional
photographs could be found in Thomas et al. (2011).
3.2 CRACKING
The pattern of cracking due to ASR is influenced by factors such as the shape or geometry of
the concrete member, the environmental conditions, the presence and arrangement of
reinforcement, and the load or stress fields (restraint) applied to the concrete. The classic
symptom of ASR is map cracking (also called pattern cracking), which takes the form of
randomly-oriented cracks on the surface of concrete elements that are relatively free
(unrestrained) to move in all directions (Figures 3.1A & 3.1B). However, drying shrinkage,
freezing/thawing cycles, and sulfate attack can also result in a pattern of cracks showing a
random orientation.
35
AAR Facts Book
Figure 3.1. A. Map-cracking in the wing wall of a 30-year-old bridge structure affected by ASR
(CSA 2000). B. Severe map-cracking and associated gel staining around cracks in a median
highway barrier affected by ASR. C. Well-defined crack pattern associated with the development
of ASR in highway pavement; the orientation of predominant cracks is longitudinal, while map- or
pattern-cracking is also identified. D. Preferred alignment of cracks in an ASR-affected concrete
column. E. Longitudinal cracking in a precast, reinforced concrete beam affected by ASR.
36
AAR Facts Book
Steel reinforcement or other restraint arising from applied compressive stress may reduce ASR
expansion in concrete. However, surface cracking due to AAR is often not significantly reduced
by the use of internal or external restraint. When expansion is restrained in one or more
directions, more expansion occurs in the direction of least confinement, and the cracks become
oriented in the same direction as the confining stresses. For example, with concrete pavements,
the expansion being restrained in the longitudinal direction, a greater amount of expansion
occurs in the transverse direction and cracks develop preferentially in the longitudinal direction
(Figure 3.1C); in the case of reinforced concrete columns, cracks tend to be aligned vertically
due to the restraint imposed by the primary reinforcement and the dead load (Figure 3.1D); for
prestressed bridge girders, the cracks will usually be aligned horizontally due to the
confinement imposed by the prestressing tendons parallel to the beam axis (Figure 3.1E).
In many cases, discoloration occurs around the cracks, often due to gel exudation in the vicinity
of the cracks (Figure 3.1B).
Cracking is usually most severe in areas of structures where the concrete has a renewable
supply of moisture, such as close to the waterline in piers, from the ground behind retaining
walls, beneath pavements slabs, elements/parts of structures exposed to rain, or by wick action
in piers or columns (Figure 3.2). Concrete members undergoing ASR and experiencing cyclic
exposure to sun, rain, and wind, or portions of concrete piles in tidal zones often show more
severe surface cracking resulting from induced tension cracking in the less expansive (due to
alkali leaching/dilution processes, variable humidity conditions, etc.) surface layer under the
expansive thrust of the inner concrete core (Stark and Depuy 1987, ACI 1998) (Figure 3.3).
37
AAR Facts Book
Figure 3.2. A&B. Map-cracking in wing walls (exposed elements) of bridge structures affected by
ASR. C. Cracking in the column and the exposed portion of a beam affected by ASR. D&E.
Cracking in the exposed portion (above ground) of the wing wall of a bridge structure affected by
ASR; the wing wall has been excavated before being demolished. The lower ground portion
shows limited cracking; discoloration occurs around the cracks above ground (E).
38
AAR Facts Book
A
Surface layer shows less reactivity
Drier
Alkali leaching
Carbonation
Humid
3-D
Exp
Water
Table
Saturated
Humid
3-D
Exp
Water
Table
Saturated
Figure 3.3. Concrete member undergoing ASR and experiencing cyclic exposure to sun, rain
and wind show more severe surface cracking resulting from induced tension cracking in the
less expansive (due to alkali leaching/dilution processes, variable humidity conditions, etc.)
surface layer under the expansive thrust of the inner concrete core.
39
AAR Facts Book
3.3 EXPANSION CAUSING MOVEMENTS AND DEFORMATIONS

The extent of ASR often varies between or within the various members/parts of an affected
concrete structure, thus causing distresses such as:
relative movement of adjacent concrete members or structural units (Figure 3.4);
deflection, closure of joints with associated squeezing/extrusion of sealing materials,
and, ultimately, spalling of concrete at joints (Figure 3.5).
40
AAR Facts Book
Figure 3.4. A. Relative movement of abutting sections of parapet wall in a bridge structure affected by
ASR (Stark 1991). B-D. Expansion of concrete pavement due to ASR (overlaid with asphalt), which
pushes against the adjacent building foundation causing shearing of concrete columns. E. Concrete
sidewalk made with alkali-carbonate reactive limestone aggregate. Originally, the sidewalk and the curb
were adjacent to each other; however, the expansion of the sidewalk created a gap that had to be filled
with asphalt.
41
AAR Facts Book
Figure 3.5. A. Expansion of bridge girder leading to loss of clearance between the girder and
embankment and eventually crushing of the girder end with localized spalling. B&C. Expansion
causing spalling at joints in concrete pavement sections incorporating highly-reactive
aggregates. D. Expansion with associated severe spalling in abutting jersey barrier sections
affected by ASR.
It is important to remember that deformations in concrete structures may be caused by a range
of different mechanisms, such as loading, thermal or moisture movements, differential
shrinkage, gravity and foundation effects, hydraulic pressure, creep, impact, and vibrations
(BCA 1992).
42
AAR Facts Book
3.4 POP-OUTS
The expansion of individual unsound or frost-susceptible aggregate particles [such as
laminated, schistose and argillaceous, clayey or porous particles or certain varieties (porous) of
chert, ironstones] at or near the concrete surface due to frost action is likely to be the main
factor for the development of pop-outs in northern countries (Figure 3.6A). Pop-outs can also
be caused by a poor bond between the cement paste and dusty coarse aggregate particles. Alkali
-silica reactive aggregates undergoing expansion near the concrete surface may also induce the
detachment of a portion of the skin of concrete leaving the reactive aggregate in the bottom
(Figure 3.6B).
43
AAR Facts Book
Figure 3.6. A. Pop-out created by the expansion of a frost-susceptible porous coarse aggregate
particle (leached chert). B. ASR-induced pop-out in a concrete pavement incorporating highlyreactive aggregates; also noted pattern cracking. C. Efflorescence and exudations of alkali-silica
gel at the surface of the concrete foundation of 25-year-old highway bridge affected by ASR. D.
Efflorescence and exudations of alkali-silica gel at the surface of a small concrete monument
affected by ASR. E. Surface discoloration and exudation associated with cracks.
44
AAR Facts Book
3.5 SURFACE DEPOSITS (GEL EXUDATION VS. EFFLORESCENCE)

Although surface gel exudation is a common and characteristic feature of ASR, the presence of
surface deposits is not necessarily indicative of ASR as other mechanisms (such as frost action
or the movement of water through cracked concrete members) can also cause surface deposits
called efflorescence (without the presence of ASR gel) (Figures 3.5C-E).
45
AAR Facts Book
REFERENCES
ACI Committee 221. 1998. Report on Alkali-Aggregate Reactivity. ACI 221.1R-98,
American Concrete Institute, Farmington Hills, MI, 30 p.
British Cement Association (BCA). 1992. The Diagnosis of Alkali-Silica Reaction Report of
a Working Party. Wexham Springs, Slough, U.K., SL3 6PL, 44 p.
CSA. 2000. Guide to the Evaluation and Management of Concrete Structures Affected by
Alkali-Aggregate Reaction. CSA A864-00, Canadian Standards Association, Mississauga,
Ontario, Canada.
Institution of Structural Engineers (ISE). 1992. Structural Effects of Alkali-Silica Reaction
Technical Guidance Appraisal of Existing Structures. Institution of Structural Engineers,
London, 45 p.
Stark, D. 1991. Handbook for the Identification of Alkali-Silica Reactivity in Highway
Structures. SHRP-C/FR-91-101, TRB National Research Council, 49 p.
Stark, D. and Depuy, G. 1987. Alkali-Silica Reaction in Five Dams in Southwestern United
States. Proceedings of the Katherine and Bryant Mather International Conference on Concrete
Durability, Atlanta, Georgia, USA, ACI SP-100, April, 1759-1786.
Thomas, M.D.A., Fournier, B., Folliard, K.J. and Resendez, Y.A. 2011. Alkali-Silica
Reactivity Field Identification Handbook. Office of Pavement Technology, Federal Highway
Administration, 72 p.
46
AAR Facts Book
Chapter 4 Test Methods
4 Test Methods
4.1 INTRODUCTION
Ever since Stanton (1940) reported his discovery of alkali-aggregate reaction (AAR) in
California, there has been keen interest in laboratory tests to (1) predict whether a given
aggregate will cause expansion and cracking in concrete and (2) to evaluate preventive
measures to allow for safe use of those aggregates found to be potentially reactive. This chapter
describes the various test methods currently being used to assess the reactivity of aggregates
(with regard to alkali-silica reaction (ASR) and alkali-carbonate reaction (ACR)).
In addition to describing and recommending appropriate test methods for evaluating aggregate
reactivity, this chapter also provides recommendations for tests aimed at evaluating measures
for preventing ASR, such as supplementary cementing materials (SCMs) and lithium nitrate.
Only limited information is provided on ACR, with the primary focus on test methods aimed at
identifying aggregates susceptible to ACR so that they can be avoided for use in concrete as
there are no available preventive measures for such aggregates. In addition to test methods
aimed at assessing aggregate reactivity and preventive measures, this chapter also includes
discussion on alkali release from aggregates and SCMs. There are no standardized tests that
adequately address this important issue of alkali release, but for completeness, relevant
information and recent trends are presented that should provide some useful insight to readers.
4.2 ASR TEST METHODS FOR EVALUATING AGGREGATE REACTIVTY
Stanton (1940) was not only the first to discover ASR in field structures, but he was also the
first researcher to develop a test method to assess aggregate reactivity, and he used this
technique to also evaluate the use of pozzolans to control ASR-induced expansion. The method
developed by Stanton, which is essentially the same as the current ASTM C 227 test method, is
still in use today by some researchers and practitioners, but a wide variety of test methods have
been developed and implemented since the time of Stantons pivotal research on ASR. Some of
these test methods have been successful, some have proven to be complete failures, and others
fall somewhere in the middle. Through research and development, as well as trial and error, test
methods have evolved over the years, and there has been a general convergence in terms of the
tests that are generally used. This chapter describes the various test methods that are currently
being used, discusses the positive and negative attributes of each test, and provides
recommendations on what current standard tests are most appropriate for testing aggregate
reactivity and preventive measures.
47
AAR Facts Book
Table 4.1 lists the most common standard test methods that have been or are currently being
used to assess aggregate reactivity, including relevant comments related to the positive and/or
negative attributes of each test. Discussion on each of these methods follows, with particular
emphasis placed on the Accelerated Mortar Bar Test (AMBT) and Concrete Prism Test (CPT)
as these two are the most commonly used ASR test methods and are integrated into AASHTO
PP65-11, Standard Practice for Determining the Reactivity of Concrete Aggregates and
Selecting Appropriate Measures for Preventing Deleterious Expansion in New Concrete
Construction, which is a newly-developed standard practice developed under FHWA Project
DTFH61-06-D-00035.
4.2.1 ASTM C 295: Standard Guide for Petrographic Examination of Aggregates for Concrete
A detailed petrographic examination, following ASTM C 295, provides useful information
regarding the potential reactivity (with respect to both ASR and ACR) of a given fine or coarse
aggregate. Within ASTM C 295 various techniques are employed that allow for the
identification and quantification (through point count) of minerals that have been known to lead
to expansion and cracking in concrete containing such aggregates. The specific minerals that
are identified and quantified in ASTM C 295 are:
Opal
Chalcedony
Cristobalite
Tridymite
Highly Strained Quartz
Microcrystalline Quartz
Volcanic Glass
Synthetic siliceous Glass.
The minerals listed above are found in a range of aggregate types, as per ASTM C 295,
including:
Chert
Gneiss
Gneissic Granite
Graywacke
Phyllites
Schist
Vein Quartz
Sandstone
Quartzite
48
AAR Facts Book
Table 4.1. Test Methods for Evaluating Aggregate Reactivity

Test Method
Comments
ASTM C 295: Standard Guide for Petrographic Useful evaluation to identify many (but not all) potentially reactive
Examination of Aggregates for Concrete
components in aggregates.
Reliability of examination depends on experience and skill of
individual petrographer.
Results should not be used exclusively to accept or reject aggregate
source findings best used in conjunction with other laboratory
tests (e.g., AASHTO T 303 and/or ASTM C 1293).
ASTM C 289: Standard Test Method for Potential Aggregate test in which crushed aggregate is immersed in 1M
Alkali-Silica Reactivity of Aggregates (Chemical
NaOH solution for 24 hours solution is then analyzed for amount
Method)
of dissolved silica and alkalinity.
Poor reliability.
Problems with test include:
- Other phases present in aggregate may affect dissolution of silica
(Brub and Fournier 1992).
- Test is overly severe, leading aggregates with good field
performance to fail the test.
- Some reactive phases may be lost during pretest processing.
ASTM C 227: Standard Test Method for Potential Mortar bar test (aggregate/cement = 2.25), intended to study cement
Alkali Reactivity of Cement-Aggregate
-aggregate combinations.
Combinations (Mortar-Bar Method)
Specimens stored in high-humidity containers at 38C.
Several reported problems with test, including excessive leaching of
alkalis from specimens.
AASHTO T 303 (ASTM C 1260): Standard Test Mortar bar test, originally designed to assess aggregate reactivity.
Method for Potential Alkali Reactivity of Aggregates Bars are soaked in 1N NaOH solution for 14 days.
(Mortar-Bar Method)
Accelerated test suitable as screening test, but because of severity
of test, it should not be used, by itself, to reject a given aggregate. If
aggregate is tested using both AASHTO T 303 and ASTM C 1293,
the results of ASTM C 1293 should govern.
ASTM C 1293: Standard Test Method for Concrete Concrete prism test, generally regarded as best indicator of field
Aggregates by Determination of Length Change of
performance, is conducted at high humidity (close to 100%) at 38
Concrete Due to Alkali-Silica Reaction
C.
Uses high-alkali cement (raised to 1.25% Na20e), with a cement
content of 420 kg/m3.
Developed as aggregate test (using non-reactive fine aggregate to
test reactivity of coarse aggregate, and vice-versa).
Test requires one year for completion this long duration limits its
use by many agencies and owners.
Cannot be used to determine the alkali threshold for a given
aggregate due to leaching of alkalis from the prisms during the
course of the test.
The identification and quantification of reactive phases within aggregates are accomplished
following ASTM C 295 through the use of optical methods (using reflective and transmitted
light) and may be aided with the complementary use of x-ray diffraction (XRD) and scanning
49
AAR Facts Book
electron microscopy (SEM). Most of the minerals that cause ASR can be identified and
quantified, but some finely divided forms of quartz are not visible using a petrographic
microscope. For example, some siliceous limestones contain less than 5 percent undetectable
(using petrographic microscope) finely divided quartz particles, which can cause ASR-induced
expansion in concrete (Fournier and Brub 1991). As such, caution should be taken in
classifying a given aggregate as being non-reactive based solely on petrographic evaluation
information from other laboratory tests and/or field performance can be combined with such
petrographic data/information to more accurately determine whether an aggregate may cause
deleterious expansion in field structures.
4.2.2 ASTM C 289: Standard Test Method for Potential Alkali-Silica Reactivity of Aggregates
(Chemical Method)
ASTM C 289, often referred to as the Quick Chemical Test, is a test that involves crushing an
aggregate (passing the 300 m sieve and retained on the 150 m sieve), soaking it in 1 N NaOH
solution for 24 hours at 80C and measuring the amount of silica that goes into solution and the
alkalinity at the completion of the test. The results of the test, specifically the amount of silica
that went into solution and the reduction in the alkalinity of the soak solution, are plotted on a
graph, contained in the Appendix to ASTM C 289. Based on how the data (amount of silica
dissolved and reduction in alkalinity) plots within this graph, the aggregate is classified as
innocuous, potentially deleterious, or deleterious.
Although ASTM C 289 is still used by some researchers and practitioners, experience has
shown that it is generally not an accurate predictor of aggregate reactivity, with many
aggregates showing discrepancies between the classification (innocuous, potentially deleterious,
or deleterious) obtained from ASTM C 289 and the performance of such aggregates in the
concrete prism test (ASTM C 1293) or in actual field structures (Hooton 1990; Grattan-Bellew
1989). The extreme severity of this test, specifically the high temperature and high alkalinity
soak solution in which the aggregates are placed, coupled with the high surface area of the
sample as a result of crushing, will cause some aggregates to be identified as reactive, even
though such aggregates perform satisfactorily in ASTM C 1293 and in field structures (Brub
and Fournier 1993). On the other hand, some reactive aggregates may lose most or all of their
reactive phases during crushing and sieving and would thus be found to be innocuous per
ASTM C 289, even though concrete containing such aggregates would exhibit significant
expansion and cracking in the laboratory and field (Brub and Fournier 1992). Because of the
above issues, ASTM C 289 is not recommended as a test method for evaluating aggregate
reactivity, even though the test is still being used by some agencies and owners.
50
AAR Facts Book
4.2.3 ASTM C 227: Standard Test Method for Potential Alkali Reactivity of Cement-Aggregate
ASTM C 227 is essentially the method Stanton (1940) developed of storing mortar bars (25 mm
x 25 mm x 285 mm) over water at 38C to provide for a high-temperature, high-humidity
atmosphere within a sealed container. Because of the small cross section of the mortar bars used
in this test, coarse aggregates must be processed by crushing, sieving, and washing in order to
produce particles small enough to be used in the mortar bars. The aggregate to cement ratio is
specified to be 2.25 (by mass), and there is no specific requirement for the cement to be used in
the test. The ASTM standard specification for concrete aggregates (ASTM C 33) classifies an
aggregate as being reactive when a given aggregate expands by more than 0.05% at 3 months or
0.10% at 6 months.
There are several technical problems associated with this test method, some of which spurred
the development of more accurate methods for evaluating aggregate reactivity. The most
significant issue with this test is that due to the small specimen cross section, appreciable
amounts of the alkalis within the mortar bars leach out and end up residing in the water in the
bottom of the container or being absorbed by the wicking material that lines the side of the
container (Swenson 1957). Slowly-reactive aggregates are particularly sensitive to the leaching
issue as the alkalis may leach from the bars before the aggregates ever have the opportunity to
react and expand. Some slowly-reactive aggregates that have been observed to pass this test but
that will fail more accurate tests and/or that will exhibit expansion and cracking in field
structures include greywackes, argillites, quartzites, and some gneisses (Grattan-Bellew 1978;
Stark 1980). Another issue with this test is that only fine aggregates can be tested so coarse
aggregates have to be crushed, graded, and washed, which may alter the textural nature of the
aggregates or remove some reactive phases.
Given the technical issues associated with this test, specifically that a wide range of known,
reactive aggregates will pass this test, it is not recommended as a test method for evaluating
aggregate reactivity.
4.2.4 AASHTO T 303 (ASTM C 1260): Standard Test Method for Potential Alkali Reactivity
of Aggregates (Mortar-Bar Method)
AASHTO T 303 (ASTM C 1260) is based on the method developed by Oberholster and Davies
(1986) at the National Building Research Institute in South Africa. The test, often referred to as
the accelerated mortar bar test (AMBT), has been adopted by various countries and agencies,
including the United States (ASTM and AASHTO) and Canada. The test entails casting mortar
bars containing the subject aggregate (either coarse or fine), which is processed to a standard
gradation. The mortar bars are then removed from their molds after 24 hours and placed in
51
AAR Facts Book
water at room temperature. The temperature of the water is then raised to 80C in an oven, and
the mortar bars are stored in this condition for the next 24 hours. Upon removing the bars from
the water, they are measured for initial length and then submersed in a 1N NaOH solution at
80C, where they are then stored for 14 days. Length change measurements are made
periodically during this storage period. The total expansion at the end of the 14-day soaking
period typically is used in specifications, although the expansion limits specified by different
agencies vary. For example, the expansion criteria established by ASTM and CSA are as
follows:
ASTM C 1260 expansion criteria:
< 0.10% is considered innocuous
0.10 to 0.20% is considered potentially reactive
> 0.20% is considered reactive
CSA A23.2-25A expansion criteria:

> 0.15% (0.10% for limestones) is considered reactive
< 0.15% (0.10% for limestones) is considered non-reactive
In ASTM C 1260, these expansion limits are proposed in an appendix to the standard test
method and are not a mandatory part of the standard, whereas in CSA, the limits form part of a
standard practice for assessing reactivity and evaluating the effectiveness of preventive
measures (CSA A23.2-27A). AASHTO PP65-11 specifies a 0.10 percent expansion limit at 14
days; aggregates with expansions below this are considered non-reactive. However, AASHTO
PP65-11 cautions that AASHTO T 303 is not as accurate as the concrete prism test (ASTM C
1293) for evaluating aggregate reactivity and that users who rely upon AASHTO T 303 are
assuming some degree of risk in doing so.
AASHTO T 303 is recognized as a very severe test method because of the extreme test
conditions, specifically the use of a highly alkaline storage solution and high temperature.
Because of this inherent severity, the test has been shown to identify some aggregates as being
reactive, sometimes with expansions as high as 0.25 percent at 14 days, even though they have
performed well in concrete prism testing (ASTM C 1293) and in field applications (Brub and
Fournier 1992). In fact, recent, unpublished work performed by the authors of this Facts Book
included the evaluation of an aggregate from Oahu that when tested according to AASHTO T
303 expanded greater than 0.40 percent at 14 days but easily passed ASTM C 1293. Because
AASHTO T 303 has been shown to fail a significant number of aggregates, especially natural
sands and gravels, an aggregate should not be rejected based solely on the test results.
More recently, a number of coarse aggregates have been found to pass AASHTO T 303 but fail
ASTM C 1293 and expand and crack in outdoor exposure blocks (Folliard et al. 2006). Arrieta
(2012) reported that the primary reason the aggregates tested by Folliard et al. (2006) passed the
AASHTO T 303 test but failed ASTM C 1293 was related to the pessimum effect. Specifically,
the primary reactive mineral was found petrographically to be chert, which is an aggregate
known to exhibit a pessimum effect. The concept of the pessimum effect is shown graphically
in Figure 4.1 (adapted from Poole (1992) by Arrieta (2012)). According to Poole (1992), the
52
AAR Facts Book
pessimum behavior proposes that ASR will cease when either the alkalis in the pore solution or
the reactive constituents within the aggregate particles are depleted. If the content of the
reactive mineral (chert, in this case) is too low compared to the availability of alkalis, very little
gel will form and hence expansion will be limited. On the contrary, if there is an overabundance
of reactive minerals compared to the availability of alkalis, the alkalis will become depleted and
only a small amount of gel with form, with minimal expansion observed. When an optimum
ratio of reactive silica to available alkalis is achieved (50 percent for the illustration in Figure
4.1, and about 10-12 percent for most cherts), expansion will be at a maximum.
Measured expansion (%)
0.3
0.25
0.2
0.15
0.1
0.05
0
0
10 20 30 40 50 60 70 80 90
Amount of reactive aggregate present (%)
100
Figure 4.1. Graphical Representation of the 'Pessimum' Behavior Concept

(adapted from Poole (1992) by Arrieta (2012))
For the aggregates studied by Folliard et al. (2006) and evaluated in more detail by Arrieta
(2012), AASHTO T 303 was modified by replacing the pessimum aggregate with varying
amounts of non-reactive sand. Figure 4.2 shows the typical response for an aggregate from
Central Texas. Based on petrographic evaluations, it was found that the chert content in this
pessimum aggregate was approximately 75% of the as-received coarse aggregate. Figure 4.2
takes the chert content into account, along with the level of replacement by the non-reactive
aggregate, to plot the approximate chert content on the x-axis versus the expansion observed
after 14 or 28 days storage in 1 N NaOH solution at 80C. As shown in this figure, the
pessimum proportion was found to be approximately 10 percent chert, which is consistent with
past studies (Nixon et al. 1989; Hudec 1990). These results are encouraging in that they show
that AASHTO T 303 can be used to not only identify pessimum aggregates but also to
determine the actual pessimum proportion. Also, the results help to explain why these coarse
aggregates failed the concrete prism test. The coarse aggregate content used in ASTM C 1293 is
70 percent of the overall aggregate content. Assuming that 75 percent of the coarse aggregate is
comprised of flint, the total amount of flint in the mixtures was about 52 percent, whereas in the
unmodified AASHTO T 303 test, the coarse aggregate content was 100 percent, hence resulting
in a chert content of 75 percent. Referring to Figure 4.2, the higher the chert content (beyond
the pessimum proportion of about 10 percent), the lower the expansion. Only when the chert
53
AAR Facts Book
content was reduced by replacing the reactive sand with a non-reactive sand did expansion
exceed the 0.10 percent expansion limit typically specified for AASHTO T 303.
ASTM C 1260 expansion (%)
0.450
0.400
0.350
0.300
0.250
Expansion limit
0.200
14 days
0.150
28 days
0.100
0.050
0.000
0
10
20
30
40
50
60
70
Chert present in the mixture (%)
80
Figure 4.2. Pessimum Graph for Coarse Aggregate from Central Texas, Expressed as Chert
Percent, when using AASHTO T 303 Testing Regime (Arrieta 2012)
In summary, AASHTO T 303 is a rapid test that in some cases yields results that agree with the
results from ASTM C 1293 tests or field experience. However, as described herein, there are a
wide range of aggregates that will give erroneous and misleading results in AASHTO T 303,
when compared to the more accurate and realistic concrete prism test. As such, when both
AASHTO T 303 and ASTM C 1293 data are available for a given aggregate, the ASTM C 1293
data should be relied upon for assessing aggregate reactivity.
Length Change of Concrete Due to Alkali-Silica Reaction
ASTM C 1293, commonly referred to as the concrete prism test, is generally considered the
most accurate and effective test in predicting the field performance of aggregates. In this test,
concrete is cast with a cement content of 420 kg/m3. The cement is required to have an
equivalent alkali content between 0.8% and 1.0%, and additional alkalis (NaOH) then are added
to the mixing water to obtain a total alkali content of 1.25% (by mass of cement), which equates
to a total alkali content in the concrete mixture of 5.25 kg/m3. Concrete prisms are cast, cured
for 24 hours at 23C, and then stored over water at 38C. Expansion measurements are taken at
regular intervals, and when testing plain concrete (without SCMs or chemical admixtures), the
test typically is run for one year (or as described later in this chapter, the test is conducted for
two years when evaluating supplementary cementing materials or lithium-based admixtures).
54
AAR Facts Book
In recent years, more countries and agencies have adopted the concrete prism test as a standard
method. An expansion limit of 0.04% (at the end of the one- or two-year test) typically is
specified, as this value has been reported to correlate well with cracking of test prisms. This
expansion limit (0.04%) is referenced in the appendix to ASTM C 1293 and in AASHTO PP65.
Although ASTM C 1293 is generally regarded as the most accurate test for assessing aggregate
reactivity, it is not without its limitations or drawbacks. Thomas et al. (2006) reported that
approximately 35 percent of the internal alkalis within concrete prisms may leach out during the
course of a one-year test, with as much as 20 percent of the alkalis leaching out in the first 90
days, as shown in Figure 4.3. In this figure, the concentration of alkalis in the water at the
bottom of the bucket (above which the prisms are stored) is shown as a function of time for a
selected mixture following ASTM C 1293. Taking into account the volume of water at the
bottom of the bucket and ignoring that some of the alkalis may be absorbed into the wicking
material on the side of the bucket, Thomas et al. (2006) calculated the total amount of alkalis
that leached from the prisms, arriving at the values shown above (20 percent after 90 days and
35 percent after one year). As such, higher alkali contents are needed to cause expansion in the
CPT than what is actually needed in field concrete or exposure blocks, which also limits the use
of the CPT to estimate the alkali threshold for a given aggregate, as discussed in more detail
next.
80
Alkali Concentration (mMol/Litre)
K
Na
60
40
20
0
0
6
Age (months)
12
Figure 4.3. Concentration of Alkalis Measured in the Water at the Bottom of an ASTM C 1293
Container as a Function of Time (after Thomas et al. 2006)
The effects of alkali leaching from concrete prisms tested during the course of an ASTM C
1293 test were clearly illustrated in a study by Folliard et al. (2006), in which a series of three
concrete mixtures with varying alkali loadings were cast using a highly-reactive sand from El
Paso, Texas. Prisms from these mixtures were then tested using ASTM C 1293, and exposure
55
AAR Facts Book
blocks were cast and stored outdoors in Austin, Texas. Table 4.2 summarizes the CPT data at
one year and the exposure block data after approximately 3.5 years of outdoor exposure. Of
most significance is the mixture containing the low-alkali cement (Na20e=0.52 percent), whose
prisms easily passed the 0.04 percent expansion criteria after the one-year ASTM C 1293 test
but whose exposure blocks expanded and cracked significantly. The reason for this discrepancy
is the leaching from the concrete prisms which dropped the concrete below its alkali threshold,
thereby suppressing expansion. The larger exposure blocks were less susceptible to leaching
and hence exhibited significant expansion and cracking.
Based on these results, and results from other researchers, it is not recommended that one use
ASTM C 1293 to determine the alkali threshold for a given aggregate or aggregate-binder
combination.
Table 4.2. Comparison of Expansions in ASTM C 1293 Compared to Outdoor Exposure
Blocks, Highlighting Importance of Leaching in Concrete Prism Test (Folliard et al. 2006)
Cement
Alkalinity
ASTM C 1293
Expansion at one year
Outdoor Exposure Blocks Expansion

after approximately 3.5 years
0.52
0.01
0.88
0.95
0.33
1.10
1.25
0.59
1.07
Another drawback to ASTM C 1293 is that the test duration (one year for aggregates, two years
for preventive measures) is often deemed to be too long, especially in the United States. As
such, very few agencies specify this test and very few practitioners run the test. The excessive
duration is due to the fact that the storage conditions (storing real concrete prisms above water
at 38C) are nowhere near extreme as the AMBT, and the source of alkalis is finite and internal.
Researchers have attempted to accelerate the test by increasing the storage temperature to 60C,
in hopes of generating relevant data in a few months instead of 1-2 years. However, as reported
recently by Ideker et al. (2010), this accelerated version of the CPT yields significantly less
expansion than the standard CPT for the following reasons:
Additional leaching of alkalis at 60C, compared to 38C
Additional prism drying observed at 60C, compared to 38C
Changes in pore solution at 60C (Ettringite begins to dissolve, contributing sulfates to
the pore solution at the expense of hydroxyl ions, thus lowering pore solution pH.)
The non-reactive sand used in conjunction with the subject coarse aggregate has
major effect on expansions at 60 C.
56
AAR Facts Book
Figure 4.4 shows the relationship between expansion after one year using the standard CPT at
38C and the expansion after 13 weeks using the accelerated CPT at 60C (Ideker et al. 2010)
for a range of aggregates (note all mixtures are plain concrete). A test duration of 13 weeks is
based on previous research by the same authors where it was observed that most aggregates
have exhibited most of their expansion by this point (see Figure 4.5). In addition, other
researchers have proposed that expansions at 13 weeks in the accelerated CPT correlate best
with one-year expansions in the standard CPT when testing plain concrete (without SCMs or
lithium-based admixtures).
Figure 4.4. Relationship between Expansion after One Year Using the Standard CPT (38C)
and the Expansion after 13 Weeks Using the Accelerated CPT (60C) (Ideker et al. 2010)
57
AAR Facts Book
Figure 4.5. Expansion of Concrete Prisms Containing Spratt Limestone and Stored
above Water at 38C vs. 60C (Folliard et al. 2006)
Referring to Figure 4.4, it is quite evident that most of the data points are well below the line of
equity, which represents expansion values that are identical for both test methods/durations. As
such, the general trend is that expansion values obtained at 60C are significantly lower than
those at 38C for the aforementioned reasons. For some of the aggregates tested by Ideker et al.
(2010), the results of testing at the two different temperatures yielded similar classifications
when using a 0.04 percent expansion limit, meaning that some aggregates passed both tests or
failed both tests. From a specification point of view, this may be acceptable in that both
versions of the CPT would conclude that the aggregate is either reactive or non-reactive (and
such requiring preventive measures). However, there were also aggregates that yielded
conflicting results when they were tested at the two different temperatures, with the
classification being fail at 38C and pass at 60C.
Because the two versions of the CPT can produce conflicting conclusions with regard to a 0.04
percent expansion criteria, caution should be taken when considering the accelerated CPT, by
itself or in parallel to the standard CPT, for a given application or project.
4.3 ASR TEST METHODS FOR EVALUATING PREVENTIVE MEASURES
When considering the use of reactive aggregates, it is essential that test methods exist that can
determine the safe level or dosage of a given product to prevent or substantially reduce ASR
58
AAR Facts Book
induced expansion and cracking. This section will discuss and critique the most commonly used
test methods that aim to achieve this goal.
Table 4.3 details the most common test methods used to evaluate preventive measures, such as
SCMs and lithium-based admixtures. Specific details are listed in the table for each test, but
additional focus of this section is on ASTM C 1567, which is a modified version of AASHTO T
303, and ASTM C 1293 as these two methods are the most commonly used test methods and
the ones that are integrated into AASHTO PP65-11.
Table 4.3. Test Methods for Evaluating Preventive Measures
Test Method
Comments
ASTM C 441: Standard Test Method for Mortar bar test, intended to assess effectiveness of SCMs in
Effectiveness of Mineral Admixtures or Ground
reducing ASR expansion.
Blast-Furnace Slag in Preventing Excessive Test uses high-alkali cement and PyrexTM glass.
Expansion of Concrete Due to the Alkali-Silica Test not very reliable because of the use of Pyrex glass, which is
Reaction
sensitive to test conditions and contains alkalis that may be released
during the test. Test does not correlate well with data from concrete
mixtures containing natural aggregates (Brub and Duchesne
ASTM C 1567: Standard Test Method for Mortar bar test, originally designed to assess aggregate reactivity.
Determining the Potential Alkali-Silica Reactivity of Bars are soaked in 1N NaOH solution for 14 days.
Combinations of Cementitious Materials and Should only be used for aggregates for which a reasonable
Aggregate (Accelerated Mortar-Bar Method)
correlation between AASHTO T 303 (ASTM C 1260) and ASTM C
1293 has been established (see AASHTO PP65-11 for details).
ASTM C 1567 can be modified as per AASHTO PP65-11 to
evaluate lithium nitrate-based admixtures, although ASTM C 1293
is deemed to be a more effective test method for determining
ASTM C 1293: Standard Test Method for Concrete Modified version of concrete prism test, allows for testing SCMs,
blended cements, and lithium admixtures.
Test duration is two years, with expansion limit of 0.04 percent.
4.3.1 ASTM C 441: Standard Test Method for Effectiveness of Mineral Admixtures or Ground
Blast-Furnace Slag in Preventing Excessive Expansion of Concrete Due to the Alkali-Silica
Reaction
This test method was developed by U.S. Bureau of Reclamation in the 1940s to evaluate SCMs
(known as siliceous admixtures or pozzolans at the time) for the construction of the Davis Dam
and other large projects (Gilliland and Moran 1949). The test is essentially the same as ASTM
C 227, where small specimens (25x25 mm in cross section) are stored above water at 38C,
with one major difference Pyrex (borosilicate) glass is used as a model reactive aggregate.
The objective of the test is to have somewhat of an internal standard, Pyrex glass, that can be
used to evaluate the relative efficacy of SCMs in reducing expansion. Specifically, the
59
AAR Facts Book
expansion after 14 days of storage above water at 38C of mortar bars containing a high-alkali
cement (0.95 to 1.05% Na2Oe) and either 25% fly ash or 50% slag cement is compared to the
expansion of control mortar bars (portland cement only). The percentage reduction in expansion
is then calculated. Although the test specifies the use of either 25% fly ash or 50% slag cement,
other SCMs and dosages may also be used. Various ASTM specifications allow for the use of
ASTM C 441 to assess the efficacy of SCMs in controlling ASR-induced expansion, as
summarized in Table 4.4.
Table 4.4. ASTM Materials Specifications and Expansion Limits for ASTM C 441
Specification
Limit
ASTM C 618: Fly Ash and Natural

Pozzolans
14-day expansion of cement/fly ash blend expansion of control mix

with low-alkali cement ( 0.60% Na2Oe)
ASTM C 989: Ground Granulated Blast If job cement and proportions of cement and slag are known: 14-day
-Furnace Slag
expansion 0.020%
Otherwise: 14-day expansion of cement/slag blend 25% of the
expansion of control mix with the high-alkali cement only (i.e., slag
should reduce expansion by at least 75% compared with the control)
ASTM C 1240: Silica Fume
14-day expansion of blend of 90% cement and 10% silica fume 20%
of the expansion of control mix with the high-alkali cement only (i.e.,
10% silica fume should reduce expansion by at least 80% compared
with the control)
ASTM C 595: Blended Hydraulic

Cement
14-day expansion of blended cement 0.020%

and
56-day expansion of blended cement 0.060%
ASTM C 1157: Hydraulic Cement

(Option R)
14-day expansion of hydraulic cement 0.020%

and
56-day expansion of hydraulic cement 0.060%
Although ASTM C 441 is referenced by all of the materials specifications shown in Table 4.4,
this test method has numerous technical issues and has fallen out of favor with most agencies
and owners. The test inherently suffers from the same major flaw described earlier for ASTM C
227, specifically leaching of alkalis from the mortar bars. But in addition to this, the use of
Pyrex glass adds other complexities and problems that have led to high variability in test
results. Although Pyrex glass was selected with the intention of it being a model or
standard reactive aggregate, several issues with the glass have limited the widespread use of
ASTM C 441 to evaluate SCMs, in favor of ASTM C 1567 (described next) and ASTM C
1293. Pyrex glass varies in its composition and particle size from source to source. Further,
Pyrex glass, itself, contains relatively high alkali contents, which are released during the course
of the test, and different sources of Pyrex glass will vary in alkali content. Hobbs (1989)
showed that the amount of SCMs needed to reduce expansion triggered by Pyrex glass does not
60
AAR Facts Book
correlate well with the SCM content needed to suppress expansion from natural aggregates used
in concrete construction. Because of the technical issues with ASTM C 441, the high level of
variability of test results, and the lack of correlation between Pyrex glass and natural
aggregates, this test is not recommended as a means of evaluating SCMs for use in controlling
ASR in concrete.
4.3.2 ASTM C 1567: Standard Test Method for Determining the Potential Alkali-Silica
Reactivity of Combinations of Cementitious Materials and Aggregate (Accelerated Mortar-Bar
Method)
ASTM C 1567 is essentially identical to AASHTO T 303, with the one exception that it is
aimed at evaluating the use of SCMs. The test utilizes the same test specimens and storage
conditions as AASHTO T 303 and is, as such, subject to the same issues and pitfalls as the
standard version of the AMBT. First and foremost, if the aggregate to be evaluated does not
produce accurate results when tested under AASHTO T 303, due to false negatives or false
positives, it should be quite evident that the AMBT is not a suitable test for evaluating the
specific aggregate in combination with SCMs to determine the dosage needed to control
expansion. This is recognized in AASHTO PP65-11, as shown in Figure 4.6, where it is
recommended that a reasonable correlation between AASHTO T 303 and ASTM C 1293 must
first be developed for a given aggregate of interest; only upon obtaining such agreement
between test methods can ASTM C 1567 then be used to evaluate preventive measures, such as
SCMs, blended cements, and lithium admixtures.
Expansion in AMBT at 14 Days (%)
1.0
If AMBT vs CPT data fall

within this range, the AMBT
may be used to evaluate
preventive measures.
0.8
0.6
0.4
0.2
0
0
0.1
0.2
0.3
0.4
0.5
0.6
Expansion in CPT at 1 Year (%)
Figure 4.6. Comparison of AMBT and CPT Data for the Purpose of Determining Whether
the AMBT is Suitable for Evaluating Preventive Measures with a Specific Aggregate
(after AASHTO PP65-11)
61
AAR Facts Book
Once it has been determined that ASTM C 1567 is an appropriate test method for evaluating a
given aggregate, past work has shown that there is generally a good agreement between ASTM
C 1567 (using a 0.10 percent expansion limit at 14 days) and ASTM C 1293 (using a 0.04
percent expansion limit at two years) when testing various SCMs (Thomas and Innis 1998), as
shown in Figure 4.7. This figure illustrates that similar SCM dosages are needed in each of
these tests to reduce the expansions below the aforementioned limits.
'Safe Level' in Concrete Test (%)
60
50
40
S - Slag
30
F - Fly Ash
20
CI - Fly Ash
CH - Fly Ash
10
N - Pozzolan
0
0
10
20
30
40
50
60
'Safe Level' in M ortar Bar Test (%)
Figure 4.7. Comparison of SCM Dosage Needed to Reduce Expansions below Test Limits for
ASTM C 1567 and ASTM C 1293 (after Thomas and Innis 1998)
When considering using lithium nitrate as an admixture in concrete, AASHTO PP65-11
recommends ASTM C 1293 as the most suitable and accurate test for determining the requisite
dosage to control expansion below an expansion limit of 0.04 percent at two years. The AMBT,
in its current form, is not suited for testing lithium compounds as any lithium present in the
mortar bars would be overwhelmed by the sodium hydroxide in the soak solution. AASHTO
PP65-11 provides guidance, based on Tremblay et al. (2008), for how to test lithium
compounds in a modified version of the AMBT. This approach recognizes that certain
aggregates cannot be tested using the AMBT in conjunction with lithium nitrate as erroneous
results are generated that do not correlate with actual performance in concrete prisms or
exposure blocks. As such, the procedure recommended in AASHTO PP65-11 identifies such
aggregates and requires that these aggregates only be tested using ASTM C 1293. Aggregates
that are suitable for being tested in combination with lithium nitrate in the modified AMBT are
also identified, and a testing regime is then specified that will arrive at an estimated dosage of
lithium to control expansion, based on an extensive database that is correlated with CPT results.
A broad overview of the approach specified in AASHTO PP65-11 is shown in Figure 4.8;
readers are directed to Tremblay et al. (2008) for a more detailed description of this
methodology for evaluating lithium nitrate in the modified AMBT (or in the CPT if it is deemed
not possible to use the modified AMBT for a given aggregate).
62
AAR Facts Book
Figure 4.8. Approach to Using the Modified Accelerated Mortar Bar Test to Determine the
Dosage of Lithium Nitrate (expressed as [Li]/ [Na+K]) to Adequately Suppress Expansion
in Concrete (Tremblay et al. 2008)
Length Change of Concrete Due to Alkali-Silica Reaction
The concrete prism test was initially developed for evaluating aggregate reactivity, but the most
recent version, ASTM C 1293, has been modified to allow for testing SCMs as means of
preventing or reducing deleterious expansion. The procedures and testing regimes are identical
to those previously described for evaluating aggregate reactivity (see section 4.2.5), except for
two significant modifications: (1) test duration is two years, and (2) SCMs may be used as
replacement for portland cement (replacement by mass).
In addition, when testing SCMs that significantly increase water demand, such as silica fume,
ASTM C 1293 allows for the use of high-range water reducers (HRWR) meeting ASTM C 494.
The dosage of HRWR may be adjusted to ensure adequate workability and sample preparation
while still achieving the target w/cm of 0.42-0.45. Conversely, if SCMs are used that
substantially reduce water demand, such as high-volume fly ash mixtures, a viscositymodifying admixture (VMA) may be used to minimize segregation and bleeding, while still
achieving the target w/cm of 0.42 to 0.45.
63
AAR Facts Book
Testing SCMs using ASTM C 1293 carries with it the same inherent shortcomings encountered
when testing aggregates in the concrete prism test. It is not possible to accurately assess the
effects of cement alkalis (and to some extent alkalis from SCMs) due to leaching of alkalis from
the prisms over a two-year period. Concerns raised over the long-term nature of ASTM C 1293
when evaluating aggregates, which takes one year, are elevated when testing SCMs, as two
years are required. This extended, two-year testing period is necessary to evaluate preventive
measures, such as SCMs and lithium-based admixtures.
However, because two years are need to test preventive measures using ASTM C 1293, it has
led most agencies and owners in the United States to rely on ASTM C 1567 (accelerated mortar
bar test) for evaluating and selecting preventive measures. As mentioned in the previous
section, ASTM C 1567 is a suitable test for evaluating preventive measures, provided that the
aggregate yields comparable results when tested using AASHTO T 303 and ASTM C 1293 (see
Figure 4.6). For the aggregates meeting this criteria, AASHTO PP65-11 recommends the use of
an expansion limit for ASTM C 1567 of 0.10 percent after 14 days as this correlates best with
0.04 percent expansion after two years testing the same SCM(s) following ASTM C 1293
(Thomas et al. 2007).
Some agencies have specified a 28-day expansion criteria when evaluating SCM/aggregate
combinations, but increasing from a 14- to 28-day expansion limit in ASTM C 1567 results in
SCM levels significantly higher (that is, by 1.5 times on average) than that required to prevent
damage in concrete, based on concrete prism tests and/or performance of outdoor exposure
blocks (Thomas et al. 2007). The perception that much higher levels of SCM are required to
control ASR may present a barrier to using such materials when potentially reactive aggregates
are used due to concerns related to slow setting and strength gain, and even deicer salt scaling
resistance.
4.3.4 Other ASR Test Methods
There are several other ASR test methods that have been developed and used, besides the ones
previously described in this chapter, since the discovery of ASR in the late 1930s. Some of
these methods are described briefly herein for completeness, including accelerated tests aimed
at evaluating aggregate reactivity and tests related to alkali-release from aggregates and SCMs.
Although the latter tests, those focusing on the release of alkalis from aggregates and/or SCMs,
are not direct tests for ASR (e.g., expansion tests), they are discussed here because of the
importance of alkali release in field structures. Although no standardized methods exist for
accurately predicting the release of alkalis from aggregates and/or SCMs, some relevant studies
and published literature are cited that should provide useful information on this important topic.
64
AAR Facts Book
Over the years, there has been interest in and research related to ultra-accelerated ASR test
methods, where ASR is accelerated through the use of an autoclave, which can subject mortar
specimens to high temperatures and pressures. Fournier et al. (1991) and Nishibayashi et al.
(1996) conducted extensive work on autoclaved mortar tests these tests generated results
within a couple days (typically less than 24 hours) through the use of extremely high alkali
loadings (2.0 to 3.5% Na2Oe by mass of cement), high autoclave pressures (up to 1.03 MPa),
and high temperatures (up to 130C). Fournier et al. (1991) and Giannini (2012) reported that
the autoclaved test was as good as or better than the accelerated mortar bar test (AASHTO T
303) at identifying reactive aggregates. There has only been limited work done on testing
concrete prisms in an autoclave test (Nishibayashi et al. 1996). Overall the interest in autoclave
testing lies in the rapid results, accomplished through high alkalinity, high temperature, and
high pressure. This test is very aggressive and unrealistic compared to field exposure, and its
only benefit would be that it generates results, for good or bad, in about a day instead of two
weeks, as in AASHTO T 303.
The Concrete Microbar Test (formerly the Chinese Accelerated Mortar Bar Method), was
introduced by Xu and co-workers to capture reactivity of alkali-carbonate rocks (Xu et al. 1998;
2000), but has since been used to identify aggregates susceptible to ASR (Lu et al. 2008; East
2007). This mortar test uses larger specimens (40 mm x 40 mm cross section) and thus allows
for testing larger aggregate particles, preserving the textural and mineralogical characteristics of
the original coarse aggregate. This test is also advantageous for aggregates in which reactive
phases are removed due to crushing and processing required in other accelerated mortar tests,
such as AASHTO T 303 and ASTM C 227.
Past laboratory and field evaluations have demonstrated that some aggregates may release
significant amounts of alkalis into the pore solution of concrete, thus increasing the likelihood
of ASR-induced expansion and cracking (Stark 1980; Thomas et al. 1992; Brub et al. 2002).
There are no standardized tests for estimating the alkali release from aggregates. There is a
general agreement that the total alkali content of an aggregate, as measured through bulk
chemical analysis, is not indicative of the available alkalis that will be released from aggregates
into the pore solution in concrete. Recent work by Brub et al. (2002) has shown promise in
better estimating the alkali release from a given aggregate by measuring the release of
potassium from the aggregate immersed in sodium hydroxide and measuring the release of
sodium from the same aggregate immersed in potassium hydroxide. This technique maintains
the type of alkaline environment in which the aggregate will reside and selectively promotes the
release of the two critical alkalis present in aggregates that may have an adverse effect on ASR.
The release of alkalis from SCMs, in general, and fly ash, in particular, into the pore solution of
concrete has been the focus of considerable research over the years. ASTM C 311 (Standard
Test Methods for Sampling and Testing Fly Ash or Natural Pozzolans for Use in Portland
65
AAR Facts Book
Cement Concrete) is a standard test that measures the total alkali content, as well as the
available alkali content of fly ash. However, research has shown that ASTM C 311 does not
accurately measure the available alkalis from within fly ash, mainly because the aqueous
solution in which the fly ash is immersed (pH = 7) is vastly different from the highly-alkaline
pore solution found in mature concrete, with pH values typically in excess of 13.2. To attempt
to generate data that is more relevant to actual alkali release from fly ash to concrete pore
solution, a method was developed by Shehata and Thomas (2006) in which mature paste
specimens, containing subject fly ash, are immersed in various alkaline solutions that are
intended to better mimic the pore solution in concrete (e.g., 0.25 OH- mol/L) undergoing ASR.
The leaching of alkalis from the fly ash paste specimens is measured at the end of a 90-day
immersion period, thereby directly measuring the alkali contribution from the fly ash to the host
solution. Although the test developed by Shehata and Thomas (2006) is not a standardized test,
it does show promise in being able to better estimate the release of alkalis from fly ash to
concrete pore solution under realistic conditions.
4.4 ACR TEST METHODS FOR EVALUATING AGGREGATE REACTIVITY
ACR, described in chapter 9, is a unique form of AAR that is highly expansive and damaging to
concrete structures. It is also a reaction for which there are no known preventive measures.
Measures that are effective in controlling ASR, such as the use of low-alkali cement, SCMs,
and lithium compounds, are ineffective in suppressing ACR-induced expansion. As such, the
only safe strategy for dealing with ACR is to identify susceptible aggregates and disallow their
use in concrete. This section briefly describes how practitioners and researchers can identify
ACR-susceptible aggregates.
AASHTO PP65-11 provides guidance, based primarily on Canadian experience, on how to
evaluate aggregates that are potentially susceptible to ACR-induced expansion and cracking.
The test methods recommended for evaluating aggregates potentially susceptible to ACR are
presented in Table 4.5.
66
AAR Facts Book
Table 4.5. Test Methods for Evaluating Alkali-Carbonate Reactivity of Aggregates

Test Method
Comments
ASTM C 295: Standard Guide for Petrographic Useful evaluation to identify many (but not all) potentially reactive
Examination of Aggregates for Concrete
components in aggregates.
Reliability of examination depends on experience and skill of
individual petrographer.
Petrographer focuses specifically on calcareous dolomites or
dolomitic limestones with clayey insoluble residues, some dolomites
essentially free of clay, and some very fine-grained limestones free
of clay and with minor insoluble residue (mostly quartz).
CSA A23.2-26A: Determination of Potential Alkali- Measures the CaO, MgO, and Al2O3 content of quarried carbonate
Carbonate Reactivity of Quarried Carbonate Rocks
rocks.
by Chemical Composition
Chemical composition plotted (see Figure 4.10) and classified as
nonexpansive or potentially expansive. Guidance, including
additional testing such as ASTM C 1105, is provided based on
classification.
ASTM C 1293: Standard Test Method for Concrete Test method developed originally for evaluating aggregates
susceptible to ASR.
Test method will also trigger ACR if such susceptible aggregates are
present in mixture petrography is required to confirm ACR is the
cause of part or all of the measured expansion.
ASTM C 1105: Standard Test Method for Length Similar test to ASTM C 1293, except lower cement alkali content is
Change of Concrete Due to Alkali-Carbonate Rock
used. ASR will not occur at lower alkali loading, and as such, any
Reaction
observed expansion is assumed to be caused by ACR.
Expansion limits are 0.025 percent at 6 months or 0.030 percent at 1
year. Aggregates exceeding limit(s) are considered to be alkalicarbonate reactive and shall not be used in concrete.
A petrographic evaluation (ASTM C 295) is recommended when assessing the potential for an
aggregate being susceptible to ACR. Aggregates prone to ACR are typically calcareous
dolomites or dolomitic limestones with clayey insoluble residues. Figure 4.9 shows a unique
petrographic feature of ACR rhombs of dolomite floating in a fine-grained matrix of clay
minerals, calcite and silica.
67
AAR Facts Book
Figure 4.9. A Unique Petrographic Feature Typical of ACR Susceptible Aggregates: Rhombs of
Dolomite Floating in a Fine-grained Matrix of Clay Minerals, Calcite and Silica
(photo courtesy of C. Rogers)
If the aggregate being assessed is a quarried carbonate rock, the potential for alkali-carbonate
reaction may be assessed on the basis of its chemical composition using the test method CSA
A23.2-26A. This test involves the determination of the lime (CaO), magnesia (MgO), and
alumina (Al2O3) content of the rock, and determining where the composition of the rock falls on
a plot of CaO/MgO ratio versus the Al2O3 content, as shown in Figure 4.10.
Rock Cylinder ExpansionTest (ASTM C 586)
Concrete Prism Expansion Test (CSA A23.2-14A)
<0.1% expansion at 4 weeks
<0.025% expansion at 1 year

in 1.25% Na2O equivalent cement
<0.1% at 4 weeks but >0.2% at 16 weeks

>0.1% expansion at 4 weeks
<0.025% expansion at 1 year

in 1.25% Na2O equivalent cement
200
Aggregates
considered
nonexpansive
CaO/MgO Ratio
100
50
Aggregates
considered
potentially
expansive
20
10
2
Aggregates considered nonexpansive
1
0
10
Al2O3 in percent
Figure 4.10. Using Chemical Composition as a Basis for Determining Potential AlkaliCarbonate Reactivity of Quarried Carbonates (from CSA A23.2-26A)
68
AAR Facts Book
If the composition falls in one of the two ranges identified as aggregates considered nonexpansive in Figure 4.10, the aggregate is not potentially alkali-carbonate reactive. However,
the aggregate should be tested to determine the potential for alkali-silica reaction using the
methods described in section 4.2.
If the composition falls in the range of aggregates considered to be potentially expansive in
Figure 4.10, the aggregate is potentially alkali-carbonate reactive and must be evaluated further.
There are two options for further testing. One option is to test the aggregate in the concrete
prism test, ASTM C 1293, to simultaneously determine the potential for alkali-carbonate and
alkali-silica reactivity. After the test, the prisms are examined by petrography to determine the
role played by the alkali-carbonate reaction. The second option is to test using the specific
concrete prism test developed for ACR-susceptible aggregates, ASTM C 1105, which employs
a reduced alkali loading to determine the potential for alkali-carbonate reaction only. If
expansion of the specific aggregate-cement combination is equal to or greater than 0.025
percent at 6 months or 0.030 percent at 1 year, the aggregate shall be considered to be alkalicarbonate reactive and shall not be used in concrete. If the specific aggregate-cement
combination passes the above expansion criteria, it is considered not to be alkali-carbonate
reactive but should still be evaluated for potential ASR reactivity (see section 4.2).
4.5 SUMMARY
This chapter briefly described the various test methods that have been or are currently being
used to evaluate aggregate reactivity (with regard to ASR and ACR) and to evaluate preventive
measures against ASR, such as the use of SCMs or lithium-based admixtures. The test methods
recommended in this chapter are consistent with those integrated into AASHTO PP 65-11, the
first nationwide recommended practice aimed at preventing ASR in new concrete construction.
It is worth noting that the expansion limits and performance criteria recommended in this
chapter are based extensively on the performance of large exposure blocks stored at outdoor
exposure sites in the United States and Canada. As new data emerges from these sites and as
blocks that had not previously expanded start to expand and crack, the expansion limits and
performance criteria will evolve. It is hoped that future evolutions will increase the accuracy of
AAR test methods and will also spur the development of improved test methods that better
predict field performance.
69
AAR Facts Book
REFERENCES
AASHTO. 2011. Standard Practice for Determining the Reactivity of Concrete Aggregates and
AASHTO. 2008. Standard Method of Test for Accelerated Detection of Potentially
Deleterious Expansion of Mortar Bars Due to Alkali-Silica Reaction. AASHTO T 303,
American Association of State and Highway Transportation Officials, Washington, DC, 6 p.
Arrieta, G. 2012. Experimental studies of the behavior of pessimum aggregates in different
test procedures used to evaluate the alkali reactivity of aggregates in concrete, M.S. Thesis, The
University of Texas at Austin, Austin, TX, USA.
ASTM. 2013. Standard Specification for Concrete Aggregate. ASTM C 33, ASTM
International, West Conshohocken, PA, 11 p.
PA, 6 p.
ASTM. 2007. Standard Test Method for Potential Alkali-Silica Reactivity of Aggregates
(Chemical Method). ASTM C 289, ASTM International, West Conshohocken, PA, 7 p.
ASTM. 2012. Standard Guide for Petrographic Examination of Aggregates for Concrete.
ASTM C 295, ASTM International, West Conshohocken, PA, 9 p.
ASTM. 2013. Standard Test Methods for Sampling and Testing Fly Ash or Natural Pozzolans
for Use in Portland-Cement Concrete. ASTM C 311, ASTM International, West
Conshohocken, PA, 10 p.
ASTM. 2011. Standard Test Method for Effectiveness of Pozzolans or Ground Blast-Furnace
Slag in Preventing Excessive Expansion of Concrete Due to the Alkali-Silica Reaction. ASTM
C 441, ASTM International, West Conshohocken, PA, 3 p.
ASTM. 2012. Standard Specification for Blended Hydraulic Cements. ASTM C 595, ASTM
ASTM. 2012. Standard Specification for Coal Fly Ash and Raw or Calcined Natural Pozzolan
for Use in Concrete. ASTM C 618, ASTM International, West Conshohocken, PA, 5 p.
70
AAR Facts Book
ASTM. 2012. Standard Specification for Slag Cement for Use in Concrete and Mortars.
ASTM. 2008. Standard Test Method for Length Change of Concrete Due to Alkali-Carbonate
Rock Reaction. ASTM C 1105, ASTM International, West Conshohocken, PA, 4 p.
ASTM. 2011. Standard Performance Specification for Hydraulic Cement. ASTM C 1157,
ASTM International, West Conshohocken, PA, 5 p.
ASTM. 2012. Standard Specification for Silica Fume Used in Cementitious Mixtures. ASTM
ASTM. 2007. Standard Test Method for Potential Alkali Reactivity of Aggregates (Mortar-Bar
Method). ASTM C 1260, ASTM International, West Conshohocken, PA, 5 p.
ASTM. 2008. Standard Test Method for Determination of Length Change of Concrete Due to
Alkali-Silica Reaction. ASTM C 1293, ASTM International, West Conshohocken, PA, 7 p.
ASTM. 2013. Standard Test Method for Determining the Potential Alkali-Silica Reactivity of
Combinations of Cementitious Materials and Aggregate (Accelerated Mortar Bar Method.
Berube, M-A., Duchesne, J., Doriona, J. and Rivest, M. 2002. Laboratory assessment of alkali
contribution by aggregates to concrete and application to concrete structures affected by alkali
silica reactivity. Cement and Concrete Research, 32: 12151227.
Brub, M.A. and Duchesne, J. 1992. Evaluation of Testing Methods Used for Assessing the
Effectiveness of Mineral Admixtures in Suppressing Expansion Due to Alkali-Aggregate
Reaction. Proceedings of the Fourth International Conference on Fly Ash, Silica Fume, Slag,
and Natural Pozzolans in Concrete, (Ed. V.M. Malhotra), ACI SP132, Vol. 1, American
Concrete Institute, Detroit, MI, 549575.
Brub, M.A. and Fournier, B. 1992. Accelerated Test Methods for Alkali-Aggregate
Reactivity. Advances in Concrete Technology, (Ed. V.M. Malhotra), CANMET/EMR, Ottawa,
Canada, 583627.
Brub, M.-A. and Fournier, B. 1993. Testing for Alkali-Aggregate Reactivity in Concrete.
Memoria del Seminario Internacional sobre Tecnologia del Concreto, (Ed. R. Rivera Villareal),
Universidad Autonoma de Nuevo Leon, Nuevo Leon, Mexico, October, 54-78.
CSA. 2009. "Test Method for Detection of Alkali-Silica Reactive Aggregate by Accelerated
Expansion of Mortar Bars. CSA A23.2-25A. A23.2-09 - Test methods and standard practices
for concrete, Canadian Standards Association, Mississauga, Ontario, Canada.
71
AAR Facts Book
CSA. 2009. "Determination of Potential Alkali-Carbonate Reactivity of Quarried Carbonate

Rocks by Chemical Composition." CSA A23.2-26A. A23.2-09 - Test methods and standard
practices for concrete, Canadian Standards Association, Mississauga, Ontario, Canada.
CSA. 2009. "Standard Practice to Identify Degree of Alkali-Reactivity of Aggregates and to
Identify Measures to Avoid Deleterious Expansion in Concrete. CSA A23.2-27A. A23.2-09 Test methods and standard practices for concrete, Canadian Standards Association,
Mississauga, Ontario, Canada.
East, B.L. 2007. Laboratory and Field Investigations on the Use of Lithium Nitrate to Prevent
or Mitigate Alkali-Silica Reaction. M.S. Thesis, The University of Texas at Austin, Austin,
TX, USA, 400 p.
Folliard, K.J., Barborak, R., Drimalas, T., Du, L., Garber, S., Ideker, J., Ley, T., Williams, S.,
Juenger, M., Thomas, M.D.A. and Fournier, B. 2006. Preventing ASR/DEF in New Concrete:
Final Report. The University of Texas at Austin, Center for Transportation Research (CTR),
CTR 4085-5.
Fournier, B., Berube, M. and Bergeron, G. 1991. A Rapid Autoclave Mortar Bar Method to
Determine the Potential Alkali-Silica Reactivity of St. Lawrence Lowlands Carbonate
Aggregates (Quebec, Canada). Cement, Concrete, and Aggregates, 13(1): 58-71.
Fournier, B. and Brub, M.A. 1991. Application of the NBRI Accelerated Mortar Bar Test to
Siliceous Carbonate Aggregates Produced in the St. Lawrence Lowlands (Qubec, Canada)
Part II: Proposed Limits, Rates of Expansion, and Microstructure of Reaction Products.
Cement and Concrete Research, 21:10691082.
Giannini, E. 2012. Unpublished data.
Gilliland, J.L. and Moran, W.T. 1949. "Siliceous admixture for the Davis Dam." Engineering
News Record, 3 February, 62.
Grattan-Bellew, P.E. 1978. Study of Expansivity of a suite of quartzites, argillites and quartz
arenites. Proceedings of the 4th International Conference on the Effect of Alkalies in Cement
and Concrete, Purdue University, Indiana, 113-140.
Grattan-Bellew, P.E. 1989. Test Methods and Criteria for Evaluating the Potential Reactivity
of Aggregates. Proceedings of the 8th International Conference on Alkali-Aggregate Reaction
in Concrete, 279294.
Hobbs, D.W. 1989. Effect of Mineral and Chemical Admixtures on Alkali-Aggregate
Reaction. Proceedings of the 8th International Conference on Alkali-Aggregate Reaction,
(Eds. K. Okada, S. Nishibayashi, and M. Kawamura), E&FN Spon, 173-186.
72
AAR Facts Book
Hooton, R.D. 1990. Case Studies of Ontario Hydros Experience with Standard Tests for
Alkali-Aggregate Reactivity. Canadian Developments in Testing Concrete Aggregates for
Alkali-Aggregate Reactivity, Ontario Ministry of Transportation Engineering Materials Report
92: 181189.
Hudec, P.P. 1990. Rapid test for predicting alkali reactivity: promises and problems.
Canadian developments in testing concrete aggregates for alkali-aggregate reactivity,
University of Windsor Geology Department, Ontario, 111-117.
Lu, D., Fournier, B., Grattan-Bellew, P., Xu, Z. and Tang, M. 2008. "Development of a
Universal Accelerated Test for Alkali-silica and Alkali-carbonate Reactivity of Concrete
Aggregates." Materials and Structures, 41(2): 235-246.
Nishibayashi, S., Kuroda, T. and Inoue, S. 1996. Expansion Characteristics of AAR in Concrete
by Autoclave Method. (Ed. A. Shayan), Melbourne, Australia, 370-376.
Nixon, P.J., Page, C.L., Hardcastle, J., Canham, I. and Pettifer, K. 1989. Chemical studies of
alkali silica reaction in concrete with different flint contents. 8th International Conference on
Alkali-Aggregate Reaction, Kyoto, Japan.
Oberholster, R.E. and Davies, G. 1986. An Accelerated Method for Testing the Potential
Reactivity of Siliceous Aggregates. Cement and Concrete Research, 16: 181189.
Poole, A.B. 1992. Introduction to alkali-aggregate reaction in concrete. The alkali-silica
rection in concrete, (Ed. R.N. Swamy), New York: Van Nostrand Reinhold, 1-28.
Shehata, M.H. and Thomas, M.D.A. 2006. "Alkali Release Characteristics of Blended
Cements." Cement and Concrete Research, 36: 11661175.
Stark, D. 1980. Alkali-silica reactions: some reconsiderations. Cement, Concrete and
Aggregates, 2(2): 92-94.
Swenson, E.G. 1957. A reactive aggregate undetected by ASTM test. ASTM Bulletin No.
226: 4850.
Thomas, M.D.A., Blackwell, B.Q. and Pettifer, K. 1992. Suppression of Damage from AlkaliSilica Reaction by Fly Ash in Concrete Dams. Proceedings of the 9th International Conference
on Alkali-Aggregate Reaction in Concrete, Vol. 2, The Concrete Society, Slough, UK, 1059
1066.
73
AAR Facts Book
Thomas, M.D.A. and Innis, F.A. 1998. Effect of Slag on Expansion Due to Alkali-Aggregate
Reaction in Concrete. ACI Materials Journal 95 (December): 71624.
Thomas, M. D. A., Fournier, B., Folliard, K.J., Shehata, M.H., Ideker, J.H. and Rogers, C.
2007. Performance Limits for Evaluating Supplementary Cementing Materials Using
Accelerated Mortar Bar Test. ACI Materials Journal, 104(2): 115-122.
Thomas, M.D.A., Fournier, B. and Folliard, K.J. 2008. Report on determining the reactivity of
concrete aggregates and selecting appropriate measures for preventing deleterious expansion in
new concrete construction. Federal Highways Administration, FHWA-HIF-09-001, National
Research Council, Washington, D.C., A34.
Tremblay, C., Brub, M.A., Fournier, B., Thomas, M.D.A., Folliard, K.J. and Nkinamubanzi,
P.C. 2008. Use of the Accelerated Mortar Bar Test to Evaluate the Effectiveness of LiNO3
Against Alkali-Silica Reaction -- Part 2: Comparison with Results from the Concrete Prism
Test. Journal of ASTM International (JAI), Online ISSN: 1546-962X, Published Online: 17
September 2008, 21 p.
Xu, Z., Shen, Y., Lu, D., Deng, M., Lan, X., Hu, R. and Tang, M. 1998. "Investigation on A
New Test Method for Determining the Alkali Silica Reactivity of Aggregates." Journal of
Nanjing University of Chemical Technology, 20(2): 1-7.
Xu, Z., Lan, X., Deng, M. and Tang, M. 2000. A New Accelerated Method for Determining
the Potential Alkali-Carbonate Reactivity. Proceedings of the 11th International Conference on
Alkali-Aggregate Reactivity, (Eds. M.A. Brub, B. Fournier, and B. Durand), CRIB, SainteFoy Qubec, Canada, June, 129-138.
74
AAR Facts Book
Chapter 5 Prevention of ASR
5 Prevention of Alkali-Silica Reaction
5.1 INTRODUCTION
This chapter discusses various strategies for preventing alkali-silica reaction (ASR) in concrete,
including avoiding reactive aggregates, controlling the alkali content of the concrete, using
supplementary cementing materials, and the use of lithium-based compounds. The chapter does
not discuss methods for preventing alkali-carbonate reaction (ACR). Strategies for controlling
ASR are not generally effective for preventing ACR and, consequently, alkali-carbonate
reactive rocks must not be used in concrete. Alkali-carbonate reaction is discussed in chapter 9.
5.2 PREVENTIVE MEASURES OPTIONS

In chapter 2 it was shown that there are three basic requirements for damaging ASR to occur in
concrete; these are:
A sufficient quantity of reactive silica (within aggregates)
A sufficient concentration of alkali (primarily from portland cement)
Sufficient moisture
Elimination of any one of these requirements will prevent the occurrence of damaging alkalisilica reaction. Exclusion of water from civil engineering structures is not practical in most
cases and so, from a consideration of the fundamental aspects of ASR discussed in chapter 2,
the most obvious options for preventing expansion due to ASR are the following:
1. Avoid the use of reactive aggregates
2. Minimize the amount of alkalis from the portland cement
Two other options are as follows:
3. Use supplementary cementing materials (SCM)
4. Use lithium-based compounds
In fact, the use of SCM is a form of Option 2 as these materials consume a portion of the alkalis
contributed by the portland cement and reduce their availability for reaction with the aggregate.
Lithium compounds work in a different manner by changing the nature of the reaction product.
Both of these mechanisms are discussed later in the chapter.
75
AAR Facts Book
5.3 USE OF NON-REACTIVE AGGREGATES

Using non-reactive, or more correctly, non-deleteriously-reactive, aggregates is certainly a
viable method of preventing ASR-induced damage if such aggregates are available. Competent
and thorough testing is required to ensure that aggregate sources are non-reactive. This can be
achieved through a combination of petrographic examination, expansion testing of mortar
(ASTM C 1260) or concrete (ASTM C 1293), and field performance. Such testing should be
performed on a regular basis to ensure that the composition (and reactivity) does not change
within a pit or a quarry. If the aggregate sources can be confirmed to be truly non-reactive, no
further precaution is required to prevent ASR.
There are a number of reasons why the option of using non-reactive aggregates is not always
feasible; these include:
Non-reactive aggregates are not available locally and the cost of shipping non-reactive
materials from other locations is prohibitive

Reactive aggregates that are otherwise wholly suitable for concrete are readily and
abundantly available at lower cost and reduced environmental impact compared with
non-reactive materials
Lack of confidence in test results (or testing laboratories) or test results are ambiguous
(e.g., different test methods do not agree)

All of the locally-available materials fail the very aggressive accelerated test (regardless
of true reactivity)
In such cases, it is necessary to adopt one of the other options for preventing damaging ASR.
Furthermore, some instances warrant extra caution even when using aggregates believed to be
non-reactive; examples include the design of critical structures (e.g., prestigious structures or
those with an extended design life) and the construction of structural elements exposed to a very
aggressive environment (e.g., structures exposed to seawater or deicing salts, which may
provide an external source of alkalis).
5.4 LIMITING THE ALKALI CONTENT OF THE CONCRETE

Stanton's (1940) formative work on ASR indicated that expansive reaction is unlikely to occur
when the alkali content of the cement is below 0.60% Na2Oe. However, later research indicated
that damaging ASR could occur both in the laboratory and in the field (e.g., Woolf 1952; Stark
1980; Blaikie et al. 1996) when low-alkali cements are used. Despite this, the 0.60% value has
76
AAR Facts Book
become the accepted maximum limit for cement to be used with reactive aggregates in the
United States, and appears in ASTM C 150 Standard Specification for Portland cement as an
optional limit when concrete contains deleteriously reactive aggregate. Appendix X1 of ASTM
C 33 Standard Specification for Concrete Aggregates includes low-alkali cement (meeting the
ASTM C 150 limit of 0.60% Na2Oe) as a measure known to prevent excessive expansion.
Limiting the alkali level of the cement takes no account of the cement content of the concrete
which, together with the cement alkali content, governs the total alkali content of concrete, and
is considered to be a more accurate index of the risk of expansion when a reactive aggregate is
used in concrete. Figure 2.12 (in chapter 2) shows the relationship between alkali content and
expansion for concretes produced with a range of cement contents and cements of varying
alkali content (Lingrd et al. 2012). The figure was produced using previously unpublished data
from the Building Research Establishment (Blackwell, private communication). The
relationship clearly shows that it is the product of cement content and cement alkali level (i.e.,
the alkali content of the concrete) that controls the alkali content rather than the cement alkali
level alone.
Figure 5.1 shows expansion of concrete prisms plotted against the alkali content of the concrete
for three different reactive aggregates (selected to demonstrate the range of behavior observed).
It can be seen that the threshold alkali content required to initiate damaging expansion in the
concrete prism test varies considerably between aggregates, with values ranging from
approximately 3.0 kg/m3 (5 lb/yd3) Na2Oe to more than 5.0 kg/m3 (8.3 lb/yd3) Na2Oe for the
aggregates shown in Figure 5.1.
77
AAR Facts Book
Figure 5.1. Effect of Alkali Content on Expansion of Concrete Prisms (Stored over Water at
38C, 100F) with Different Aggregates (data from Thomas et al. 1996 and Figure 2.12)
A number of specifications have employed a maximum concrete alkali content as an option to
control expansion in concrete containing reactive aggregates. Nixon and Sims (1992) reported
that maximum permissible alkali contents between 2.5 and 4.5 kg/m3 (4.2 to 7.5 lb/yd3) Na2Oe
have been specified by various countries and agencies, with the allowable alkali content
sometimes varying depending on aggregate reactivity.
Unfortunately, traditional concrete prism tests tend to underestimate the threshold alkali content
for aggregates as a result of alkali being leached from the concrete during exposure. This
phenomenon has been discussed in literature (Thomas et al. 2006; Lindgrd et al. 2012) and a
further example is shown in Figure 5.2. The figure shows the expansion of concrete blocks
(0.38 x 0.38 x 0.71 m, 15 x 15 x 28 in.) containing reactive (Jobe) sand and various levels of
alkali, and stored on an outdoor exposure site at the University of Texas in Austin, compared
with the expansion of concrete prisms from the same mix stored over water (ASTM C 1293). It
is evident that the concrete prism test will yield a higher threshold alkali content than may be
observed for larger elements stored under field conditions.
78
AAR Facts Book
Figure 5.2. Effect of Alkali Content on the Expansion of Blocks Stored Outdoors and Prisms
Stored over Water (unpublished data from the University of Texas at Austin)
In 2000, CSA A23.2-27A introduced a sliding-scale for the alkali limit used as a preventive
measure, the value varying from 1.8 to 3.0 kg/m3 (3 to 5 lb/yd3) Na2Oe as follows:
Table 5.1. Alkali Limits Specified in CSA A23.2-27A
Level of prevention
required
Mild (W)
Moderate (X)
Strong (Y)
Exceptional (Z)
Alkali limit (Na2Oe)

kg/m3
lb/yd3
3.0
5.0
2.4
4.0
1.8
3.0
1.8 + SCM
3.0 + SCM
As the data available from laboratory tests at the time was considered unreliable for the purpose
of establishing threshold alkali values, the maximum alkali content limits were based on
published data from structures and field experience. Thomas (1996a) had previously reported
damaging ASR in concrete dams in the U.K. and Canada where the estimated alkali content was
in the range of 2.0 to 2.4 kg/m3 (3.3 to 4.0 lb/yd3) Na2Oe. Rogers et al. (2000) reported results
from a study of 8-year-old concrete blocks (0.6 x 0.6 x 2 m, 2 x 2 x 6.6 ft) stored on an
exposure site in Ontario, Canada. Specimens produced with high-alkali cement showed very
significant expansion and cracking after 8 years. Specimens with low-alkali cement (0.46%
79
AAR Facts Book
Na2Oe) and a calculated alkali content of 1.91 kg/m3 (3.18 lb/yd3) Na2Oe showed significantly
less expansion; however, the expansion did exceed 0.04% at 8 years and small cracks were
evident. Subsequent investigations (Hooton et al. 2006; MacDonald et al. 2012) of these blocks
at later ages (14 and 20 years) confirmed the presence of significant ASR-induced damage in
the blocks produced with low-alkali cement and measured expansion of approximately 0.08% at
20 years (see Figure 5.3). Based on the information available, a maximum alkali limit of 1.8 kg/
m3 (3.0 lb/yd3) Na2Oe was selected for Prevention Level Y (e.g., concrete with highly reactive
aggregate, exposed to moisture with a service life up to 75 years), and it was decided to require
the same limit plus incorporate minimum levels of supplementary cementing material (SCM)
for Level Z. This was supported by anecdotal evidence available to the committee preparing the
guidelines, there being no known case of ASR in concrete structures with lower alkali contents.
In the absence of any other data, the same alkali limits were adopted in AASHTO PP65-11; this
is discussed further in chapter 6.
Figure 5.3. Expansion of Concrete Prisms in the Laboratory and Concrete Blocks and Slabs on
the Kingston Exposure Site (data from MacDonald et al. 2012)
Aggregates that are used in concrete with an alkali content below the threshold for expansion
may cause damaging expansion if the alkali content of the concrete increases at some locations
during the service life of the structure. This may occur through alkali concentration caused by
drying gradients, alkali release from aggregates, or the ingress of alkalis from external sources,
such as deicing salts or seawater (Nixon et al. 1987). Stark (1978) reported increases in soluble
alkali from 1.1 to 3.6 kg/m3 Na2Oe close to the surface of some highway structures. Migration
of alkalis due to moisture, temperature, and electrical gradients has also been demonstrated by
80
AAR Facts Book
laboratory studies.
Alkali migration can occur very early in the life of concrete and has been implicated as the
cause of popouts at the surface of slabs (Cong et al. 2004; Landgren and Hadley 2002). Alkalis
may be carried upwards with the bleed water in freshly placed concrete, and this migration may
be exacerbated during and after setting by the evaporation of water from the surface of the slab
especially during hot, dry weather. The concentration of alkalis at the surface leads to ASR in
the near surface of the concrete where the formation of gel and expansion of aggregates can
lead to the fracture of a small conical-shaped mortar fragment (popout) overlying the reacting
aggregate. Oftentimes the damage is restricted to the surface of the concrete where the alkalis
are concentrated as there is insufficient alkali in the bulk concrete to cause damage. Some
studies have reported surface concentrations that are as much as six times the alkali
concentration in the bulk concrete (Nixon et al. 1979). This phenomenon can occur even when
low-alkali cement or pozzolans are used (Cong et al. 2004).
A number of workers have demonstrated that many aggregates contain alkalis that may be
leached out into the concrete pore solution, thereby increasing the risk of alkali-aggregate
reaction. Stark and Bhatty (1986) reported that, in extreme circumstances, some aggregates
release alkalis equivalent to 10% of the portland cement content. A comprehensive review on
alkali release from aggregates, including methods for determining the releasable or available
alkali in aggregates, has been presented by Berube et al. (2002).
Supplementary cementing materials (SCM), such as fly ash, silica fume, slag, and natural
pozzolans may also contain significant quantities of alkali. However, these alkalis generally do
not need to be included in the calculation of the concrete alkali content as SCMs tend to reduce
the alkalis that are available for reaction with the aggregate; this is discussed in the next section.
5.5 USE OF SUPPLEMENTARY CEMENTING MATERIALS

One of the most efficient means of controlling ASR in concrete containing reactive aggregates
is the appropriate use of supplementary cementing materials (SCM). Such materials include
pozzolans (e.g., fly ash, silica fume, calcined clay, or shale) and ground-granulated blast
furnace slag. The potential use of pozzolans to control ASR dates back as far as the discovery
of ASR, having been reported in the first major publication on the phenomenon (Stanton 1940).
In this paper, Stanton not only demonstrated that damaging reaction would only occur if there
was a sufficient quantity of alkalis in the portland cement and reactive silica in the aggregate,
but also that expansion was reduced when a pozzolanic cement was used. Ten years later,
Stanton (1950) further demonstrated that partially replacing portland cement with a sufficient
quantity of pozzolan (pumicite or calcined shale) eliminated deleterious expansion whereas
81
AAR Facts Book
replacement with similar quantities of ground quartz (Ottawa) sand did not, indicating that the
beneficial action of the pozzolan extended beyond merely diluting the cement alkalis. In the
early 1950s, various studies (Cox et al. 1950; Barona 1951; Buck et al. 1953) showed that other
supplementary cementing materials (SCM), namely fly ash and slag, were also effective in
reducing expansion.
Since these early studies there have been literally hundreds of studies and technical papers
dealing with the effects of SCM on ASR, and it is now generally recognized that the use of a
sufficient quantity of a suitable SCM is one of the more efficient preventive measures for
controlling expansion when a deleteriously reactive aggregate is used in concrete (Thomas et al.
2008). Thomas (2011) reviewed selected published works dealing with (i) the mechanisms by
which SCM controls ASR, (ii) the effect of SCM composition on its efficacy in this role, and
(iii) test methods for determining the amount of SCM required to minimize the risk of
damaging expansion to an acceptable level.
Thomas (2011) showed that almost any SCM can be used to control ASR provided it is used at
a sufficient level of replacement. The amount required varies widely depending on, among
other things, the following:
The nature of the SCM (especially mineralogical and chemical composition); more
SCM is required as its silica content decreases or as its alkali and calcium content
increase,
The nature of the reactive aggregate; generally, the more reactive the aggregate, the
higher the level of SCM required,

The availability of alkali within the concrete (i.e., from the portland cement and other
sources); the amount of SCM required increases with the amount of available alkali, and
The exposure conditions of the concrete; concrete exposed to external sources of alkali
may require higher levels of SCM.

Figure 5.4 shows the (conceptual) relationship between the (long-term) expansion of concrete
and the level of replacement for different SCMs. Generally, as the level of replacement
increases with a particular SCM, the expansion decreases and eventually reaches an acceptable
level at which no damage occurs. SCMs that are very high in reactive silica, such as silica fume
and metakaolin, tend to be very efficient in controlling expansion and are only required at
relatively low levels of replacement (e.g., 10 to 15%). On the other hand, SCMs with lesser
amounts of silica, such as Class C fly ash and slag, need to be used at higher levels of
replacement (e.g., 40%). Some SCMs (e.g., some Class C fly ashes) may produce a pessimum
effect by increasing the amount of expansion (compared to concrete without SCM) if they are
used at low levels of replacement, but decreasing expansion at higher levels of replacement.
82
AAR Facts Book
Level of replacement required to

control ASR expansion increases as:
SiO2 of SCM decreases
CaO of SCM increases
Alkali content of SCM
Alkali available in the concrete
increases
Reactivity of aggregate increases
Figure 5.4. Expansion versus SCM Content - Conceptual Relationship (Thomas 2011)
It is generally considered that the principal mechanism by which SCMs control ASR expansion
is by reaction with and consumption of the alkalis in the concrete pore solution which reduces
the alkali available for reaction with the aggregates. However, the use of SCMs also results in a
reduction in the availability of calcium (due to the dilution of and consumption of calcium
hydroxide) and in a refinement in the pore structure, which leads to reduced ionic and moisture
diffusivity; these effects may also be beneficial in terms of minimizing the risk and extent of
ASR (Thomas 2011).
This section begins with a discussion on the effect of SCMs on the composition of the pore
solution and the availability of alkalis, and then provides examples on the impact of SCMs on
the expansion of concrete. The role of SCM composition, cement alkalis, and aggregate
reactivity on the expansion of concrete containing SCMs is also discussed.
5.5.1 Effect of SCM on the Availability of Alkalis
Although all SCMs contain some level of alkali and some may contain significantly more alkali
than the portland cement that they partially replace, the main mechanism by which SCMs
reduce expansion due to ASR is by reducing the alkalis that are available to the concrete pore
solution. Once the alkalis in the binder phase (portland cement + SCM) of concrete are
released by hydration they may be present in one of three ways: dissolved within the pore
solution, bound by the hydration products, or incorporated in alkali-silica gel. In the absence of
83
AAR Facts Book
reactive aggregate, alkalis will not be consumed by ASR, and the partition of the alkalis
between the pore solution and the hydrates is largely a function of the composition of the
binder.
Numerous workers have shown that SCMs have a significant impact on the concentration of
alkalis in the pore solution (see Thomas 2011). Studies on the effect of fly ash and slag on the
pore solution of pastes have been reviewed by Thomas (1996b), and studies involving silica
fume have been reviewed by Thomas and Bleszynski (2001). These studies show that the
incorporation of most SCMs leads to a reduction in the concentration of alkali hydroxides in the
pore solution of pastes, mortar, and concretes, the amount of reduction increasing with higher
SCM replacement levels. Figure 5.5 shows the evolution of the hydroxyl ion concentration of
the pore solution extracted from sealed paste samples with w/cm = 0.50, and Figure 5.6 shows
the OH- concentration at 2 years as a function of the level of SCM (Shehata et al. 1999;
Ramlochan et al. 2000; Bleszynski 2002; Shehata and Thomas 2002). Silica fume is the most
efficacious SCM in this role, at least initially, followed by metakaolin, low-calcium fly ash, and
slag. High-calcium or high-alkali fly ashes are less effective and have to be used at relatively
high levels of replacement to produce a significant reduction in the pore solution alkalinity. It is
interesting to note that in the case of the paste with 10% silica fume the OH- concentration
drops rapidly over the first 28 days but then starts to increase slowly with time beyond 3
months; similar behavior was observed in pastes containing 5% silica fume (Shehata and
Thomas 2002). This behavior is not observed for pastes containing any of the other SCMs. As
shown in Figure 5.5, the long-term increase in the OH- concentration seems to be prevented in
pastes containing 5% silica fume by the addition of either slag (25%) or fly ash (15%). It is
conjectured that the presence of alumina in the SCM possibly contributes in some way to
prevent the long-term release of alkalis back in to the pore solution. Hong and Glasser (2002)
showed that introducing alumina into C-S-H, to form C-A-S-H, markedly increases its alkalibinding capacity, and they suggest that this partially explains the beneficial effects of aluminous
SCM with regards to reducing pore solution alkalinity and the potential for ASR.
84
AAR Facts Book
Figure 5.5. Evolution of the Pore Solution in Pastes Containing SCM (Shehata et al. 1999;
Ramlochan and Thomas 2000; Bleszynski 2002; Shehata and Thomas 2002)
Figure 5.6. Effect of Type and Amount of SCM on Pore Solution Composition (Shehata et al.
1999; Ramlochan et al. 2000; Bleszynski 2002; Shehata and Thomas 2002)
85
AAR Facts Book
Figure 5.7 shows an empirical relationship between the OH- concentration of the pore solution
extracted from 2-year-old sealed pastes with w/cm = 0.50 (Thomas and Shehata 2004; Thomas
and Folliard 2007) and a chemical index derived from the chemical composition of the
binder.
Figure 5.7. Relationship Between Pore Solution Composition and the Chemical
Composition of the Binder (Thomas 2011)
A total of 79 different binders were tested including the following:

100% portland cement at a range of different alkali contents (0.36 to 1.09% Na2Oe),
Binary mixes with 25 to 70% fly ash using 18 different fly ashes with a range of
chemical compositions (1.1 to 30.0% CaO, 1.4 to 9.7% Na2Oe),

Binary mixes with 25 to 50% slag, 5 to 10% silica fume and 10 to 20% metakaolin; each
SCM came from a single source, and

Ternary mixes containing silica fume blended with either slag or fly ash; both low-
calcium and high-calcium fly ash were used.

The relationship was derived empirically to find the chemical index that was most reliably
correlated to the OH- concentration based on a least-square fit. The best-fit index was found to
be the product of the equivalent alkalis and calcium divided by the square of the silica content
of the binder (Na2Oe x CaO)/(SiO2)2. Although this is an empirical relationship it makes sense
intuitively because the alkalinity of the pore solution can be expected to be a function of the
alkalis in the binder and the ability of the hydrates to bind alkalis, which has been shown to be a
86
AAR Facts Book
function of the calcium-to-silica ratio of the binder (see discussion below). The alumina content
of the binder was not found to be statistically significant based on the empirical analysis
conducted using this dataset, despite the apparent benefit of alumina discussed above.
Analyzing the composition of the pore solution extracted from a paste sample only provides one
point on the equilibrium curve between bound and free alkalis. If the alkali content of the pore
solution in concrete decreases, due perhaps to leaching or reaction with reactive silica in the
aggregate, a portion of the bound alkalis may be released to regain equilibrium. It is important
to establish what portion of the alkalis in a binder are available to a solution at a pH that is
just able to sustain the alkali-silica reaction, as it is these alkalis that are available to fuel the
reaction. Shehata and Thomas (2006) studied the alkali release characteristics of pastes
produced with high-alkali cement and combinations of silica fume and various fly ashes. Paste
samples, 1 to 3 years of age, were immersed in solutions of alkali hydroxide at initial molar
concentrations of 0, 0.10, 0.25, and 0.40 and the change in concentration was observed to
determine how much alkali was leached from the binder. Figure 5.8 shows a selection of the
data produced in this study. When mature paste samples were immersed in distilled water (pH =
7.0) almost all of the alkalis present in the binder (80 to 90%) were released regardless of
composition. As the alkali concentration of the leaching solution increased, the amount of alkali
released from the binder decreased and was strongly dependent on the composition of the
binder. Pastes containing 100% portland cement, 5% silica fume, or 25% high-CaO fly ash still
released a significant portion of the alkalis present in the binder (50 to 80%) even in the
solution of the highest initial alkali concentration (0.40M). Much less alkali (~ 20%) was
released from pastes containing 25% low-CaO fly ash or ternary blends containing silica fume
with either low-CaO or high-CaO fly ash. A correlation of the data (Shehata and Thomas 2006)
available for 24 different binders showed that the amount of alkali released to a solution with an
initial alkali hydroxide concentration of 0.25M (assumed to be the concentration necessary to
sustain alkali-silica reaction) was related to the chemical composition of the binder as
represented by the parameter (Na2Oe x CaO)/(SiO2).
87
AAR Facts Book
Figure 5.8. Available Alkalis in Pastes with SCM Stored in Solutions of Varying Reactivity
(Shehata and Thomas 2006)
88
AAR Facts Book
The ability of SCMs to reduce the pore solution alkalinity is linked to their effect on the
composition and alkali-binding capacity of the hydrates (especially C-S-H). Bhatty and
Greening (1978) found that C-S-H with a low Ca/Si ratio was able to retain more alkali (Na +
K) compared to hydrates of higher lime to silica ratios. The addition of fly ash reduces the Ca/
Si ratio of the C-S-H hydrates, and there is a concomitant increase in the alkali content.
Rayment (1982) observed significant differences in the C-S-H composition of portland cement
and fly ash pastes after just 8 days curing at 20C. However, Uchikawa et al. (1989) found little
difference in pastes after 91 days at 20C but substantial changes due to the incorporation of fly
ash after 60 days at 40C, indicating the role of the pozzolanic reaction in the CSH composition.
Thomas et al. (1991), reporting results for 7-year-old concretes containing reactive flint sand,
showed that the alkali binding capacity of C-S-H hydrates in concretes was increased
significantly by the addition of fly ash. Uchikawa et al. (1989) showed that slag has a similar
effect to low-calcium fly ash on hydrate composition. Glasser & Marr (1985) explain the
differences in alkali absorption on the basis of the surface charge on the C-S-H, which is
dependent on the Ca/Si ratio. At high ratios, the charge is positive and the C-S-H tends to repel
cations. As the Ca/Si ratio decreases the positive charge reduces, becoming negative at low Ca/
Si ratios, e.g., less than 1.3 (Glasser 1992). Negatively charged C-S-H has an increased capacity
to sorb cations, especially alkalis. Hong and Glasser (1999) confirmed the importance of the
Ca/Si ratio on the alkali-binding capacity of synthesized single-phase C-S-H but subsequently
showed that the binding capacity could be greatly increased by introducing alumina into the C
S-H to form C-A-S-H (Hong and Glasser 2002).
Many of the studies on the alkali-binding of C-S-H have involved microanalysis (e.g., using
scanning electron microscopy equipped with energy dispersive x-ray analysis) of the innerproduct C-S-H forming around remnant alite and belite grains. However, outer-product C-S-H
also forms by reaction between Ca(OH)2 and pozzolans, but this phase is more difficult to
identify and analyze separately than the inner-product C-S-H. The pozzolanic reaction is
actually very similar to the alkali-silica reaction. The reactive silica in the pozzolan reacts first
with the alkali-hydroxides in the pore solution, and alkali-silica gel containing small amounts of
calcium is formed. Over time, calcium exchanges for alkali in the gel, and C-S-H forms with a
relatively low Ca/Si ratio compared to that formed in portland cement paste. The only
substantial differences between this pozzolanic reaction and the alkali-silica reaction is the
timescale over which the reactions occur and the absence of any detectable expansion due to the
pozzolanic reaction. The lack of expansion can perhaps be explained by the fact that pozzolans
are very-finely divided materials and the alkali-silica gel that forms and is subsequently
converted to C-S-H is distributed throughout the cement paste, whereas the presence of reactive
aggregate particles leads to the accumulation of larger deposits of alkali-silica gel in discrete
locations that can become sites of expansion. The importance of the size and distribution of the
reactive silica can be demonstrated in two ways. Firstly, it has been known since the formative
work of Stanton (1940) that if a reactive aggregate is ground to sufficient fineness (sub-180
89
AAR Facts Book
microns in Stantons studies) expansion is eliminated. Figure 5.9 shows data from Thomas
(2011) relating to the effect of ground Vycor glass (sub-100 micron) on the expansion of mortar
bars containing sand-sized Vycor glass as a reactive aggregate. The sand-sized Vycor glass
behaves as a reactive aggregate causing expansion of the mortar with portland cement as the
only binder. However, the same material, when ground, behaves like a pozzolan when it is used
to replace 20% of the portland cement and prevents expansion. Secondly, if finely-divided
pozzolans agglomerate and form sand-sized particles, these particles will behave like reactive
aggregates and may result in expansion and cracking. This effect has been observed with
agglomerated silica fume both in the field and in the laboratory. Figure 5.10 (from Maas et al.
2007) shows a back-scattered electron image of a mortar bar containing agglomerated silica
fume after storage in 1M NaOH solution at 80C for 14 days (Maas et al. 2007). The mortar,
which contained non-reactive sand, expanded during test, and the expansion was attributed to
the reaction of the agglomerated silica fume.
Figure 5.9. Role of Particle Size on the Behavior of Vycor Glass (Thomas 2011)
90
AAR Facts Book
Figure 5.10. Agglomerated Silica Fume Particle Behaving as a Reactive Aggregate and
Source of ASR Expansion (Maas et al. 2007)
5.5.2 Effect of SCM on the Expansion of Concrete

Stanton proposed a test method for evaluating the potential for cement-aggregate combinations
to expand due to ASR in his first major paper on the subject (Stanton 1940). This involved the
manufacture of small mortar bars and storing them over water in sealed containers. Subsequent
modifications included elevating the temperature to 38C (100F), and the test was eventually
standardized to become ASTM C 227. Pozzolans were often evaluated using a modified version
of this test with either the job aggregate or a standard reactive aggregate, Pyrex glass; the test
with Pyrex became standardized as ASTM C 441. ASTM C 227 is no longer widely used as it
fails to detect many slowly-reacting aggregates because the small sample size and test
conditions promote the leaching of alkalis from the bars; such a drawback is also a problem for
evaluating SCMs (Thomas et al. 2006). ASTM C 441 is a much more rapid test (typically 14 to
56 days) because of the high reactivity of the Pyrex glass, and leaching is less significant during
the test. However, this test fails to account for the nature of the reactive aggregate which is
known to impact the amount of SCM required, and tends to overestimate the amount of SCM
required to control expansion with natural aggregates (Thomas et al. 2006). However, the use of
Pyrex glass does allow a comparative evaluation of pozzolans, and a number of investigators
have used the test to observe the reduced efficacy of high-calcium ash compared with lowcalcium ash (Dunstan 1981; Buck and Mather 1987; Klieger and Gebler 1987; Smith 1988;
91
AAR Facts Book
Carrasquillo and Snow 1987). Dunstan's (1981) work was the most comprehensive with regards
to fly ash composition. He reported results from Pyrex mortar bar tests for 17 ashes of varying
chemistry and showed a reliable correlation between the calcium content of the ash and the
expansion of mortar bars at 14 days.
The test methods most commonly used today to evaluate the efficacy of SCM in controlling
ASR expansion are the concrete prism test (ASTM C 1293) and the accelerated mortar bar test
(ASTM C 1567). The accelerated test is by far the most widely used test, however, it is only
intended as a screening test and should not be relied upon for phenomenological studies. The
test involves the immersion of small mortar bars in 1 M NaOH solution at 80C, and this tends
to mask the importance of the alkalis in the system under test. Since SCMs control ASR
expansion mainly by reducing the availability of alkalis, providing an inexhaustible supply of
alkalis is not desirable as eventually the beneficial effects of the SCM will be swamped by the
ingress of alkalis from the storage solution. Furthermore, the very high temperature used in this
test is not representative of the conditions that concrete encounters in the field. The review
presented here focuses on expansion tests involving concrete exposed either to field conditions
or to accelerated conditions (up to 38C) in the laboratory (e.g., ASTM C 1293 concrete prism
tests and similar tests). Data from ultra-accelerated tests, such as the accelerated mortar bar test
(ASTM C 1567), are not included.
5.5.2.1 Effect of SCM Composition on Expansion
As discussed above and illustrated in Figure 5.4, almost any SCM can be used to control
expansion due to alkali-silica reaction provided it is used in sufficient quantity. The amount of
SCM required is, of course, dependent on the composition of the SCM, but also on the
reactivity of the aggregate, the quantity of alkalis supplied by the portland cement (and other
sources), and whether the concrete will be exposed to alkalis (e.g., seawater, deicing chemicals)
during service. On one extreme, a highly efficient pozzolan with a high level of reactive silica
and negligible alkali content may be expected to eliminate damaging expansion with a
moderately reactive aggregate when used with a moderate-alkali cement at replacement levels
of about 10%; this scenario is represented by the left-hand curve in Figure 5.4. On the other
extreme, as represented by the right-hand curve in Figure 5.4, an SCM with a higher alkali and
lower silica content might need to be used at a replacement level of 50 to 60% or more with a
highly reactive aggregate and high-alkali cement.
Figure 5.11 shows the expansion of concretes at 2 years as a function of the type and amount of
SCM used; the tests were performed in a single laboratory using a reactive aggregate from a
single source (siliceous limestone from the Spratt quarry in Ontario, Canada) and were,
generally, performed in accordance with ASTM C 1293. Silica fume and metakaolin are the
most efficient with regards to reducing the expansion at 2 years, followed by low-calcium fly
92
AAR Facts Book
ash. Slag, high-calcium fly ash, and high-alkali fly ash were less efficient and had to be used at
significantly higher replacement levels to control expansion to below 0.040% at 2 years.
Generally, the effect of the different SCMs on the expansion of concrete prisms was consistent
with their effect on pore solution alkalinity.
Figure 5.11. Effect of SCMs on Two-Year Expansion of Concrete Containing Siliceous

Limestone (Shehata and Thomas 2002; Ramlochan et al. 2000; Thomas and Innis 1998)
Figure 5.12 (modified from Shehata and Thomas 2000) shows the 2-year expansion of concrete
containing 25% fly ash and a siliceous limestone (Spratt) aggregate; fly ashes from 29 different
sources are shown in this figure. Fly ashes with low to moderate alkali ( 4% Na2Oe) and
calcium contents ( 20% CaO) are generally effective in controlling expansion below 0.040%
at 2 years. As the calcium content increases above 20% CaO there is a marked increase in
expansion with increasing calcium content. Fly ashes with high alkali contents (> 5% Na2Oe)
are not effective in controlling expansion when used at a replacement level of 25% regardless of
the calcium content of the fly ash. High-calcium fly ashes may be effective in controlling ASR
expansion when used at increased levels of replacement. Figure 5.13 shows the expansion of
concrete prisms at 2 years plotted as a function of the level of fly ash replacement for Class F
fly ashes with less than 10% CaO and Class C fly ashes with more than 20% CaO. Whereas 20
to 25% low-CaO Class F fly ash was sufficient to control expansion, replacement levels of 50%
or more were required with the high-CaO Class C fly ashes.
93
AAR Facts Book
Figure 5.12. Effect of the Calcium Content of Fly Ash on the Two-Year Expansion of Concrete
Containing Siliceous Limestone (Shehata and Thomas 2002; Thomas 2011)
Figure 5.13. Effect of Fly Ash Replacement Level on Expansion of Concrete

94
AAR Facts Book
Ternary blends of portland cement with two SCMs are also effective in controlling ASR
expansion. Figure 5.14 shows the expansion of concrete containing 5% silica fume in
combination with fly ashes with different calcium contents. Combinations of 5% silica fume
with either 10 to 15% low-CaO fly ash or 20 to 30% high-CaO fly ash are effective in
controlling expansion ( 0.040% at 2 years). Similarly Figure 5.15 shows that combinations of
moderate amounts of silica fume (2 to 6%) are effective with moderate amounts of slag (15 to
35%).
Figure 5.14. Expansion of Concrete with Blends of Silica Fume and Fly Ash
95
AAR Facts Book
Figure 5.15. Expansion of Concrete with Blends of Silica Fume and Slag
(Bleszynski 2002; Bleszynski et al. 2002)
Figure 5.16 shows an empirical relationship between the expansion of concrete at 2 years and a
chemical index derived from the chemical composition of the cementing materials to produce
132 different concrete mixes which were tested in accordance with ASTM C 1293 (Thomas and
Shehata 2004). The cementing materials used to produce these concretes were the same as those
used for the pore solution study discussed above. The reactive coarse aggregate was siliceous
limestone (Spratt). The best fit between expansion and chemical composition was found to be
with the following index: [(Na2Oe)0.33 x CaO]/(SiO2)2. This relationship is not intended as a
method for predicting expansion based on the chemical composition of the binder phase, but
merely to examine what constituents of the binder tend to most influence ASR expansion. The
relationship is likely quite different if a different reactive aggregate or, even, a different test
method is used. However, the relationship does indicate that expansion is likely to increase as
the alkali and calcium content of the binder increase or as the silica content decreases, and this
is somewhat intuitive. It is interesting that the alkali content of the binder appears to play a less
important role in determining expansion compared with the pore solution composition, but this
is likely an artifact of the test conditions as significant leaching of alkalis occurs during the
concrete prism test, and this may reduce the apparent importance of the initial alkali content.
This effect can be observed when looking at the expansion data for the concrete mixes produced
with low-alkali cement. The expansion is lower than that expected based on the chemical
composition. However, it is known that the concrete prism test will likely underestimate the
96
AAR Facts Book
expansion with low-alkali cement because of leaching (Thomas et al. 2006). As with the
relationship with pore solution, the alumina content of the binder does not appear to
significantly affect the expansion of concrete. The role of alumina, however, is not yet well
understood, and further study is needed to determine its impact.
Figure 5.16. Effect of Binder Composition on the Expansion of Concrete
Containing Siliceous Limestone
5.5.2.2 Effect of Cement Alkalis

One of the drawbacks of the concrete prism test is that significant alkali leaching occurs during
the test, which means it cannot generally be used to determine the threshold alkali content
required to initiate expansion with a specific aggregate or to determine how the minimum
amount of SCM required varies as the alkali content of the cement changes (Thomas et al.
2006). Larger samples exposed under natural conditions should be relatively immune from the
effects of alkali leaching, however, much longer testing periods are required.
Figure 5.17 shows the expansion of 300-mm concrete cubes containing a reactive hornfels
aggregate and stored outdoors at the National Building Research Institute (NBRI) in South
Africa (Oberholster and Davies 1987; Oberholster 1989). Two series of mixes were cast with
cementitious contents of approximately 350 and 450 kg/m3. Within each series 5% or 10% of
the portland cement by mass was replaced with an equal volume of silica fume, resulting in
silica fume levels of 3.5% and 7.0% by mass. The active alkali content was maintained at a
constant level within a given series by addition of alkali hydroxide (using the same Na2O to
97
AAR Facts Book
K2O ratio as the cement). The active alkalis included the available cement alkalis (using
ASTM C 311) plus the alkali hydroxide, but excluded alkalis in the silica fume. The use of
3.5% or 7% silica fume delayed the onset of expansion and time to cracking in all cases.
However, only the mixture at the lower cement content with 7% silica fume failed to expand
after just over 7 years field exposure. These data clearly show the effect of alkali content on the
efficacy of silica fume in controlling expansion. A replacement level of 7% silica fume appears
to have been sufficient to provide long-term prevention of expansion when the active alkalis
of the mix were just less than 4 kg/m3 Na2Oe but not at the higher alkali content of 5 kg/m3
Na2Oe.
Figure 5.17. Effect of Cement Alkalis and Expansion of Concrete Containing Silica Fume
(Oberholster 1989)
Figure 5.18 shows data from Fournier and co-workers (2004) for concrete blocks stored
outdoors in Ottawa, Canada. The blocks contain a reactive greywacke coarse aggregate
(Springhill Quarry in New Brunswick) and 420 kg/m3 of cementing material. High-alkali
portland cement with 0.90% Na2Oe was used to manufacture the blocks and in some cases the
alkali content of the portland cement component of the concrete was boosted to 1.25% Na2Oe
by the addition of NaOH to the mix water as per ASTM C 1293. The data in Figure 5.18 show
that although the differences in expansion between boosted and unboosted blocks is not large,
increased amounts of SCM are clearly needed to control expansion to acceptable levels when
the alkali content of the mixture is increased.
98
AAR Facts Book
Figure 5.18. Effect of Cement Alkalis on Expansion of Concrete Containing Fly Ash
or Silica Fume (Fournier et al. 2004)
5.5.2.3 Effect of Aggregate Reactivity
Figure 5.19 shows 2-year expansion data for concrete prism tests containing various reactive
aggregates and different amounts of slag (Thomas and Innis 1998) or metakaolin (Ramlochan et
al. 2000). It can be seen that the amount of slag required to limit expansion below 0.040% at 2
years varies between 35% and 50% depending on the aggregate type. Figure 5.20 shows similar
data for blocks exposed on an outdoor exposure site in Ottawa (Fournier et al. 2004); the blocks
contain high-alkali cement (boosted to 1.25% Na2Oe), either low-CaO Class F fly ash or silica
fume, and aggregates of varying reactivity. For three of the reactive aggregates, which produced
expansion levels between 0.097% and 0.219% when tested with 100% portland cement, a fly
ash replacement level of 20% or a silica fume replacement level of 7.5% was sufficient to
reduce expansion below 0.040%. For the aggregates that produced an expansion of 0.338%
when tested with 100% portland cement, neither 30% fly ash nor 10% silica fume was quite
sufficient to reduce the expansion below 0.040%, although expansions were much reduced
(0.051% and 0.046% with 30% fly ash and 10% silica fume, respectively). For the aggregate
that produced the largest expansion when tested with portland cement (0.386%), significant
expansion (0.148%) still occurred with 30% fly ash. In this figure, it would appear that the
amount of fly ash required increases as the reactivity (as determined by the expansion when
tested with portland cement alone) of the aggregate increases.
99
AAR Facts Book
0.30
Expansion at 2 Years (%)
Spratt
0.25
Sudbury
Potsdam
0.20
Slag
Granite
Spratt
0.15
Sudbury
Metakaolin
0.10
0.05
0.00
0
20
40
60
Metakaolin or Slag Replacement Level (%)
Figure 5.19. Effect of Aggregate Type on the Amount of Metakaolin or Slag Required to
Control ASR Expansion (Thomas and Innis 1996; Ramlochan et al. 2000)
Expansion at latest measurement (%)
0.40
0.30
NM - FA
Con - FA
Su - FA
Su - SF
Al - FA
Al - SF
Sl - FA
Sl - SF
0.20
0.10
0.00
0
10
20
30
Silica Fume (SF) or Fly Ash (FA) Replacement Level (%)
Figure 5.20. Effect of Aggregate Type on the Amount of Silica Fume (SF) or Fly Ash (FA)
Required to Control ASR Expansion (Fournier et al. 2004). NM, Con, Su, Al and Sl are
different aggregates. FA and SF are fly ash and silica fume, respectively.
100
AAR Facts Book
5.5.3 Summary on Effect of SCM on ASR

Supplementary cementing materials are an effective means for controlling expansion due to
alkali-silica reaction and most, if not all, SCMs can be used in this role provided they are used
at a high enough level of replacement. The level of SCM required generally increases with the
following parameters:
The alkali available from the portland cement increases (If significant alkalis are available
from the aggregates e.g., feldspars, greywackes or from external sources, this will likely
also increase the level of SCM required.)
The alkali from the SCM increases
The CaO/SiO2 of the SCM increases
The reactivity of the aggregate increases
The amount of SCM required to prevent damaging ASR expansion generally falls in the ranges
below (modified from Thomas and Folliard 2007):
Table 5.2. Required Levels of SCM
Type of SCM
Level required (%)
Low-calcium fly ash (< 8% CaO)
20 to 30
Moderate-calcium fly ash (8 - 20% CaO)
25 to 35
High-calcium fly ash (> 20% CaO)
40 to 60
Silica fume
8 to 15
Slag
35 to 65
Metakaolin (calcined kaolin clay)
10 to 20
However, the level of SCM required may exceed these values under exceptional conditions
(e.g., extremely reactive aggregate, high alkali availability in concrete including alkali
contribution from aggregates, concrete exposed to high concentrations of alkali in service,
critical structure with extended service life).
101
AAR Facts Book
5.6 USE OF LITHIUM

The ability of lithium to control deleterious expansion due to alkali-silica reaction (ASR) in
mortar and concrete was first demonstrated by McCoy and Caldwell (1951). They showed that,
out of more than 100 chemical compounds tested, various salts of lithium (e.g., LiCl, Li2CO3,
LiF, Li2SiO3, LiNO3, and Li2SO4) were the most promising and could virtually eliminate the
expansion of mortar containing Pyrex glass provided they were used at sufficient levels of
replacement. Since then, there have been numerous studies which corroborate this earlier
discovery (Feng et al. 2005).
It is somewhat paradoxical that lithium compounds are effective suppressants of ASR as lithium
is an alkali metal like sodium and potassium. The precise mechanism by which lithium controls
ASR is not known, although many theories have been put forward (Feng et al. 2005). The
simplest and most commonly used explanation is that lithium salts will react with reactive silica
in a similar way to sodium and potassium salts, but the reaction product is an insoluble lithiumsilicate with little propensity to imbibe water and swell. The lithium silicate forms around
reactive aggregate particles and protects the underlying reactive silica from attack by alkali
hydroxides.
The initial work of McCoy and Caldwell (1951) showed that the amount of lithium required to
control expansion was a function of the availability of other alkalis (Na + K) in the system, and
they concluded that the expansion of mortar bars containing reactive Pyrex glass could be
effectively suppressed provided that the lithium-to-sodium-plus-potassium molar ratio was
greater than 0.74, i.e., [Li]/[Na+K] > 0.74. Since then numerous workers have demonstrated a
similar relationship between the amount of lithium required and the amount of alkali available,
but the minimum value of [Li]/[Na+K] has been shown to vary depending on a number of
issues such as the form of lithium, nature of reactive aggregate, and, perhaps, the method of test
used (Feng et al. 2005).
Although most lithium compounds have a beneficial effect, lithium nitrate (LiNO3) is
considered to be the most efficient form for suppressing ASR (Stokes et al. 1997). Lithium
nitrate solution is commercially available from a number of companies in North America, being
marketed as an ASR-suppressing admixture. Currently the product is sold as a 30% solution
of LiNO3. To achieve a lithium-to-sodium-plus-potassium molar ratio of [Li]/[Na+K] = 0.74
using a 30% solution of LiNO3 requires a dose of 4.6 gallons of LiNO3 solution per 1 kg of
Na2Oe (0.55 gallons of solution per 1 lb Na2Oe). This has been referred to as the standard
dose of lithium nitrate solution.
102
AAR Facts Book
Recent research (Tremblay et al. 2007) has highlighted the influence of aggregate type on the
amount of lithium required to suppress expansion due to ASR. Figure 5.21 shows the 2-year
expansion of concrete prisms with 12 different reactive aggregates and 1 non-reactive aggregate
(NF), and various levels of lithium (standard dose is [Li]/[Na+K] = 0.74). For 6 of the 12
aggregates 75% to 100% of the standard dose was sufficient to control expansion ( 0.040% at
2 years). For 3 of the aggregates 125% to 150% of the standard dose was required; however, for
the remaining 3 aggregates expansion could not be controlled even at 150% of the standard
dose.
Figure 5.21. Effect of Lithium Dose on the Expansion of Concrete with Different Reactive
Aggregates (from Tremblay et al. 2007)
As the effectiveness of lithium appears to be extremely aggregate dependent, it is not possible

to prescribe a single dose for controlling ASR, and the minimum dose must be determined by
testing lithium with the specific reactive aggregate being considered for use. At this point in
time there is no consensus regarding the appropriateness of accelerated tests for determining the
correct lithium dose, and it is recommended that the concrete prism test is used for this purpose.
103
AAR Facts Book
REFERENCES
ASTM. 2012. Standard Specification for Portland Cement. ASTM C 150, ASTM
PA, 6 p.
ASTM. 2013. Standard Test Methods for Sampling and Testing Fly Ash or Natural Pozzolans
for Use in Portland-Cement Concrete. ASTM C 311, ASTM International, West
Conshohocken, PA, 10 p.
Barona de la O, F. 1951. "Alkali-aggregate expansion corrected with Portland-slag cement."
Journal of the American Concrete Institute, 22(7): 545-552.
Berube, M-A., Duchesne, J., Dorion, J.F. and Rivest, M. 2002. Laboratory assessment of alkali
contribution by aggregates to concrete and application to concrete structures affected by alkali
silica reactivity. Cement and Concrete Research, 32: 1215-1227.
Bhatty, M.S.Y. and Greening, N.R. 1978. Interaction of alkalis with hydrating and hydrated
calcium silicates. Proceedings of the Fourth International Conference on the Effects of Alkalis
in Cement and Concrete, Purdue, 87-112.
104
AAR Facts Book
Blaikie, N.K., Bowling, A.J. and Carse, A. 1996. The assessment & management of alkalisilica reaction in the Gordon River Power Development intake tower. Alkali-Aggregate
Reaction in Concrete (Ed. A. Shayan), Proceedings of the 10th International Conference on
Alkali-Aggregate Reaction, Melbourne, 500- 507.
Bleszynski, R.F. 2002. The performance and durability of concrete with ternary blends of
silica fume and blast-furnace slag. PhD Thesis, University of Toronto.
Bleszynski, R., Hooton, R.D., Thomas, M.D.A. and Rogers, C.A. 2002. Durability of Ternary
Blend Concrete with Silica Fume and Blast-Furnace Slag: Laboratory and Outdoor Exposure
Site Studies. ACI Materials Journal, 99(5): 499-508.
Buck, A.D., Houston, B.J. and Pepper, L. 1953. "Effectiveness of mineral admixtures in
preventing excessive expansion of concrete due to alkali-aggregate reaction." Journal of the
American Concrete Institute, 30(10): 1160.
Buck, A.D. and Mather, K. 1987. "Methods for controlling effects of alkali-silica reaction in
concrete." U.S. Army Corps of Engineers, Waterways Experiment Station, Technical Report
No. SL-87-6, U.S. Army Engineer Waterways Experiment Station, Vicksburg.
Carrasquillo, R.L. and Snow, P.G. 1987. "Effect of fly ash on alkali-aggregate reaction in
concrete." ACI Materials Journal, 84(4): 299-305.
Cong, D.X., Lawrence, B.L., Deno, D.W. and Patty, T.S. 2004. ASR-induced surface defects.
Proceedings of the 12th International Conference on alkali-aggregate reaction in concrete,
October 15-19, 2004, Beijing, China, Vol. 2. 1142-1147.
Cox, H.P., Coleman, R.B. and White, L. 1950. "Effect of blastfurnace-slag cement on alkaliaggregate reaction in concrete." Pit and Quarry, 45(5): 95-96.
Dunstan, E.R. 1981. "The effect of fly ash on concrete alkali-aggregate reaction." Cement,
Concrete and Aggregates, 3(2): 101-104.
Farny, J.A. and Tarr, S.M. 2008. Concrete Floors on Ground. PCA EB075, Portland Cement
Association, Skokie, IL, 239 p.
105
AAR Facts Book
Feng, X., Thomas, M.D.A., Bremner, T.W., Balcom, B.J. and Folliard, K.J. 2005. Studies on
lithium salts to mitigate ASR-induced expansion in new concrete: a critical review. Cement
and Concrete Research, 35: 1789-1796.
Fournier, B., Nkinamubanzi, P-C. and Chevrier, R. 2004. Comparative Field and Laboratory
Investigations on the Use of Supplementary Cementing Materials to Control Alkali-Silica
Reaction in Concrete, Proceedings of the 12th International Conference on Alkali-Aggregate
Reaction in Concrete, Vol. 1, (Eds. T. Mingshu and D. Min), International Academic
Publishers/World Publishing Corp., Beijing, China, 528-537.
Glasser, F.P. 1992. "Chemistry of the alkali-aggregate reaction." The Alkali-Silica Reaction in
Concrete (Ed. R.N. Swamy), Blackie, London, 96-121.
Glasser, F.P. and Marr, J. 1985. The alkali binding potential of OPC and blended cements. Il
Cemento, 82: 85-94.
Hong, S-Y and Glasser, F.P. 1999. Alkali binding in cement pastes Part I. The C-S-H phase.
Hong, S-Y and Glasser, F.P. 2002. Alkali sorption by C-S-H and C-A-S-H gels Part II. Role of
alumina. Cement and Concrete Research, 32: 1101-1111.
Hooton, R.D., Rogers, C.A. and Ramlochan, T. 2006 The Kingston Outdoor Exposure Site for
ASR - After 14 Years What Have We Learned? Proceedings, Marc-Andr Brub
Symposium, Seventh CANMET/ACI International Conference on Durability, Montreal, May
31-June 2, 2006, 22 p.
Klieger, P. and Gebler, S. 1987. "Fly ash and concrete durability." Concrete Durability,
Katherine and Bryant Mather International Conference, (Ed. J.M. Scanlon), ACI SP-100, Vol.
1, American Concrete Institute, Detroit, 1043-1069. Discussion, ACI Materials Journal, 85(3),
1988, 214-215.
Landgren, R. and Hadley, D.W. 2002. Surface Popouts Caused by Alkali-Aggregate
Reactions. Research and Development Bulletin RD121, Portland Cement Association, Skokie,
IL, 13 p.
Lindgrd, J., Andic-Cakir, O., Fernandes, I., Ronning, T.F. and Thomas, M.D.A. 2012. Alkali
silica reactions (ASR): Literature review on parameters influencing laboratory performance
testing. Cement and Concrete Research, 40: 223-243.
Maas, A.J., Ideker, J.H. and Juenger, M.C.G. 2007. Alkali silica reactivity of agglomerated
silica fume. Cement and Concrete Research, 37: 166174.
106
AAR Facts Book
MacDonald, C.A., Rogers, C. and Hooton, R.D. 2012. The relationship between laboratory
and field expansion observations at the Kingston outdoor exposure site for ASR after twenty
years. Proceedings of the 14th International Conference on Alkali-Aggregate Reaction in
Concrete, Austin, TX, USA, May.
McCoy, W.J. and Caldwell, A.G. 1951. New approach to inhibiting alkali-aggregate
expansion. Journal of the American Concrete Institute, 22(9): 693-706.
Nixon, P.J. and Sims, I. 1992. Alkali Aggregate Reaction - Accelerated Tests Interim Report
and Summary of National Specifications. Proceedings of the 9th International Conference on
Alkali-Aggregate Reaction in Concrete, Vol. 2, The Concrete Society, Slough, 731-738.
Nixon, P.J., Collins, R.J. and Rayment, P.L. 1979. The Concentration of Alkalies by Moisture
Migration in Concrete - A Factor Influencing Alkali Aggregate Reaction. Cement and
Concrete Research. 9: 417-423.
Nixon, P.J., Canham, I. and Page, C.L. 1987. Aspects of the Pore Solution Chemistry of
Blended Cements Related to the Control of Alkali-Silica Reaction. Cement and Concrete
Research, 17(5): 839-844.
Oberholster, R.E. and Davies, G. 1987. "The effect of mineral admixtures on the alkaliaggregate expansion of concrete under outdoor exposure conditions." Proceedings of the 7th
International Conference on Concrete Alkali-Aggregate Reactions, (Ed. P.E. Grattan-Bellew),
Noyes Publications, New Jersey, 60-65.
Oberholster, R.E. 1989. "Alkali-aggregate reaction in South Africa. Some recent developments
in research." Proceedings of the 8th International Conference on Alkali-Aggregate Reaction,
(Eds. K.Okada, S. Nishibayashi and M. Kawamura), Kyoto, 77-82.
Ramlochan, T., Thomas, M.D.A. and Gruber, K.A. 2000. The effect of metakaolin on alkalisilica reaction in concrete. Cement and Concrete Research, 30(3): 339-344.
Rayment, P.L. 1982. "The effect of pulverized-fuel ash on the C/S molar ratio and alkali
content of calcium silicate hydrates in cement." Cement and Concrete Research, 12(2): 133
140.
Rogers, C., Lane, B. and Hooton, D. 2000, Outdoor exposure for validating the effectiveness
of preventative measures for Alkali-Silica Reaction. Proceedings of the 11th International
Conference on Alkali-Aggregate Reaction, (Eds. M-A. Brub, B. Fournier and B. Durand),
Quebec, June, 743-752.
Shehata, M., Thomas, M.D.A. and Bleszynski, R.F. 1999. The effect of fly composition on the
chemistry of pore solution. Cement and Concrete Research, 29(12): 1915-1920.
107
AAR Facts Book
Shehata, M.H. and Thomas, M.D.A. 2000. The effect of fly ash composition on the expansion
of concrete due to alkali-silica reaction. Cement and Concrete Research, 30: 1063-1072.
Shehata, M.H. and Thomas, M.D.A. 2002. Use of ternary blends containing silica fume and fly
ash to suppress expansion due to alkali-silica reaction in concrete. Cement and Concrete
Research, 32(3): 341-349.
Shehata, M.H. and Thomas, M.D.A. 2006. Alkali release characteristics of blended cements.
Smith, R.L. 1988. "Is the available alkali test a good durability predictor for fly ash concrete
incorporating reactive aggregate?" Fly Ash & Coal Conversion By-Products: Utilization &
Disposal IV, MRS Symposia Proceedings, Vol. 113, Materials Research Society, Pittsburg, 249
-256.
Stanton, T.E. 1950. "Studies of use of pozzolans for counteracting excessive concrete expansion
resulting from reaction between aggregates and the alkalies in cement." Pozzolanic Materials in
Mortars and Concretes, ASTM STP 99, American Society for Testing and Materials,
Philadelphia, 178-203.
Stark, D. 1978. Alkali-silica Reactivity in the Rocky Mountain Region. Proceedings of the
4th International Conference on Effects of Alkalis in Cement and Concrete, CE-MAT-1-78,
Purdue University, W. Lafayette, Indiana, 235-243.
Stark, D. 1980. Alkali-Silica Reactivity: Some Reconsiderations. Cement, Concrete, and
Stark, D. and Bhatty, M.S.Y. 1986. Alkali-silica Reactivity: Effect of Alkali in Aggregate on
Expansion. Alkalis in Concrete, ASTM STP 930, American Society for Testing and Materials,
Philadelphia, 16-30.
Stokes, D.B., Wang, H.H. and Diamond, S. 1997. A lithium-based admixture for ASR control
that does not increase the pore solution pH. Proceedings of the 5th CANMET/ACI Int. Conf.
on Superplasticizers and Other Chemical Admixtures in Concrete, (Ed. V.M. Malhotra), ACI
SP-173, American Concrete Institute, Detroit, 855-867.
Thomas, M.D.A. 1996a. "Field studies of fly ash concrete structures containing reactive
aggregates." Magazine of Concrete Research, 48(177): 265-279.
108
AAR Facts Book
Thomas, M.D.A. 1996b. "Review of the effect of fly ash and slag on alkali-aggregate reaction
in concrete." Building Research Establishment Report, BR314, Construction Research
Communications, Ltd, Watford, U.K., 1996.
Thomas, M.D.A., Blackwell, B.Q. and Nixon, P.J. 1996. Estimating the alkali contribution
from fly ash to expansion due to alkali-aggregate reaction in concrete. Magazine of Concrete
Research, 48(177): 251-264.
Thomas, M.D.A. 2011. The effect of supplementary cementing materials on alkali-silica
reaction: A review. Cement and Concrete Research, 41: 1224-1231.
Thomas, M.D.A. and Bleszynski, R.F. 2001. The use of silica fume to control expansion due
to alkali-aggregate reactivity in concrete a review. Materials Science of Concrete VI, (Ed. S.
Mindess and J. Skalny), American Ceramics Society, Westerville, OH, 377-434.
Thomas, M.D.A., Fournier, B., Folliard, K.J., Shehata, M., Ideker, J. and Rogers, C.A. 2007.
Performance limits for evaluating supplementary cementing materials using the accelerated
mortar bar test. ACI Materials Journal, 104(2): 115-122.
Thomas, M.D.A. and Folliard, K.J. 2007. Concrete aggregates and the durability of concrete.
Durability of concrete and cement composites, (Ed. C.L. Page and M.M. Page), Woodhead,
Cambridge, U.K. 247-281.
Thomas, M.D.A. and Innis, F.A. 1998. Effect of slag on expansion due to alkali-aggregate
reaction in concrete. ACI Materials Journal, 95(6): 716-724.
Thomas, M.D.A. and Shehata, M. 2004. Use of blended cements to control expansion of
concrete due to alkali-silica reaction. Proceedings of the 8th CANMET/ACI Int. Conf. on Fly
Ash, Silica Fume, Slag and Natural Pozzolans in Concrete, Supplementary Papers, Las Vegas,
591-607.
Thomas, M.D.A., Nixon, P.J. and Pettifer, K. 1991. "The effect of pulverized fuel ash with a
high total alkali content on alkali silica reaction in concrete containing natural U.K. aggregate."
Proceedings of the 2nd CANMET/ACI International Conference on Durability of Concrete,
(Ed. V.M. Malhotra), Vol. 2, American Concrete Institute, Detroit, 919-940.
Tremblay, C., Berube, M-A., Fournier, B., Thomas, M.D.A. and Folliard, K.F. 2007.
Effectiveness of lithium-based products in concrete made with Canadian reactive aggregates.
ACI Materials Journal, 104(2): 195-205.
109
AAR Facts Book
Uchikawa, H., Uchida, S. and Hanehara, S. 1989. "Relationship between structure and
penetrability of Na ion in hardened blended cement paste mortar and concrete." Proceedings of
the 8th International Conference on Alkali-Aggregate Reaction, (Ed. K.Okada, S. Nishibayashi
and M. Kawamura), Kyoto, 121-128.
Woolf, D.O. 1952. Reaction of aggregate with low-alkali cement. Public Roads, August, 50
56.
110
AAR Facts Book
Chapter 6 AAR Specifications
6 Alkali-Aggregate Reactions: Specifications
6.1 INTRODUCTION
This chapter discusses specifications for alkali-aggregate reaction focusing on AASHTO PP65
11 Standard Practice for Determining the Reactivity of Concrete Aggregates and Selecting
Appropriate Measures for Preventing Deleterious Expansion in New Concrete Construction.
Requirements within ASTM and ACI are also discussed.
Essentially there are two approaches to specifying preventive measures for controlling alkaliaggregate reaction. In the prescriptive approach, reactive aggregates are permitted provided
they are used with prescribed amounts of ameliorative materials such as a minimum level of fly
ash or slag (for example). In a performance-based approach, a certain combination of materials
(reactive aggregates and preventive measures) are permitted provided they are tested and shown
to meet certain performance requirements such as meeting the expansion limits of mortar-bar or
concrete-prism expansion tests. AASHTO PP65-11 has both prescriptive and performance
options.
6.2 EXISTING SPECIFICATIONS IN USA

6.2.1 ASTM
ASTM C 33 Standard Specification for Concrete Aggregates states that concrete aggregates
for use in concrete shall not contain any materials that are deleteriously reactive with the
alkalies in the cement in an amount sufficient to cause excessive expansion of mortar or
concrete unless one of the following three conditions is met:
the aggregate is used with a cement containing less than 0.60% alkalies calculated as
sodium oxide equivalent (Na2Oe = Na2O + 0.658 x K2O),
there is a satisfactory service record evaluation, or
the aggregate is used with the addition of a material that has been shown to prevent
harmful expansion due to the alkali-aggregate reaction.
Within the ASTM concrete standards there are number of laboratory test methods that can be
used to determine the potential for aggregates to be deleteriously reactive or the effectiveness of
different preventive measures against ASR; these are:
111
AAR Facts Book
C 295 Petrographic Examination of Aggregates

C 227 Standard Test Method for Potential Alkali Reactivity of Cement-Aggregate
C 289 Standard Test Method for Potential Alkali-Silica Reactivity of Aggregates
(Chemical Method)
C 441 Standard Test Method for Effectiveness of Pozzolans or Ground Blast-Furnace
Slag in Preventing Excessive Expansion of Concrete Due to the Alkali-Silica Reaction
C 1105 Standard Test Method for Length Change of Concrete Due to Alkali-Carbonate
Rock Reaction
C 1260 Standard Test Method for Potential Alkali Reactivity of Aggregates (Mortar-Bar
Method)
C 1293 Standard Test Method for Determination of Length Change of Concrete Due to
Alkali-Silica Reaction
C 1567 Standard Test Method for Determining the Potential Alkali-Silica Reactivity of
Combinations of Cementitious Materials and Aggregate (Accelerated Mortar-Bar
Method.
Appendix X1 of ASTM C 33 provides some limited guidance on interpretation of the results
from these tests and measures for mitigating alkali-silica reaction5. If the aggregate produces
little or no expansion in C 1260 or C 1293, or has a satisfactory service history (with similar
cementitious materials) no mitigation is necessary. On the other hand, if the aggregates are
considered to be deleteriously alkali-silica reactive Appendix X1 (section 4.3) recommends one
of the following preventive measures:
Use of ASTM C 150 portland cement meeting the low-alkali option ( 0.60% Na2Oe)
Use of ASTM C 595 blended cement meeting the optional mortar-bar-expansion
requirement
Use of hydraulic cement meeting the ASTM C 1157 performance specification
including Option R - Low Reactivity with Alkali-Reactive Aggregates.
Use of pozzolans or slag meeting the optional requirements of the relevant material
specifications (C 618 for fly ash and natural pozzolans, C 1240 for silica fume, and C
989 for slag) for preventing excessive expansion due to ASR.
5
ASTM C 33 Appendix X1 also provides guidance on mitigation of alkali-carbonate reaction. The guidance is
limited to avoiding reactive carbonate rocks; selective quarrying; diluting reactive rock to less than 20 % of the
aggregate in the concrete; use of smaller maximum size; and the use of very low alkali cement.
112
AAR Facts Book
The optional requirements for blended cements, C 1157 cements, pozzolans, and slag all make
use of C 441 (Pyrex) mortar-bar method for demonstrating the effectiveness in controlling
expansion due to ASR. However, each material specification has different performance
requirements and these may be summarized as follows:
ASTM Specification
Expansion Limit of Mortars with Pyrex Glass
C 595 Blended cements
Maximum expansion of 0.020% at 14 days and 0.060%

at 56 days
C 1157 Hydraulic cements

(performance-based specification)
Maximum expansion of 0.020% at 14 days
C 618 Fly ash and natural pozzolans
Expansion of fly ash mortars not greater than expansion

of control mortars with low-alkali cement ( 0.60%
Na2Oe) at 14 days
C 989 Slag
Expansion of the job cement plus slag should not expand

by more than 0.02% at 14 days or, if the job cement is
not known, the slag should reduce the 14-day expansion
of a mixture with high-alkali cement by at least 75%
when compared with a mix with high-alkali cement on
its own.
C 1240 Silica fume
Blend of high-alkali cement plus silica fume must reduce

expansion by at least 80% compared with high-alkali
cement alone
A joint C09/C01 Task Group recently recognized that numerous ASTM specifications provide
requirements and guidance for avoiding deleterious ASR expansion for individual concrete
materials (e.g., cementitious materials, aggregates, and SCMs) while, together, these
specifications do not provide coherent guidance for preventing deleterious expansion in
concrete; also, the specifications utilize standard test methods that are in many cases considered
to be unreliable (Struble 2010). The Task Group thus recommended that clear and consistent
guidance/specification for the prevention of ASR be developed that addresses performance at
the concrete level and includes requirements for aggregates. A joint ASTM technical
subcommittee, C01/C09-50, was then formed to develop new global requirements for the Risk
Management of Alkali-Aggregate Reactions. This new subcommittee started its activities in
2010.
6.2.2 ACI
ACI 301 Specifications for Structural Concrete (ACI 301-10) states that potentially reactive
aggregates may be used either with low-alkali cement ( 0.60% Na2Oe) or supplementary
113
AAR Facts Book
cementing materials in an amount shown to be effective in preventing harmful expansion due

to alkali-aggregate reaction in accordance with ASTM C441 and defers to ASTM C 33.
ACI 318 Building Code Requirements for Structural Concrete (ACI 318-08) requires
aggregates to meet the requirements of ASTM C 33 but makes no reference to alkali-aggregate
reactions (ASR or ACR) or the use of deleteriously-reactive aggregates in concrete.
Guidance for preventing damage due to AAR is provided in ACI 201 Guide to Durable
Concrete (ACI 201.2R-08) and ACI 221 Report on Alkali-Aggregate Reactivity (ACI 221.1R
98).
6.2.3 State Specifications
Many state transportation agencies have developed specifications for minimizing the risk of
deleterious alkali-aggregate reactions in concrete. These are too numerous to discuss here but
examples of these specifications are available for review at the following link:
http://www.fhwa.dot.gov/pavement/concrete/asr/reference.cfm?main_category=Specifications.
6.3 AASHTO PP65-11

A version of AASHTO PP65-11 Standard Practice for Determining the Reactivity of Concrete
Aggregates and Selecting Appropriate Measures for Preventing Deleterious Expansion in New
Concrete Construction was first published in 2010. As the title suggests, these guidelines can be
broken down into two steps as follows:
Evaluating aggregate reactivity (for alkali-silica and alkali-carbonate reactive aggregates)
Selecting preventive measures (for alkali-silica reactive aggregates only)
The practice is based on the 2004 version of the practice developed in Canada: CSA A23.2-27A
Standard practice to identify degree of alkali-reactivity of aggregates and to identify measures
to avoid deleterious expansion in concrete. Further information on the development of this
approach can be found in Report No. FHWA-HIF-09-001 (Thomas et al. 2008). The rationale
for preventive measures recommended in the prescriptive approach is also documented in a
report by the Federal Highway Administration (Thomas et al. 2012).
6.3.1 Evaluating Aggregate Reactivity
Aggregate reactivity is evaluated by considering one or more of the following options: (i) field
performance history, (ii) petrographic assessment, (iii) chemical composition (for quarried
114
AAR Facts Book
carbonates), (iv) data from accelerated mortar bar tests (AASHTO T 303), and (v) data from
concrete prism tests (ASTM C 1293). A flow chart showing the sequence of testing is presented
in Figure 6.1. Aggregates may be classified as non-deleteriously-reactive and can be used
without preventive measures provided they meet one of the following conditions:
There is an established history of satisfactory field performance (minimum 15 years)
with the aggregate used in the same exposure conditions with similar cementitious
materials.
Petrographic examination of the aggregate by a skilled and experienced petrographer
indicates that potentially deleterious minerals are not present in sufficient concentration
to cause deleterious reaction and expansion.
Expansion of mortar bars (AASHTO T 303; ASTM C 1260) is not greater than 0.10%
after 14 days immersion in 1 M NaOH at 176F (80C).
Expansion of concrete prisms (ASTM C 1293) is not greater than 0.040% at 1 year.
Although AASHTO PP65-11 allows aggregates to be accepted solely based on field history
and/or petrography it strongly recommends expansion testing using either the accelerated
mortar bar test or, preferably, the concrete prism test. If, after testing, the aggregate is identified
to be alkali-silica reactive, AASHTO PP65-11 requires that it is either rejected for use or used
together with appropriate preventive measures.
Prior to expansion testing, quarried carbonates must be subjected to chemical analysis to
determine whether there is a risk of alkali-carbonate reaction (ACR). The composition of the
aggregate is plotted on a graph of CaO/MgO vs. Al2O3 (see Figure 6.2) and if the results of the
analysis indicate a potential for ACR the aggregate must be tested by the concrete prism test as
the accelerated mortar bar test is not capable of detecting alkali-carbonate reactive aggregate.
Two options are allowed for the concrete prism test. One is to follow ASTM 1105 using a total
alkali of no more than 1.8 kg/m3 (3 lb/yd3) Na2Oe. If the expansion of the concrete prisms is
equal to or greater than 0.025 % at 6 months or 0.030 % at 1 year, the aggregate shall be
considered to be alkali-carbonate reactive and shall not be used in concrete. If the expansion of
the concrete is below these limits, it is not considered to the alkali-carbonate reactive and can be
evaluated in the same manner as non-carbonate rocks for alkali-silica reaction. The second
option is to follow ASTM C 1293 (5.25 kg/m3, 8.75 lb/yd3 Na2Oe). If the expansion of concrete
prisms exceeds 0.040% at one year, the aggregate is deemed to be reactive and the concrete
prism must be subjected to a petrographic examination to determine if ACR has contributed to
the expansion. If ACR is considered to have contributed to the expansion the aggregate must be
rejected for use in concrete; however, if there is no sign of ACR, the aggregate is deemed to be
alkali-silica reaction and can be used in concrete with appropriate preventive measures. If the
aggregate does not produce expansion in excess of 0.040% at 1 year in the concrete prism test it
115
AAR Facts Book
is considered to be non-deleteriously reactive and may be used in concrete without preventive

measures.
The type of reaction only needs to be determined after the concrete prism test if the aggregate being tested is a
quarried carbonate that has been identified as being potentially alkali-carbonate reactive by chemical composition
in accordance with test method CSA A23.2-26A.
The solid lines show the preferred approach. However, some agencies may want to reduce the amount of testing
and accept a higher level of risk and this can be achieved by following the direction of the hashed lines.
Figure 6.1. Sequence of Testing in AASHTO PP65-11 for Evaluating Aggregate Reactivity
116
AAR Facts Book
Figure 6.2. Potential for Alkali-Carbonate Reactivity of Quarried Carbonates based on
Chemical Composition (from CSA A23.2-26A)
6.3.2 Selecting Preventive Measures

PP65-11 provides two approaches for selecting preventive measures, these being: (i) a
performance approach based on laboratory testing, and (ii) a prescriptive approach based on a
consideration of the reactivity of the aggregate, type and size of structure, exposure conditions,
and the composition of cementitious materials being used. The two test methods used for the
performance-based approach are the accelerated mortar bar test for evaluating combinations of
cementing materials and aggregates (ASTM C 1567) and the concrete prism test (ASTM C
1293), with preference given to the latter test. The rationale behind the use of these two tests
and the preference for the concrete prism test is presented in FHWA-HIF-09-001 (Thomas et al.
2008) and in Thomas et al. (2006).
The options for preventive measures included in the prescriptive approach of PP65 are to (i)
control the alkali content of the concrete to a maximum allowable level, (ii) use a minimum
level of supplementary cementing material (SCM) or combination of SCMs, or (iii) use a
combination of these two options (that is controlling the alkali content of the concrete and using
SCM). The precise level of alkali permitted or SCM required depends on a number of factors
including the aggregate reactivity, type and size of structure, exposure conditions, and even the
composition of the cement and SCM being used.
117
AAR Facts Book
A brief summary of the prescriptive and performance options are described here.
6.3.2.1 Prescriptive Approach for Selecting Preventive Measures
The prescriptive approach of AASHTO PP65 can be summarized in the following steps:
Step 1. Determine aggregate reactivity class: The aggregate is tested in either the
accelerated mortar bar test (AMBT), AASHTO T 303, or, preferably, the concrete
prism test (CPT), ASTM C 1293. The criteria in Table 6.1 are used to classify the
aggregate reactivity, which can range from R0 non-reactive to R3 very
highly reactive.
Step 2. Determine level of ASR risk: Based on the aggregate-reactivity class determined
in Step 1 (Table 6.1) and the size and exposure conditions of the concrete under
construction, the level of ASR risk is determined using the criteria in Table 6.2.
The risk may range from Level 1 (lowest or negligible risk) to Level 6 (highest
risk).
Step 3. Determine level of prevention: Based on the level of ASR risk determined in Step
2 (Table 6.2) and the classification of the structure6, the level of prevention
required is determined using Table 6.3. The level of prevention required may
range from Level V (no measures necessary) to Level ZZ (extreme preventive
measures necessary).
Step 4. Identification of preventive measures: Based on the level of prevention required
that was determined in Step 3 (Table 6.3), a number of options are presented as
acceptable measures for preventing ASR; these are:
Option 1 limiting the alkali content of the concrete (Table 6.5)
Option 2 using supplementary cementing materials, SCM7 (Table 6.6 and Table
6.7)
Option 3 limiting the alkali content of the concrete and using SCM (Table 6.8)
The prescriptive approach does not allow the option for using lithium compounds as a
preventive measure. Research has shown that the efficacy of lithium compounds in controlling
expansion due to ASR is highly influenced by the nature of the reactive aggregate (Tremblay et
al. 2007). Currently, it is not possible to prescribe the required lithium dose based on aggregate
reactivity or mineralogy and, consequently, lithium compounds must be tested using the
prescriptive approach to determine the minimum dose required with a specific aggregate.
6
Table 6.4 is intended to provide guidance in selecting the Class of Structure.

For Option 2, the minimum amount of SCM determined from Table 6.6 may be adjusted based on the alkali level
of the portland cement using Table 6.7.
118
AAR Facts Book
Table 6.1. Classification of Aggregate Reactivity

AggregateReactivity Class
Description of Aggre
gate Reactivity
One-Year Expansion 14-Day Expansion in

in CPT (%)
AMBT (%)
R0
Non-reactive
0.04
0.10
R1
Moderately reactive
> 0.04, 0.12
> 0.10, 0.30
R2
Highly reactive
> 0.12, 0.24
> 0.30, 0.45
R3
Very highly reactive
> 0.24
> 0.45
Table 6.2. Determining the Level of ASR Risk

Aggregate-Reactivity Class
Size and exposure conditions
R0
R1
R2
R3
Non-massive1 concrete in a dry2

environment
Level 1
Level 1
Level 2
Level 3
Massive1 elements in a dry2

environment
Level 1
Level 2
Level 3
Level 4
All concrete exposed to humid air,

buried or immersed
Level 1
Level 3
Level 4
Level 5
All concrete exposed to alkalis in

service3
Level 1
Level 4
Level 5
Level 6
A massive element has a least dimension > 3 ft (0.9 m).
A dry environment corresponds to an average ambient relative humidity lower than

60%, normally only found in buildings.
Examples of structures exposed to alkalis (sodium and potassium) in service include

marine structures exposed to seawater and highway structures exposed to deicing salts
(e.g., NaCl) or anti-icing salts (e.g., potassium acetate, potassium formate, sodium
acetate, sodium formate, etc.).
119
AAR Facts Book
Table 6.3. Determining the Level of Prevention

Classification of Structure (Table 6.4)
Level of ASR
Risk (Table 4)
S1
S2
S3
S4
Risk Level 1
Risk Level 2
Risk Level 3
Risk Level 4
Risk Level 5
ZZ
Risk Level 6
ZZ
It is not permitted to construct a Class S4 structure (see Table 6.4) when the
risk of ASR is Level 6. Measures must be taken to reduce the level of risk in
these circumstances.
The level of prevention V, W, X, Y, Z and ZZ are used in Tables 6.5 to 6.8.
120
AAR Facts Book
Table 6.4. Structures Classified on the Basis of the Severity of the Consequences Should ASR1
Occur (Modified for Highway Structures from RILEM TC 191-ARP)
Class
S1
S2
S3
S4
Consequences of
Safety, economic, or
Some deterioration
environmental
from ASR may be
consequences small
tolerated
or negligible
Some safety,
economic, or
environmental
consequences if
major deterioration
Significant safety,
economic, or
environmental
consequences if
minor damage
Serious safety,
economic, or
environmental
consequences if
minor damage
Examples2
Acceptability of
Moderate risk of
ASR is acceptable
Minor risk of ASR

acceptable
ASR cannot be
tolerated
Non-load-bearing elements inside

buildings
Temporary structures (e.g. < 5 years)
Sidewalks, curbs and gutters
Service-life < 40 years
Pavements
Culverts
Highway barriers
Rural, low-volume bridges
Large numbers of precast elements

where economic costs of replacement
are severe
Service life normally 40 to 75 years
Major bridges
Tunnels
Critical elements that are very

difficult to inspect or repair
Service life normally > 75 years
This table does not consider the consequences of damage due to ACR. This practice does not permit the use of
alkali-carbonate reactive aggregates.
The types of structures listed under each Class are meant to serve as examples. Some owners may decide to use
their own classification system. For example, sidewalks or curbs and gutters may be placed in the Class S3.
121
AAR Facts Book
Table 6.5. Maximum Alkali Contents in Portland Cement Concrete to Provide Various
Levels of Prevention
Prevention Level
Maximum Alkali Content of Concrete (Na2Oe)

lb/yd3
No limit
V
W
5.0
3.0
4.0
2.4
3.0
1.8
Z1
Table 6.8
ZZ1
kg/m3
SCMs must be used in prevention levels Z and ZZ.
122
AAR Facts Book
Table 6.6. Minimum Levels of SCM to Provide Various Levels of Prevention
Type of SCM1
Fly ash
(CaO 18%)
Slag
(% by mass of cementitious material)
(% Na2Oe)
Level W
Level X
Level Y
Level Z
3.0
15
20
25
35
> 3.0, 4.5
20
25
30
40
1.0
25
35
50
65
1.2 x LBA
1.5 x LBA
1.8 x LBA
2.4 x LBA
or
or
or
or
2.0 x KGA
2.5 x KGA
3.0 x KGA
4.0 x KGA
Silica Fume
(SiO2 85%)
Minimum Replacement Level3
Alkali level
of SCM
1.0
Level ZZ
Table 6.7
The SCM may be added directly to the concrete mixer or it may be a component of a blended cement. SCMs
should meet the requirements of AASHTO M 295, M 302, or M 307. Blended cements should meet the
requirements of AASHTO M 240 or ASTM C 1157.
The minimum level of silica fume (as a percentage of cementitious material) is calculated on the basis of the
alkali (Na2Oe) content of the concrete contributed by the portland cement and expressed in either units of lb/yd3
(LBA in Table 6.6) or kg/m3 (KGA in Table 6.6). LBA is calculated by multiplying the cement content of the
concrete in lb/yd3 by the alkali content of the cement divided by 100. For example, for a concrete containing 500
lb/yd3 of cement with an alkali content of 0.81% Na2Oe, the value of LBA = 500 x 0.81/100 = 4.05 lb/yd3. For
this concrete, the minimum replacement level of silica fume for Level Y is 1.8 x 4.05 = 8.1%. KGA is calculated
by multiplying the cement content of the concrete in kg/m3 by the alkali content of the cement divided by 100.
For example, for a concrete containing 300 kg/m3 of cement with an alkali content of 0.91% Na2Oe, the value of
KGA = 300 x 0.91/100 = 2.73 kg/m3. For this concrete, the minimum replacement level of silica fume for Level
X is 2.5 x 2.73 = 6.8%. Regardless of the calculated value, the minimum level of silica fume shall not be less
than 7% when it is the only method of prevention.
The use of high levels of SCM in concrete may increase the risk of problems due to deicer salt scaling if the
concrete is not properly proportioned, finished, and cured.
123
AAR Facts Book
Table 6.7. Adjusting the Minimum Level of SCM Based on the Alkali Content
in the Portland Cement
Cement Alkalis
Level of SCM
(% Na2Oe)
0.70
Reduce the minimum amount of SCM given in

Table 6.6 by one prevention level1
> 0.70, 1.00
Use the minimum levels of SCM given in Table 6.6
> 1.00, 1.25
Increase the minimum amount of SCM given in

Table 6.6 by one prevention level
> 1.25
No guidance is given
The replacement levels should not be below those given in Table 6.6 for prevention level W, regardless of the
alkali content of the portland cement.
Table 6.8. Using SCM and Limiting the Alkali Content of the Concrete to Provide
Exceptional Levels of Prevention
SCM as Sole Prevention
Prevention
Level
Limiting Concrete Alkali Content Plus SCM
Minimum SCM Level
Maximum Alkali
Content, lb/yd3 (kg/m3)
Minimum SCM Level
SCM level shown for

Level Z in Table 6.6
3.0 (1.8)
SCM level shown for

Level Y in Table 6.6
ZZ
Not permitted
3.0 (1.8)
SCM level shown for

Level Z in Table 6.6
6.3.2.2 Performance Approach for Selecting Preventive Measures

AASHTO PP65-11 recommends that the concrete prism test (ASTM C 1293) is used to
evaluate the efficacy of supplementary cementing materials or lithium-based compounds for
controlling alkali-silica reaction. Preferably tests should be conducted at a range of SCM levels
124
AAR Facts Book
or lithium doses to determine the safe level. The preventive measure is considered to be
effective provided the expansion of concrete prisms is not greater than 0.040% after 2 years.
If there is insufficient time to conduct concrete tests, ASTM C 1567 (modified accelerated
mortar bar test for evaluating combinations of reactive aggregate and SCM) may be used to
evaluate SCMs. The SCM is deemed to be effective in controlling ASR expansion with the
aggregate under test provided the expansion of mortar bars is not greater than 0.10% after 14
days immersion in 1 M NaOH at 176F (80C).
AASHTO PP65-11 recommends that, before using ASTM C 1567 to determine the amount of
SCM (or lithium required), the aggregate is tested in both the concrete prism test and the
accelerated mortar bar test (ASTM C 1260) to ensure that the latter test gives a reliable
prediction of the aggregate reactivity. The expansions from the two tests are plotted on the
graph shown in Figure 6.3 to determine the validity of using the accelerated test.
Figure 6.3. Comparison of Accelerated Mortar Bar Test (AMBT) and Concrete Prism Test
(CPT) Data for the Purpose of Determining Whether the AMBT is Suitable for Evaluating
Preventive Measures with a Specific Aggregate
125
AAR Facts Book
AASHTO PP65-11 also includes a modified version of the accelerated mortar bar test for
evaluating lithium-based compounds based on the approach proposed by Tremblay et al (2008).
The approach is summarized as follows:
Test the aggregate using the standard accelerated mortar bar test (AASHTO T 303,
ASTM C 1260). Extend the duration of the test such that the mortar bars are exposed to
sodium hydroxide for a period of 28 days. Let E1 = expansion of bars without lithium
nitrate at 28 days.
Test the aggregate in a modified version of the accelerated mortar bar test. In this test
add sufficient lithium nitrate to the mortar bar mixture and the soak solution to achieve
lithium-to-alkali molar ratios of [Li]/[Na+K] = 0.74 in the mortar and [Li]/[Na+K] =
0.148 in the soak solution (see Appendix A2 on calculation of lithium nitrate additions).
Conduct the rest of the test in accordance with AASHTO T 303, extending the period in
sodium hydroxide to 28 days. Let E2 = expansion of bars with lithium nitrate at 28 days.
Note: To achieve [Li]/[Na+K] = 0.74 in the mortar add 4.6 liters of 30%-LiNO3
solution for every 1 kg of alkali (as Na2Oe) in the mix (70.4 fl.oz. of 30%-LiNO3 solution
for every 1 lb of alkali).
There are few data available for calibrating this test method, and it is strongly recommended
that the concrete prism test be used to evaluate lithium-aggregate combinations.
126
AAR Facts Book
REFERENCES
AASHTO. 2012. Standard Specification for Blended Hydraulic Cement. AASHTO M 240,
AASHTO. 2011. Standard Specification for Coal Fly Ash and Raw or Calcined Natural
Pozzolan for Use in Concrete. AASHTO M 295, American Association of State and Highway
Transportation Officials, Washington, DC, 4 p.
AASHTO. 2012. Standard Specification for Ground Granulated Blast-Furnace Slag for Use in
Concrete and Mortars. ASHTO M 302, American Association of State and Highway
Transportation Officials, Washington, DC, 9 p.
AASHTO. 2005. Standard Specification for Silica Fume Used in Cementitious Mixtures.
ASHTO M 307, American Association of State and Highway Transportation Officials,
AASHTO. 2008. Standard Method of Test for Accelerated Detection of Potentially
Deleterious Expansion of Mortar Bars Due to Alkali-Silica Reaction. AASHTO T 303,
ACI Committee 201. 2008. Guide to Durable Concrete. ACI 201.2R-08, American Concrete
Institute, Farmington Hills, MI, 49 p.
ACI Committee 221. 1998. Report on Alkali-Aggregate Reactivity. ACI 221.1R-98,
American Concrete Institute, Farmington Hills, MI, 30 p.
ACI Committee 301. 2010. Specifications for Structural Concrete. ACI 301-10, American
Concrete Institute, Farmington Hills, MI, 77 p.
ACI Committee 308. 2008. Building Code Requirements for Structural Concrete. ACI 318
08, American Concrete Institute, Farmington Hills, MI, 473 p.
ASTM. 2013. Standard Specification for Concrete Aggregate. ASTM C 33, ASTM
127
AAR Facts Book
ASTM. 2012. Standard Specification for Portland Cement. ASTM C 150, ASTM
PA, 6 p.
ASTM. 2012. Standard Guide for Petrographic Examination of Aggregates for Concrete.
ASTM. 2012. Standard Specification for Blended Hydraulic Cements. ASTM C 595, ASTM
ASTM. 2012. Standard Specification for Coal Fly Ash and Raw or Calcined Natural Pozzolan
for Use in Concrete. ASTM C 618, ASTM International, West Conshohocken, PA, 5 p.
ASTM. 2012. Standard Specification for Slag Cement for Use in Concrete and Mortars.
ASTM. 2008. Standard Test Method for Length Change of Concrete Due to Alkali-Carbonate
Rock Reaction. ASTM C 1105, ASTM International, West Conshohocken, PA, 4 p.
ASTM. 2011. Standard Performance Specification for Hydraulic Cement. ASTM C 1157,
ASTM International, West Conshohocken, PA, 5 p.
ASTM. 2012. Standard Specification for Silica Fume Used in Cementitious Mixtures. ASTM
128
AAR Facts Book
Struble, L. 2010. Opening remarks. Workshop on specifications on alkali-aggregate reaction.
Sponsored by ASTM Committees C09 on Concrete and Concrete Aggregates and C01 on
Cement, 6 June, St. Louis, MO, USA. (Available on ASTM website.)
Thomas, M.D.A., Fournier, B. and Folliard, K.J. 2008. Report on determining the reactivity of
concrete aggregates and selecting appropriate measures for preventing deleterious expansion in
new concrete construction. Federal Highway Administration, FHWA-HIF-09-001, National
Research Council, Washington, D.C., 20 p.
Thomas, M.D.A., Fournier, B. and Folliard, K.J. 2012. Selecting Measures to Prevent
Deleterious Alkali-Silica Reaction in Concrete: Rationale for the AASHTO PP65 Prescriptive
Approach. Federal Highway Administration, FHWA-HIF-13-002, National Research Council,
Washington, D.C., 47 p.
Tremblay, C., Brub, M-A., Fournier, B., Thomas, M.D.A. and Folliard, K.F. 2007.
Effectiveness of lithium-based products in concrete made with Canadian reactive aggregates.
ACI Materials Journal, 104(2): 195-205.
Tremblay, C., Berube, M-A., Fournier, B., Thomas, M.D.A. and Folliard, K.J. 2008. Use of
the Accelerated Mortar Test to Evaluate the Effectiveness of LiNO3 against Alkali-Silica
Reaction. Part 2: Comparison With Results from the Concrete Prism Test. Journal of ASTM
International, 5(8).
129
AAR Facts Book
130
AAR Facts Book
Chapter 7 Diagnosis and Prognosis
7 Diagnosis and Prognosis of Alkali-Silica Reaction in Concrete Structures
7.1 INTRODUCTION
Alkali-silica reaction (ASR) is one of the many factors that might be fully/partly responsible for
the deterioration and premature loss in serviceability of concrete infrastructure. Comparative
field and laboratory investigations can be carried out to confirm that AAR is the main cause or a
significant contributor to the deterioration observed, thus aiming at selecting appropriate
remedial actions. Such investigations will likely include one or several of the following steps:
the field inspection of the structure(s) under study to identify the presence/distribution
and severity of the defects affecting the various structural elements (especially those
features diagnostic of AAR), as well as the exposure conditions to which the above
components of the structure are subjected;
the in-situ monitoring of deterioration (especially signs of expansion/deformation, extent
and progress of cracking, spalling, etc.); and
a range of laboratory tests (including petrographic characterization, chemical, physical,
and mechanical tests) on samples collected from one or several components of the
affected concrete structure (especially those showing features of deterioration typically
associated with or diagnostic of AAR).
7.2 MANAGEMENT PROGRAM FOR ASR-AFFECTED STRUCTURES

A global approach was recently proposed by FHWA for the diagnosis and prognosis of ASR in
transportation structures (Fournier et al. 2010); the main features of the above protocol are
illustrated in Figure 7.1. This step-by-step approach aims at evaluating the cause of concrete
distress (diagnosis) and the progress or potential for future expansion/damage (prognosis),
both elements providing information for the selection of appropriate mitigation measures in
ASR-affected structures. The extent to which each of the various methods proposed in the
above approach will need to be implemented in a particular case will depend upon different
factors, including the nature/extent of the problem, the criticality of the structure, the potential
impact on the safety of users, etc.
131
AAR Facts Book
Figure 7.1. Global Approach for the Diagnosis and Prognosis of ASR in Concrete Structures
(modified from Fournier et al. 2010)
7.2.1 ASR Investigation Program Step 1 - Diagnosis
7.2.1.1 Condition Survey
Signs of premature deterioration in concrete transportation structures that could be related to
ASR can generally be detected during routine site inspections of concrete structures (condition
survey). For example, FHWA recently proposed Alkali-Silica Reactivity Surveying and
Tracking Guidelines (Thomas et al. 2012a) that are intended to assist engineers and inspectors
in tracking and surveying ASR-induced features of deterioration in bridges, pavements, and
tunnels.
In the case of bridge structures, basic inspection data can first be collected for the various
elements/components of the structure, for example in accordance with the new AASHTO Guide
Manual for Bridge Element Inspection (AASHTO 2011). The concept is to identify defects that
are specific/unique enough so the manifestation of distress may be attributed to ASR. The
presence and extent of ASR-related defects are then noted and quantified in accordance with the
four condition states described in Table 7.1. The Alkali-Silica Reactivity Field Identification
Handbook (Thomas et al. 2011, updated from Stark (1991)) provides detailed coverage on the
132
AAR Facts Book
causes and effects of ASR and a wide range of photographs illustrating common features of
ASR, as well as examples of condition states (or severity ratings) for a range of transportation
structures. Such features include map cracking, aligned cracking, gel exudation, and relative
dislocation/ misalignment of adjacent sections. Table 7.2 summarizes the proposed ASR-related
defects, defect descriptions, and criteria/threshold for bridge element types using a similar
approach as in the AASHTO guide manual (2011). More detailed descriptions of the above
defects are provided in Table 7.3; examples of defects with different condition states are given
in Figure 7.2. In addition, environmental conditions, especially temperature, relative humidity,
exposure to sun and winds, and precipitation, can be tracked and coupled with other inspection
findings to attempt linking the specific climatic conditions to the progress of ASR. For instance,
concrete that is subjected directly to moisture and exposed to sun (i.e., wetting-drying cycles) is
likely to exhibit more severe symptoms of ASR (e.g., cracking) than concrete that is less
exposed to moisture in the same structure of even the same element (see chapter 3).
Table 7.1. Standard Condition States for Defects in Bridge Elements (after AASHTO 2011)
Condition State #
Condition State
Good
Fair
Poor
4*
Severe
* Condition state 4 (severe) is typically reserved for conditions

that warrant safety concerns and that are beyond the range of
defects described in condition states 1 through 3.
133
AAR Facts Book
Table 7.2. Recommended Defects and Condition States for Bridge Elements Potentially
Affected by ASR (Thomas et al. 2012a)
Defect
Condition
State 1
Condition
State 2
Condition
State 3
Map
Cracking
None to
hairline
Narrow size or
density, or both
Medium size or density, or

both
Aligned
Cracking
None to
hairline
Narrow size or
density, or both
Medium size or density, or

both
Gel
Exudation
None
Moderate
Severe (with gel staining)
Relative
Dislocation/
Misalignment
None
Tolerable
Approaching or exceeding
limits (including causing
local crushing)
Condition State 4
The condition is beyond the

limit state of Condition State 3,
warrants a structural review to
determine the strength or
serviceability of the element or
bridge, or both
Table 7.3. Recommended Defects and Condition States for Bridge Elements Potentially
Defect
Hairline-Minor
Narrow-Moderate
Medium-Severe
Map Cracking
Crack width < 0.0625

(1.6 mm)
% Map Cracking < 5%
Crack width: 0.0625 (1.6 mm)

0.1250 (3.2 mm)
Crack width > 0.1250 (3.2 mm)
Crack width < 0.0625

(1.6 mm)
Crack width: 0.0625 (1.6 mm)

0.1250 (3.2 mm)
Crack width: > 0.1250 (3.2

mm)
None
Gel visible on surface (< 20%

of concrete surface, with no
build-up of gel)
Gel build-up on surface (> 20%

of concrete surface), typically at
or near cracks; gel staining
visible (especially once structure
dries after a rain event)
None
Tolerable (movement is visible

but no loss of clearance,
exudation of sealants at joints,
or local crushing)
Movement is visual, with loss of

clearance, exudation of sealants
at joints, or local crushing
Aligned
Cracking
Gel Exudation
Relative
Dislocation/
Misalignment
% Map Cracking: 5 to 25%
134
% Map Cracking > 25%
AAR Facts Book
Figure 7.2. ASR-related defects in bridge elements and severity assessment. A. Map cracking
on bridge abutment (condition state 2). B. Map cracking on bridge abutment (condition state 3).
C. Map cracking in bridge deck (condition state 1). D. Aligned (longitudinal) cracking in
bridge deck (condition state 2). E. Aligned (longitudinal) cracking in post-tensioned concrete
bridge girder (condition state 2). F. Aligned (longitudinal) cracking in reinforced concrete beam
(condition state 3).
135
AAR Facts Book
Figure 7.2 (contd). ASR-related defects in bridge elements and severity assessment. G. Map cracking
on a parapet wall (condition state 3); moderate gel exudations are also present, which give the
impression of permanent dampness. H. Map cracking in massive column foundation block of a bridge
structure (condition state 3). I. Aligned (vertical/longitudinal) cracking in reinforced concrete column
(condition state 2). J. Severe (condition state 3) aligned (longitudinal) cracking connected by moderate
map cracking in a reinforced concrete column. K. Aligned (longitudinal) cracking in the top chord of a
reinforced concrete arch bridge (condition state 3 to 4).
136
AAR Facts Book
Figure 7.2 (contd). ASR-related defects in bridge elements and severity assessment. L. Closure
of joint between adjacent parapet wall sections causing the buildup of concentrated stresses
which led to crushing of the concrete. M. Severe relative dislocation/misalignment, leading to
loss of clearance and localized spalling (photo from SHRP C-315 1991). N. Aligned
(longitudinal) cracking with associated white exudations (mainly calcium carbonate with lesser
amounts of potassium indicating both lime and alkali leaching from the concrete) on the
underside of the arch shown in Figure 7.2K. O. Moderate aligned cracking and exudations in
the exposed portions of abutment and wing wall of a bridge structure (condition state 2).
A similar approach can be adopted for pavements. The distresses most commonly associated
with ASR in pavements include map cracking, joint deficiencies/deterioration, and, in some
cases, popouts. The latter are generated by reactive aggregate particles causing expansion close
to the surface of the pavement (or other concrete elements), thus inducing the detachment of a
conical portion of the surface leaving the reactive aggregate in the bottom (Figure 7.3J).
Popouts can also be caused by the expansion of frost-susceptible aggregate particles (such as
laminated, schistose and argillaceous, clayey or porous particles or certain varieties of chert,
137
AAR Facts Book
ironstones), as well as by a poor bond between the cement paste and dusty coarse aggregate
particles.
Table 7.4. Recommended Distress and Condition States for Pavements Potentially
Distress
Condition
State 1
Condition
State 2
Condition
State 3
Map Cracking
None to
hairline
Narrow size or
density, or
both
Medium size or
density, or both
Joint Sealant Failure
None
Moderate
Severe
Joint Deterioration
None
Moderate
Severe
Popouts
None
Moderate
Severe
Condition
State 4
The condition is beyond the

limit state of Condition State 3,
warrants a structural review to
determine the strength or
serviceability of the pavement,
or both.
Table 7.5. Recommended Distress and Distress Description for Pavements Potentially
Defect
Map Cracking
Joint Sealant
Failure
Joint
Deterioration
Popouts
Hairline-Minor
Narrow-Moderate
Medium-Severe
Crack width < 0.0625 (1.6 mm) Crack width: 0.0625 (1.6
mm) 0.1250 (3.2 mm)
% Map Cracking < 5%
% Map Cracking: 5 to 25%
Crack width > 0.1250

(3.2 mm)
Joint sealant failure in less than

10% of joints.
Joint sealant failure in 10 to

50% of joints.
Joint sealant failure in

greater than 50% of joint
None or only minor cracking

near corners/joints
Wide, open cracks exist and

mass loss has occurred in
joint region (less than 5%
of joints). No patching
applied.
Wide, open cracks and mass

loss has occurred in joint
region (greater than 5% of
joints). Patching has been
applied.
None
Popouts isolated and few

[< 1 popout per 10 ft]
Popouts prevalent
% Map Cracking > 25%
[> 1 popout per 10 ft]
* Popout data generally not collected and not included in LTPP. Estimates are in []
138
AAR Facts Book
Expansion
Expansion
Figure 7.3. ASR-related defects in pavements and severity assessment. A. Minor cracking in a concrete
pavement (condition state 1). B. Cracking at the corner of paving slabs (condition state 2). C. Map
cracking of concrete pavement; cracking is much more evident after rainy periods (condition state 2).
D. Longitudinal cracking connected by map cracking in severely damaged concrete pavement
(condition state 3) (photo from SHRP C-315 1991). E. Enlargement (red rectangle) of a section of the
pavement in figure D showing the aligned (longitudinal) cracking in pavement with connecting map
cracking.
139
AAR Facts Book
Figure 7.3 (contd). ASR-related defects in pavements and severity assessment. F. Expansion (with
related map cracking) in pavement slabs causing the extrusion of joint sealant material. G. Spall
occurring at a joint in a concrete pavement; as ASR advances the cracked concrete at the vicinity of the
joint may be further distressed by the action of freezing and thawing, and vehicular loading. Eventually
spalling starts to occur at the joint. H. Joint Sealant Failure in Concrete Pavement (condition state 3). I.
Joint deterioration in concrete pavement (condition state 3). J. Popout involving a porous (leached)
chert particle (with complements of C.A. Rogers).
140
AAR Facts Book
It is important to mention that if the condition survey points out issues that can impair the
integrity of the structure or public safety (related or not to ASR), immediate action should be
taken in consultation with experts in the respective fields.
Conclusions from the Condition Survey
Signs of premature deterioration in concrete pavement and bridge structures that could be
related to ASR can generally be detected during routine site inspections that are performed
regularly by trained personnel (e.g., inspectors from State Highway Authorities) (Figure 7.1).
Visual symptoms of deterioration are noted and compared to those commonly observed on
structures affected by ASR (Figures 7.2 and 7.3).
It is often difficult to determine only from field observations whether ASR is the only/main
factor responsible for the observed distresses since some of the visual signs of deterioration
generally associated with ASR may have been caused by other processes such as internal
sulphate attack, or plastic or drying shrinkage. Table 7.6 classifies the occurrence of the features
obtained from the condition survey as indicative of low, medium, and high potential of ASR
contribution in the deterioration observed. The assessment of the exposure conditions should
also contribute to support the observations of the symptoms of distress listed in Table 7.6, as
follows:
Low potential for ASR: Element in dry and sheltered environment.
Medium potential for ASR: Element exposed outdoors but sheltered from constant wetting.
High potential for ASR: Parts of components frequently exposed to moisturee.g., rain,
groundwater, or water due to natural function of the structure (e.g., hydraulic dam or
bridge).
141
AAR Facts Book
Table 7.6. Classification System for the Condition Survey (modified from CSA A864-00)
Feature
Probability for ASR

Low
Medium
High
Expansion
and/or
displacement
of elements
None
Some evidence (e.g., closure of joints

in pavements, jersey barriers, spalls,
misalignments between structural
members)
Fair to extensive signs of volume increase

leading to spalling at joints, displacement
and/or misalignment of structural members
Cracking and
crack pattern
None
Some cracking pattern typical of ASR

(e.g., map cracking or cracks aligned
with major reinforcement or stress)
Extensive map cracking or cracking aligned

with major stress or reinforcement
Surface
discoloration
None
Slight surface discoloration associated

with some cracks
Many cracks with dark discoloration and

adjacent zone of light colored concrete
Exudations
None
White exudations around some cracks;

possibility of colorless, jelly-like
exudations
Colorless, jelly-like exudations readily

identifiable as ASR gel associated with
several cracks
If the probability of ASR is low or no visual signs suggestive of ASR are noted during the
routine inspection program, further work can be postponed until the next inspection. However,
when the visual signs of deterioration observed on the structure(s) examined are such that AAR
is a possibility, sampling of the structure under evaluation is recommended to confirm the first
diagnostic obtained from the visual survey.
7.2.1.2 Documentation
Any documents (i.e., testing reports of materials, construction and/or inspection reports) related
to the structure examined should be gathered as they may provide valuable information in the
appraisal process. This activity could also be carried out either in preparation for the condition
survey or following it, i.e., for structures where some signs of deterioration potentially
indicative of ASR have been noticed. Useful information could include the following (CSA
2000):
Type and location of the structure and, hence, its likely exposure conditions due to the
nature of operation(s) and geography.
Age of the structure and details and dates of any modifications or repairs. ASR may take
from 3 to even more than 25 years to develop significantly in concrete structures
depending on factors such as the nature (reactivity level) of the aggregates used, the
moisture and temperature conditions, and the concrete alkali content.
Plans, drawings, and original specifications.
Details of concrete mixes used, particularly mix proportions, source of cement and
aggregates, and details of any analyses or tests carried out on concrete materials. The
availability of samples of these materials should also be checked; some agencies store
142
AAR Facts Book
samples of cements and aggregates used in major projects.

Previous inspection/testing reports, especially dates when deterioration was first
observed.
Information from the inspection of other local structures that may have been constructed
with similar materials, especially if these structures are exhibiting signs of deterioration
typical of ASR.
Details regarding the concrete materials, especially the composition and proportion of the
cement and the type of aggregate used, are most useful when assessing the likelihood of ASR. It
is recognized that information of this nature is often not available or lacks specific details in the
case of many structures; however, it is important to collect whatever data are available.
7.2.1.3 Sampling of Structural Components
At this stage of the program, sampling is carried out on component(s) of the structures showing
typical defects suggestive of ASR, which most often correspond to structural components
exposed to a constant or renewable supply of moisture, with/without cycles of wetting and
drying (e.g., Figure 7.4A-7.4C). This is essentially done to determine whether or not the
concrete contains petrographic evidence of ASR. For comparison purposes, it will also be
useful/appropriate to collect cores from structures/components that are less deteriorated than the
structure in question, not deteriorated, or not exposed to the severe environmental elements.
143
AAR Facts Book
Figure 7.4. A-C. Sampling of concrete structures/elements affected by ASR for laboratory
investigations (e.g., petrographic examination, mechanical testing). D. Cores are properly
identified and wrapped in plastic film/bag to avoid moisture loss.
144
AAR Facts Book
The samples collected should be labeled carefully, photographed, and, immediately after their
preliminary examination (for obvious signs of degradation.), wrapped in a plastic film and
sealed in a plastic bag to prevent alkali-silica gel and surfaces to carbonate, become
contaminated, or dry out during subsequent transport and storage (Figure 7.4D). More details
regarding sampling (i.e., type and sample size) and preliminary treatment/analysis of samples
can be found in Fournier et al. (2010).
7.2.1.4 Diagnosis Petrographic Examination
Petrographic examination is a powerful technique in the diagnosis of the cause of concrete
deterioration. ASTM C 856 outlines procedures for the petrographic examination of samples of
hardened concrete. Interesting information regarding petrographic features of ASR-affected
concrete can be found in several publications, including BCA (1992), St. John et al. (1998),
CSA (2000), Walker et al. (2006) and Fournier et al. (2010).
Macroscopic Features of ASR
Macroscopic signs of concrete deterioration due to ASR (but sometimes not exclusively) can be
detected by examining the cores immediately after the extraction or in the laboratory in an as
received condition. Such features can consist of (Figure 7.5): macrocracks in the outer portion
(or "skin") of the concrete member turning into microcracking in the "internal" part of the
concrete member, gel staining surrounding surface cracks, dark reaction rims at the periphery of
reacted aggregate particles, cracks within reactive aggregate particles sometimes extending into
the cement paste (with/without reaction products gels), alkali-silica gel in air voids of the
cement paste, and deposits of reaction products on the cracked surfaces of cores.
145
AAR Facts Book
Figure 7.5. Macroscopic observations on cores (as received condition). A. Cores extracted from a
concrete pavement affected by ASR and showing macrocracks penetrating from the upper and lower
parts of the pavement. B. Gel staining surrounding cracks and gel exudations at the surface of a core
extracted from a sidewalk section affected by ASR. C. Fine cracking pattern showing up after re-wetting
of the core. D. Dark rim surrounding reactive aggregate particles. E. Macrocracks in reactive coarse
aggregate particles. F. Deposits of alkali-silica gel in a void on core surface. G. Deposits of alkali-silica
reaction products on the broken surface of a core extracted from a highway bridge structures affected by
ASR.
146
AAR Facts Book
Microscopic Features of ASR

Petrographic features of ASR can be detected on various types of specimens prepared from the
drilled cores (e.g., polished sections or slices, broken (fresh) surfaces, and thin sections).
Although not necessarily exclusive to ASR, these features generally consist of the following:
Microcracking in aggregates and/or cement paste
Reaction product gel
Reaction rims
Loss of the cement paste-aggregate bond
Microcracking - BCA (1992) and St. John et al. (1998) compared idealized cracking patterns
in concrete specimens affected by various deleterious mechanisms (Figure 7.6). In the case of
concrete affected by ASR when the reactive fraction is part either of the fine aggregate (Figure
7.6E) or coarse aggregate (7.6F) fraction, a network of microcracks develops in the inner part of
the concrete, with only a few macrocracks being observed in its outer (surficial) portion. The
microcracks are found connecting the aggregate particles; when the reactive material is found in
the coarse aggregate particles, cracks typically run through the particles (Figure 7.6F).
In the early stages of the reaction, microcracks are generally limited to the reacting aggregate
particles and the cement paste-aggregate interface. With the progress of expansion,
microcracks, more or less filled with alkali-silica gel, will extend from the aggregate particles
into the cement paste (Figures 7.7A-C); depending on the extent of expansion, the cracks will
cover considerable distances through the paste where they are often filled with secondary
reaction products (Figures 7.7A, 7.7D, 7.7E). In badly deteriorated concrete specimens, cracks,
even filled with gel, may run through non-reactive aggregate particles. Consequently, great care
should be taken to correctly identify the sites (or the aggregate particles) that have generated the
expansive forces.
147
AAR Facts Book
Figure 7.6. Typical cracking patterns in concrete affected by different deleterious mechanisms (from
BCA 1992). A. Internal crack pattern which can be induced by drying shrinkage. B. Internal crack
pattern which can be caused by internal sulfate attack for delayed ettringite formation (DEF), or
from sulfates derived from the aggregates. C. Internal crack pattern which can be induced by
shrinkage of the coarse aggregate. D. Internal crack pattern which can be induced by frost attack.
E. Internal crack pattern which can be caused by ASR: reactive silica in the sand fraction. F.
Internal crack pattern which can be caused by ASR: reactive silica in the coarse aggregate.
148
AAR Facts Book
2,5 cm
B (distance between the vertical lines = 1 cm)
C (distance between the vertical lines = 1 cm)
D (distance between the vertical lines = 1 cm)
E (distance between the vertical lines = 1 cm)
Figure 7.7. Petrographic features of ASR observed on polished concrete sections. Cracks with reaction
products in coarse (A, C, E) and fine (C) aggregate particles. Cracks running from one aggregate particle
to another through the cement paste (A-C). Reaction rims around reactive coarse aggregate particles (E,
B). Reaction products (ASR gel) in cracks (B, D) and voids (A, B, E) of the cement paste.
149
AAR Facts Book
0.5 mm
2 mm
J
5 mm
10 mm
Figure 7.7 (contd). Petrographic features of ASR. F&G. Polished concrete sections treated with uranyl
acetate solution to enhance the presence of alkali-silica gel (F. natural light; G. under UV illumination
showing the gel in greenish-yellow staining color filling cracks in the cement paste in the vicinity of
reactive aggregate particles). H&I. Thin sections micrograph showing reaction products (ASR gel) in
cracks of coarse aggregate particles and of the cement paste (H), as well as in an adjacent air void of the
cement paste (I). J&K. Micrographs of concrete sample under the stereomicroscope showing alkali-silica
reaction products on broken surfaces of reactive coarse aggregate particles and in voids (J), or covering
cracked surfaces of the cement paste (K).
150
AAR Facts Book
Reaction product gel - ASR generates secondary reaction products containing silica, alkalis
and calcium as typical constituents. The so-called alkali-silica gel will be found filling cracks
within the aggregate particles (Figures 7.7B-7.7E), lining or filling voids (Figures
7.7B,7.7H,7.7J) and fractured surfaces of the cement paste and the aggregate particles (Figures
7.7J,7.7K). These deposits will cover more or less important surfaces depending on many
factors, such as the extent of the reaction-expansion processes that have occurred, the
availability of water, etc. However, the abundance of gel deposits is not necessarily indicative
of the magnitude of any resultant expansion and cracking (BCA 1992); cracking due to ASR
has been observed in many concrete structures while very little gel was found in concrete
specimens taken from the affected members.
The confirmation of the presence and the nature of reaction products is not easy. This is often
done under the scanning electron microscope equipped with an energy dispersive X-Ray
analyzer (Fournier et al. 2010). Staining techniques have also been proposed to facilitate
identification of the reaction product gel in concrete affected by ASR (Natesaiyer et al. 1991;
Stark 1991; Guthrie and Carey 1997). Natesaiyer et al. (1991) proposed a method that consists
of applying an uranyl acetate solution on polished or fresh broken surfaces of concrete
specimens to be examined followed by a visual observation of the section under a UV light.
Stark (1991) indicated that by applying the uranyl acetate solution to a surface containing the
gel, the uranyl ion substitutes for alkali in the gel, thereby imparting a characteristic yellowishgreen glow when viewed in the dark using short wavelength ultraviolet light ASR gel fluoresces
much brighter than cement paste due to the greater concentration of alkali and, subsequently,
uranyl ion in the gel (Figures 7.7F-7.7G). This technique, which has also been applied to field
structures (Stark 1991; AASHTO 1993; ASTM C 856-02), should be used with great care
following appropriate health and safety procedures because of the potentially hazardous nature
of the product. Also, technically speaking, the results of the test should be interpreted with great
care since some aggregates fluoresce naturally, which can incorrectly suggest the presence of
alkali-silica gel through macroscopic or microscopic examinations. Guthrie and Carey (1997)
proposed a method that consists in treating fresh concrete surfaces to successive applications of
Sodium Cobaltinitrite and Rhodamine B. Upon treatment, regions affected by ASR stain either
yellow or pink. According to Guthrie and Carey (1997), yellow staining would be associated
with massive ASR-related precipitate with gel-like morphology as well as granular precipitate
consisting of crystals that have grown from the gel. Yellow stained regions would correspond to
alkali-bearing siliceous reaction products resulting from ASR.
Reaction rims - Dark reaction rims are observed at the internal periphery of a number of alkali
-silica reactive aggregates in deteriorated concrete specimens; these are particularly evident on
polished sections or slabs of affected concrete cores (Figures 7.7B,7.7E). However, these rims
must not be mixed up with dark rims that are often found in the outer (but also internal) portion
of "weathered" gravel particles. When concrete cores are fractured for examining fresh
151
AAR Facts Book
broken surfaces, cracks that have formed within the aggregate particles and the cement paste,
due to the ASR processes, will form zones of weakness where the core will preferentially break.
The fractured surfaces thus created (which in many cases correspond to ASR cracking
surfaces) often show a dark rim surrounding internal deposits of whitish color (Figure 7.7J).
Such a feature does not correspond to a "reaction rim" per se; it actually corresponds to a
typical arrangement of reaction products deposited on the cracking surface, i.e., 1) a dark rim
covering the immediate internal periphery of the particle, and 2) white deposits going through
the central portion of the particle showing a powdery aspect.
Loss of the cement paste-aggregate bond - The interfacial region between the cement paste
and the aggregate particles certainly represents, because of its nature and the arrangement of
hydrates that form herein, a preferential zone of weakness where cracks will initiate and run.
Loss of the cement paste-aggregate bond has been reported as a petrographic consequence but
is not necessarily indicative of AAR.
Conclusions from the Petrographic Examination
Table 7.7 classifies the occurrence of the features obtained from the petrographic examination
as indicative of low, medium, and high probability of ASR. When petrographic evidence of
ASR is confirmed, a decision on further steps to be taken is based on factors such as the
severity of the damage and the "criticality" of the structure (Fournier et al. 2010) (Figure 7.1).
In some cases, it may be decided that additional technical investigations are not required and
some remedial actions could/should already be implemented. Examples of such cases are
further discussed in Fournier et al. (2010). Also, in some cases, the extent of the damage is such
that no immediate action is needed; the structure will then be re-examined as part of the routine
condition survey (Figure 7.1). However, in the case of critical structures (e.g., large size/
major highway bridges and pavements, hydraulic dams) or when the extent of deterioration is
significant, a detailed laboratory and/or in-situ investigation program may be desirable/
necessary to determine the potential for the progress of ASR in the structure (prognosis).
152
AAR Facts Book
Table 7.7. Classification System for Petrographic Examination (Fournier et al. 2010)
Probability of ASR
Nature and Extent of Features

No potentially reactive rock types (from petrographic examination of thin sections):
Low
no alkali-silica gel present (or only in a very few air voids), no (or very few)
reaction rims, no (or very few) sites of expansive reaction, very limited cracking
within the aggregate particles that extends, or not, into the cement paste;
presence of other indicative features rarely found.
Presence of some features generally consistent with AAR:
Medium
damp patches on core surfaces;

presence of potentially reactive rock types;
cracking/microcracking within a fair number of aggregate particles; some of the
cracks may extend into the cement paste;
alkali-silica gel observed in cracks within a fair number of aggregate particles
and/or cracks within the cement paste and/or air voids;
darkening of cement paste around reactive aggregate particles, cracks or voids;
reaction rims around the internal periphery of a fair number of reactive particles.
Presence of extensive signs of ASR (as described in the previous section but observed
in larger frequency), for instance:
High
evidence of site of expansion reaction, i.e., where evidence or reaction and

emanation of swelling pressure can be positively identified; and/or
presence of gel in cracks and voids associated with several reactive particles and
readily visible to the unaided eye or under low magnification.
7.2.2 ASR Investigation Program Step 2 - Prognosis

This part of the study aims at generating additional technical information, leading to a more
complete assessment of the degree of damage due to ASR in the concrete structure and the
selection of most appropriate remedial measures (Figure 7.1). The selection of the activities will
depend on the criticality of the structure, the amount of time/funding available to generate the
data, and the degree of precision expected.
7.2.2.1 In-Situ Investigations
An in-situ investigation program, which could include monitoring of expansion, deformation,
availability of moisture (to account for seasonal temperature changes), and non-destructive
testing, can provide prognostic information for ASR-affected structural members when
carried out at regular intervals over several years. The in-situ investigation program can include
one or several of the activities described hereafter. Details on the procedures/methods can be
found in Fournier et al. (2010).
153
AAR Facts Book
Surface Cracking
The extent of surface cracking on severely exposed/cracked sections of concrete elements is
somewhat related to the overall amount of expansion reached by the affected member. The
Cracking Index (CI) method consists of the measurement and summation of crack widths along
a set of lines drawn perpendicularly on the surface of the concrete element investigated (LCPC
1997, 1999) (Figure 7.8).
154
AAR Facts Book
O
B
A
C
Interval (10 cm)

OA
Vertical
direction
Horizontal
direction
BC
OB
AC
10
-
-
-
0.2
0.3, 0.5
-
-
0.1, 0.2
-
0.4
0.1
0.2
0.4
-
-
-
0.6
-
0.1
-
0.3
0.6
0.3
0.2
0.1, 0.1 0.2

0.4
-
0.1
0.4
-
0.3
-
-
0.4
0.3
0.5
0.2
0.05
0.05
Base
#
Length
cracks
(m)
Crack opening (mm)

Total
Avg.
Avg.
C.I.
sum
/crack
/m
mm/m
0.5
1.8
0.3
3.6
0.5
1.4
0.2
2.8
0.5
2.8
0.4
5.6
3.2
4.6
0.5
1.6
0.2
3.6
Figure 7.8. Measurements of the Cracking Index on ASR-affected pavement (A) and jersey barrier (B).
The cracks can be examined with a magnifying lens (C). Using a transparent crack card to measure their
width along the measurement lines (D). An example of the data generated in the Cracking Index method
is given in (E); the values of C.I. are given separately for the vertical and horizontal measurements.
The method gives a quantitative assessment of the extent of cracking in structural members,
either punctually (when obtained at a specific time) or as a progressive process when
performing the measurements on a regular basis at the exact same location (rate of expansion
prognosis).
155
AAR Facts Book
For example, Francoeur et al. (2012) used the CI method to study the progress of expansion and
cracking in metric size concrete blocks exposed outdoors and suffering from deleterious
aggregates reactions (oxidation of sulphide-bearing aggregates). The authors found a good
correlation between the expansion of the concrete blocks and the progress of the surface
cracking measured through the CI Method (Figure 7.9).
A
0,130
0,090
Cracking Index (mm/m)
Expansion (%)
0,110
4,00
Jan2012
May2012
Aug2012
0,070
0,050
0,030
0,010
-0,010
CIMay 2011
CIMay 2012
CIAug2012
3,50
3,00
2,50
2,00
1,50
1,00
0,50
0,00
1
Blocks
Blocks
Figure 7.9. Measurements of Expansion and Cracking Index on Concrete Housing Foundations
Suffering from Deleterious Aggregate Reactions (Francoeur et al. 2012)
Expansion and Deformation Measurements

Measurements of expansion/deformation can be performed using different means (e.g., demec
points, metallic references drilled into the surface of selected structural members,
extensometers, invar wires/rods or optical systems (leveling)) (Fournier et al. 2010; LCPC 2003,
2009; Thomas et al. 2012b) (Figure 7.10). Fiber-optic and vibrating wire systems can also be
used, deformation measurements being performed and the data transmitted automatically to
central servers for further treatment. On a larger scale, instruments such as inclinometers and
inverted pendulums can be installed in strategic parts of the structure to evaluate the relative
156
AAR Facts Book
movement, deflections, clearances at joints, etc. (e.g., Thompson et al. 1995; Danay et al. 1993;
Gaudreault 2000).
157
AAR Facts Book
Figure 7.10. In-situ measurements of expansion and deformation in ASR-affected concrete

elements. A-E. Length-change measurements in a concrete highway barrier wall, reinforced
concrete columns and a concrete pavement affected by ASR. F-G. Invar-bar extensometer used
for the in-situ monitoring of deformation/movement associated with ASR (Gaudreault 2000).
158
AAR Facts Book
Temperature and Humidity Measurements

The relative humidity is one of the essential conditions to maintain ASR in a concrete structure;
it can be measured over time with depth or laterally in different concrete elements using various
techniques (e.g., wooden stick, portable or permanent probes) (Thomas et al. 2012b; Jensen
2004) (Figure 7.11). The monitoring of relative humidity provided very useful information for
the understanding of the beneficial effect of silane treatments on median barriers affected by
ASR (Brub et al. 2002a).
A
Figure 7.11. In-situ measurements of humidity in concrete. A. Wooden-stick method. B-C.

Portable humidity probes. D. Automatic in-situ monitoring of temperature, humidity and
expansion (vibrating wire) in a bridge deck affected by ASR (Siemes and Gulikers 2000).
Non-Destructive Testing
Periodic measurements, such as pulse velocity, impact echo, acoustic methods, etc. can be made
on specific members of the affected structure (at the surface or in the bottom of drilling holes)
to determine the evolution/extent of internal cracking or deterioration (Thomas et al. 2012b;
Moradi-Marani et al. 2011; Tajari et al. 2011; Sargolzahi et al. 2010) (Figure 7.12).
159
AAR Facts Book
Figure 7.12. Non-destructive testing of ASR-affected concrete structures. A. Impact-echo.

B. Non-linear acoustic technique.
7.2.2.2 Advanced Laboratory Investigations
A series of tests can be performed on core samples extracted from structural members showing
different degrees of deterioration and/or of exposure conditions. The data will provide
information for further evaluating the current concrete condition, the expansion reached to date,
the current rate of expansion, and the potential for future expansion of the concrete.
Quantitative Petrography
The Damage Rating Index (DRI), a semi-quantitative petrographic technique, evaluates the
condition of concrete by counting, under the stereomicroscope (~16x magnification), the
number of typical petrographic features of ASR in a grid system drawn at the surface of
polished concrete sections (Grattan-Bellew 1992) (Figure 7.13). The DRI represents the
normalized value (to 100 mm2) of the presence of these features after the count of their
abundance over the surface examined has been multiplied by weighing factors representing
their relative importance in the overall deterioration process (Grattan-Bellew and Mitchell
2006) (Table 7.8). The method gives a quantitative assessment of the extent of internal damage
in structural concrete members, exposed or not to moisture, either punctually (i.e., at a specific
time - diagnosis) or as a progressive process when performing the measurements on a regular
basis on cores extracted at the same location (prognosis).
160
AAR Facts Book
1cm
1cm
Figure 7.13. Examination of the polished concrete section under the stereomicroscope for the
determination of the Damage Rating Index. Petrographic features of deterioration (Table 7.8;
Figure 7.7) are counted in a one cm x one cm grid system drawn at the surface of the polished
concrete section.
Table 7.8. Petrographic Features and Weighing Factors for the DRI
(Grattan-Bellew and Mitchell 2006)
Petrographic feature
Abbreviation
Weighing factor
CrCA
x 0.75
OCrCA
x 4.0
Cr + RPCA
x 2.0
CAD
x 3.0
RR
x 0.5
CrCP
x 2.0
Cement paste with cracks and reaction products
Cr + RPCP
x 4.0
Air voids lined or filled with reaction products
RPAV
x 0.50
Coarse aggregate with cracks

Open crack in coarse aggregate
Coarse aggregate with cracks and reaction products
Coarse aggregate debonded
Reaction rims around aggregate
Cement paste with cracks
Thomas et al. (2012) found a strong correlation between the extent of ASR (obtained from the
DRI) and the exposure conditions of structural bridge members (Figure 7.14). Rivard et al.
(2000) used the method to estimate the amount of expansion reached by concrete specimens
cored from a large concrete dam affected by ASR.
161
AAR Facts Book
Site 1
Sites 1,2,3
Site 3
528
MS 3
195
MS 2
133
MS 1
100
200
300
CrCA
OCrCA
Cr+RPCA
CrCP
Cr+RPCP
CAD
RR
RPAV
400
500
600
Damage Rating Index

E (distance between the vertical lines = 1 cm)
F (distance between the vertical lines = 1 cm)
Figure 7.14. A. Highway bridge affected by ASR in Bangor, Maine. Coring sites in the
abutment (not exposed (site 1 - B) and the wing wall (exposed (site 3 - C)). D. DRI results for
cores obtained from sites 1 to 3; the higher the DRI, the higher the degree of ASR damage in
the concrete. E. Polished section from site 1 showing very limited signs of deterioration. F.
Polished section from site 3 showing extensive cracking filled with alkali-silica gel in aggregate
particles and the cement paste.
162
700
AAR Facts Book
Mechanical Testing
Several authors have shown that tensile strength is a much better indicator of the progression of
ASR than compressive strength (Blight et al. 1981; Swamy and Al-Asali 1986; Clayton 1989;
ISE 1992; Siemes and Visser 2000; Smaoui et al. 2006) which is generally affected only at
much higher expansion levels (Figures 7.15 and 7.16). Direct tensile strength (traction tests) is
also more sensitive to ASR than splitting tensile (Brazilian) strength (Clayton 1989; ISE 1992;
Siemes and Visser 2000; Smaoui et al. 2006). The use of the tensile-to-compressive strength
ratio was suggested as a good indication of internal damage due to AAR (Nixon and
Bollinghaus 1985). The tensile-to-compressive strength ratio of sound concrete typically varies
from 0.07 to 0.11. In investigations dealing with AAR, it was suggested that a ratio less than
0.06 was indicative of internal deterioration due to ASR. Since the ultimate compressive
strength of concrete supplied in the field is generally greater than the designed 28-day strength,
a reduction in compressive strength due to AAR is unlikely to be critical in current practice, and
failure through loss in compressive strength is also unlikely (ISE 92; Swamy 1995).
Figure 7.15. Residual mechanical properties for concrete affected by ASR, as percentage of
values obtained at the same age from unaffected concrete; the expansions of the companion test
prisms are also given (Pleau et al. 1989)
163
AAR Facts Book
Figure 7.16. Lower Bound of Residual Mechanical Properties for ASR-Affected Concrete
Specimens, Expressed as a Percentage of Values Obtained for Unaffected Concrete at 28 Days
(ISE 1992)
A number of tests have shown that losses in elastic modulus and flexural strength could lead to
substantial reductions in flexural rigidity and structural stiffness of affected members. Hobbs
(1986) and Smaoui et al. (2006) noted that the modulus of elasticity is more affected than the
compressive and indirect tensile strengths, both of which having been quite similarly affected
by ASR. The rate of reduction in modulus of elasticity can vary according to the type of
reactive aggregate (Smaoui et al. 2006); important reduction in the modulus of elasticity can
occur even at a low level of AAR expansion actually even when compressive strength of the
affected concrete is still increasing (Pleau et al. 1989) (Figure 7.15). According to Clark (1990),
there would be no significant reduction in modulus of elasticity for expansions less than 0.05%.
Losses in modulus of elasticity and flexural strength between 20 and 60% were reported for
expansions ranging from 0.1 to 0.3% (ISE 1992), and could even reach 80% for the former at
very high expansion levels. For the two reactive aggregates tested in their study, Smaoui et al.
(2006) found relatively similar reductions in the modulus of elasticity measured in compression
and in traction. Swamy and Al-Asali (1986, 1988) observed that the dynamic modulus of
elasticity is affected at an early age, even before reaching the 0.1% expansion level.
The Stiffness Damage Test (SDT), which consists in subjecting concrete cores to 5 cycles of
uniaxial loading/unloading up to a maximum of 5.5 MPa (Chrisp et al. 1989) or 10 MPa
(Smaoui et al. 2004) (Figure 7.17A), can be used for assessing the ASR expansion attained to
date through (1) the energy dissipated during the first cycle (hysteresis loop), and (2) the
accumulated plastic strain after the 5 load/unload cycles (Figure 7.17B). Recent research
164
AAR Facts Book
indicates that more accurate results are obtained when the SDT is carried out at a percentage of
the design (28-day) strength (40%) instead of a fixed load (Sanchez et al. 2012). Calibration
curves correlating the above output parameters and expansion obtained on laboratory specimens
can be used for estimating the amount of expansion reached to date by the field concrete
(Smaoui et al. 2004) (Figure 7.17C).
A
Hysteresis loop
(1st cycle)
Total strain
(5 cycles)
Strain (mm/m)
Dissipated energy (joules x 10-3/m3)
C
1.2
1.0
0.8
Quebec
R2=0.98
NM
R2=0.99
Potsdam
R2=1
Texas
R2=0.92
0.6
0.4
0.2
Limeridge
R2=0.83
0.0
0.000 0.080
0.160 0.240
Expansion (%)
0.320
0.400
Figure 7.17. Mechanical testing of concrete cores using the Stiffness Damage Test (SDT). A.
Cores subjected to 5 loading/unloading cycles up to 10 MPa. B. The best output parameters
correspond to the energy dissipated during the first cycle (hysteresis loop) and the accumulated
plastic strain after the 5 cycles. C. Calibration curves for different reactive aggregates (Smaoui
et al. 2004).
165
AAR Facts Book
Similar to the DRI, the SDT can provide a quantitative assessment of the extent of internal
damage in structural concrete members, exposed or not to moisture, either punctually (i.e., at a
specific time - diagnosis) (e.g., Figure 7.18) or as a progressive process when performing the
measurements on a regular basis on cores extracted at the same location (prognosis).
B Core from Site 3
60,00
60,00
50,00
50,00
Load (KN)
Load (KN)
A Core from Site 1
40,00
30,00
20,00
10,00
40,00
30,00
20,00
10,00
0,00
0,00
200
400
600
800
1000
600
800
1000
D
3000
2500
2000
A
B
C
Average
First Cycle hysteresis
1500
1000
500
0
Site 1
Site 3
Core samples
Modulus of elasticity (GPa)
Dissipated Energy (J/m3)
400
Strain (microstrain)
Strain (microstrain)
200
35
30
25
20
15
Modulus of elasticity
A
B
Average
10
5
0
Site 1
Site 2
Core samples
Figure 7.18. Results of Stiffness Damage Testing on cores extracted from the highway bridge
structure illustrated in Figure 7.14. A. SDT result for a core extracted from site 1 (abutment not
exposed). B. SDT result for a core extracted from site 3 (wing wall exposed). C. Comparison of
SDT results, for instance the first cycle hysteresis, for cores extracted from sites 1 (under bridge
deck - less deteriorated) and 3 (exposed to moisture more deteriorated). Higher dissipated
energy values are obtained for ASR-affected concrete cores (site 3). D. Comparison of SDT
results, for instance the average values of the modulus of elasticity for the 2nd and 3rd cycles,
for cores extracted from sites 1 (under bridge deck - less deteriorated) and 3 (exposed to
moisture more deteriorated). Lower modulus of elasticity values are obtained from ASRaffected concrete cores (site 3).
Expansion Tests on Cores
The most commonly used method involves the testing of concrete cores maintained at 38C and
95% R.H. They are expected to provide an estimate of the potential for further expansion of
166
AAR Facts Book
ASR-affected concrete over a relatively short period of time, e.g., six months to one year
(Fournier et al. 2010; Fecteau et al. 2012; Brub et al. 2002b; CSA 2000) (Figure 7.19). After
an initial conditioning period during which the concrete core will reach a volumetric
equilibrium with respect to its new condition (i.e., unrestrained, high temperature and humidity,
additional expansion of preexisting ASR gel due to unrestrained and very humid test
conditions), expansion is measured and tentatively related to a potential for further expansion
due to ASR. However, difficulties are often encountered in the interpretation of the test results
(Brub et al. 2002b), mainly because of: (1) unknown true correlation between free
expansion of cores and the actual expansion in reinforced concrete members; (2) uncertain
correlation of ASR expansion with respect to temperature, which is normally lower in the field;
(3) extreme humidity in the expansion tests; and (4) the possibility that the tested concrete may
be abnormally cracked and porous with respect to the overall field concrete member under
study. A further complication arises from the leaching of alkalis from relatively small
specimens stored at 100% relative humidity; this can lead to an underestimation of the residual
potential for ASR. Immersion tests on cores in 1N NaOH solution at 38C are generally
considered to indicate the absolute degree of reactivity of the aggregates present in the concrete
under study (Brub et al. 2002b).
167
AAR Facts Book
Figure 7.19. Residual expansion testing carried out on core samples (38oC, R.H. > 95%). A.
Expansion measurement of the core. B. Mass measurements are also carried out in parallel to
the expansion measurements to determine the moisture uptake capacity of the core. C. Example
of comparative mass and residual expansion measurements for cores. This allows
differentiating the hydric reequilibration of the concrete at the beginning of the test (due to
moisture uptake only) and the residual expansion potential due to ASR.
Alkali Content of Concrete
The measurement of the available/residual alkali content in concrete can yield interesting
information in assessing whether the concrete tested contains sufficient alkalis to sustain this
reaction (prognosis). The available/ residual alkali content in concrete can be obtained by hot
168
AAR Facts Book
-water extraction or, when possible, using pore solution extraction devices (Fournier et al.
2010; Brub and Tremblay 2004). Over time, significant alkalis can be released from
aggregates and contribute to an increase in the concrete alkali content (Brub et al. 2002c).
Since the alkali content can be subject to considerable variation within a single concrete
element, or between different concrete components of the same structure, separate
determinations should be made on a number of samples taken from different components at
different depths, and representing concrete showing varying severity of deterioration and
subjected to different exposure conditions (rain, sun, buried or underwater portions of the
structures, etc.).
7.2.2.3 Collective Assessment of In-situ and Laboratory Investigations
Fournier et al. (2010) provides a scheme for the analysis of the results obtained from the in-situ
and laboratory investigations for prognostic evaluation of ASR. In summary, the authors
suggest that in the case of reinforced concrete members (e.g., bridges), the potential for further
expansion due to ASR will be expressed by the number of years before the reinforcing steel
yield (in the direction of lower or lack of restraint) could occur, which requires data on the ASR
expansion attained to date and the current expansion rate. In the case of concrete pavements,
the potential for further expansion due to ASR will be expressed by the number of years before
the joints could close, which requires information on the current expansion rate and widths of
joints. The urgency of applying remedial actions will then be partly based on criteria related to
the delay before steel yielding in reinforced concrete members (expansion criteria of > 0.20% is
proposed), or the delay before the closure of expansion joints occur in the case of concrete
pavements. It is recommended that further action be taken when the delay before steel yielding
or joint closure is estimated to be less than 5 years, for example by starting an in-situ
monitoring program of expansion, with measurements at least on a yearly basis, and/or by
performing a structural assessment of the member/structure; it would be appropriate to confirm
an assessment that has been based essentially on expansion tests on cores rather than on in-situ
monitoring.
7.3 CONCLUSION
Unexpected or premature concrete deterioration due to alkali-silica reactivity (ASR) is a
widespread problem worldwide. Routine site inspections performed on a regular basis may
permit identification of the problem; however, ASR in concrete cannot generally be diagnosed
without detailed laboratory and site investigations. Such investigations would include
determination of the distribution and severity of the various defects affecting the concrete
structure, as well as laboratory testing (petrography) of samples collected from the affected
concrete structures. For critical structures such as large dams and fair to large size highway
169
AAR Facts Book
bridges, detailed investigations including a more extensive sampling program might be

necessary to quantify the current condition of the concrete, and to evaluate the potential for
future deterioration (prognosis). Such investigations can involve a detailed sampling program
for further testing in the laboratory and in-situ monitoring of the progress of expansion/
deterioration. The results of the above processes of investigation will then be analyzed to
propose appropriate management actions to be taken for each of the particular applications.
170
AAR Facts Book
REFERENCES
AASHTO. 1993. Standard Method of Test for Rapid Identification of Alkali-Silica Reaction
Products in Concrete. AASHTO T 299, American Association of State and Highway
Transportation Officials, Washington, DC.
AASHTO. 2011. AASHTO guide manual for bridge element inspection, 1st Edition, American
Association of State Highway and Transportation Officials, Washington, DC, 172 p.
ASTM. 2003. Standard Practice for Petrographic Examination of Hardened Concrete. Annual
Book of ASTM Standards, Section Four, Vol. 04.02 Concrete and Aggregates, ASTM C856-02,
ASTM International, 434-450.
Brub, M.A. and Tremblay, C. 2004. Chemistry of pore solution expressed under high
pressure influence of various parameters and comparison with hot-water extraction method.
12th International Conference on AAR in Concrete, Beijing, China, October, Int. Academic
Publishers, Beijing World Publishing Corp., 833-842.
Brub, M.A., Chouinard, D., Pigeon, M., Frenette, J., Rivest, M. and Vzina, D. 2002a.
Effectiveness of Sealers in Counteracting Alkali-Silica Reaction in Highway Median Barriers
Exposed to Wetting and Drying, Freezing and Thawing, and Deicing Salts. Canadian Journal
of Civil Engineering, 29(2): 329-337.
Brub, M.A., Frenette, J. and Rivest, M. 2002b. Laboratory Assessment of the Potential Rate
of ASR Expansion of Field Concrete. Cement, Concrete and Aggregates, 24(1): 28-36.
Blight, G.E., McIver, J.R., Schutte, W.K. and Rimmer, R. 1981. The Effects of AlkaliAggregate Reaction on Reinforced Concrete Structures Made with Witwatersrand Quartzite
Aggregate. Proceedings of the 5th International Conference on Alkali-Aggregate Reaction in
Concrete, S252/15, Cape Town, South Africa.
Brub, M.A., Duchesne, J., Dorion, J.F. and Rivest, M. 2002c. Laboratory Assessment of
Alkali Contribution by Aggregates to Concrete and Application to Concrete Structures Affected
by Alkali-Silica Reactivity. Cement and Concrete Research, 32: 1215-1227.
British Cement Association (BCA). 1992. The Diagnosis of Alkali-Silica Reaction Report of
a Working Party. Wexham Springs, Slough, U.K., 44 p.
CSA. 2000. Guide to the Evaluation and Management of Concrete Structures Affected by
Alkali-Aggregate Reaction. CSA A864-00, Canadian Standards Association, Mississauga,
Ontario, Canada.
171
AAR Facts Book
Chrisp, T.M., Wood, J.G.M., Norris, P. 1989. Towards Quantification of Microstructural

Damage in AAR Deteriorated Concrete. International Conference on Recent Developments
on the Fracture of Concrete and Rocks, Elsevier Applied Science, London, U.K., 419-427.
Clark, L.A. 1990. Structural Aspects of Alkali-Silica Reaction. Structural Engineering
Review, 2(2), June.
Clayton, N. 1989. Structural Performance of ASR Affected Concrete. Proceedings of the 8th
International Conference on Alkali-Aggregate Reaction in Concrete, (Eds. K. Okada, S.
Nishibayashi and M. Kawamura), Kyoto, Japan, 671-676.
Danay, A., Adeghe, L. and Hindy, A. 1993. Diagnosis of the Cause of the Progressive
Concrete Deformations at Saunders Dam. Concrete International, 25-33.
Fecteau, P.L., Fournier, B. and Duchesne, J. 2012. Residual expansion testing : new aspects
on cores extracted from exposure blocks submitted to environmental conditions. 14th
International Conference on AAR in Concrete, 20-25 May, Austin, TX, electronic.
Fournier, B., Brub, M.A., Folliard, K.J. and Thomas, M.D.A. 2010. Report on the diagnosis,
prognosis and mitigation of alkali-silica reaction (ASR) in transportation structures. FHWA
HIF-09-004, Federal Highway Administration.
Francoeur, J., Fournier, B. and Duchesne, J. 2012. tude de la dtrioration du bton en
prsence de granulats incorporant des sulfures de fer. ACI Quebec and Eastern Ontario Chapter
Annual Seminar on Progress in Concrete Technology, November, Boucherville, Qubec.
Gaudreault, M. 2000. The St. Lawrence Seaway (Quebec, Canada): A Case Study in the
Management of Structures Affected by Alkali-Aggregate Reaction. Proceedings of the 11th
International Conference on Alkali-Aggregate Reaction in Concrete, (Eds. M.A. Brub, B.
Fournier and B. Durand), Quebec City, Canada, 1293-1302.
Grattan-Bellew, P.E. 1992. Comparison of Laboratory and Field Evaluation of Alkali-Silica
Reaction in Large Dams. First International Conference on Concrete AAR in Hydroelectric
Plants and Dams, September-October, Fredericton, NB, Canada, 23 p.
Grattan-Bellew, P.E. and Mitchell, L.D. 2006. Quantitative petrographic analysis of concrete
The Damage Rating Index (DRI) method, a review. Marc-Andr Brub symposium on AAR
in concrete, CANMET/ACI Advances in concrete technology, Montral, Canada, 321-334.
Guthrie, G.G., Jr. and Carey, J.W. 1997. A Simple Environmentally Friendly, and Chemically
Specific Method for the Identification and Evaluation of the Alkali-Silica Reaction. Cement
and Concrete Research, 27(9): 1407-1417.
172
AAR Facts Book
Hobbs, D.W. 1986. Alkali-Silica Reaction in Concrete. The Structural Engineer, 64A(12):
381-383.
Institution of Structural Engineers (ISE). 1992. Structural Effects of Alkali-Silica Reaction
Technical Guidance Appraisal of Existing Structures. Institution of Structural Engineers,
London, 45 p.
Jensen, V. 2004. Measurements of cracks, relative humidity and effects of surface treatments
on concrete structures damaged by alkali-silica reaction. 12th International conference on
AAR in Concrete, Beijing, China, October, Int. Academic Publishers, Beijing World Publishing
Corp., 1245-1253.
Laboratoire Central des Ponts et Chausses (LCPC). 1997. Dtermination de lindice de
fissuration dun parement de bton. Mthode no. 47, Ministre de lquipement, des transports
et du logement, Paris.
Laboratoire Central des Ponts et Chausses (LCPC). 1999. Manuel didentification des
ractions de dgradation interne du bton dans les ouvrages dart. Ministre de lquipement,
des transports et du logement, Paris.
Laboratoire Central des Ponts et Chausses (LCPC). 2003. Aide la gestion des ouvrages
atteints de ractions de gonflement interne. Ministre de lquipement, des transports et du
logement, Paris.
Laboratoire Central des Ponts et Chausses (LCPC). 2009. Mthodes de suivi dimensionnel et
de suivi de la fissuration des structures. Ministre de lquipement, des transports et du
logement, Paris.
Moradi-Marani, F., Kodjo, S.A., Rivard, P. and Lamarche, C.P. 2011. Application of the
mechanical perturbation produced by traffic as a new approach of nonlinear acoustic technique
for detecting microcracks in concrete : a laboratory simulation. Conference on Review of
progress in quantitative nondestructive evaluation (QNDE), Burlington, USA, 17-22 July.
Natesaiyer, K., Stark, D. and Hover, K.C. 1991. Gel Fluorescence Reveals Reaction Product
Traces. Concrete International, January, 25-28.
Nixon, P.J. and Bollinghaus, R. 1985. The effect of Alkali-Aggregate Reaction on the Tensile
and Compressive Strength of Concrete. Durability of Building Materials, 2: 243-248.
Pleau, R., Brub, M. A., Pigeon, M., Fournier, B. and Raphal, S. 1989. Mechanical Behavior
of Concrete Affected by AAR. 8th International Conference on AAR in Concrete. Kyoto,
Japan, 721-726.
173
AAR Facts Book
Rivard, P., Fournier, B. and Ballivy, G. 2000. Quantitative Petrographic Technique for
Concrete Damage Due to ASR: Experimental and Application. Cement, Concrete and
St-John, D.A., Poole, A.W. and Sims, I. 1998. Concrete Petrography: A handbook of
investigative techniques. Arnold, London, 474 p.
Sanchez, L., Fournier, B. and Jolin, M. 2012. Study of the parameters of the stiffness damage
test for assessing concrete damage due to alkali-silica reaction. Proceedings of the 14th
International Conference on AAR in Concrete, Austin,Texas, 20-25 May, electronic.
Sargolzahi, M., Kodjo, S.A., Rivard, P. and Rhazi, J. 2010. Effectiveness of nondestructive
testing for the evaluation of ASR in concrete. Construction & Building Materials, 24(8): 1398
1403.
Siemes, T. and Gulikers, J. 2000. Monitoring of Reinforced Concrete Structures Affected by
Alkali-Silica Reaction. Proceedings of the 11th International Conference on Alkali-Aggregate
Reaction in Concrete (Eds. M.A. Brub, B. Fournier and B. Durand), Quebec City, Canada,
1205-1214.
Siemes, T. and Visser, J. 2000. Low Tensile Strength in Older Concrete Structures with Alkali
-Silica Reaction. Proceedings of the 11th International Conference on Alkali-Aggregate
Reaction in Concrete (Eds. M.A. Brub, B. Fournier and B. Durand), Quebec City, Canada,
1029-1038.
Sims, I. and Nixon, P.J. 2010. Outline guide to the use of RILEM methods in assessments of
alkali-reactivity potential. RILEM Recommended Test method AAR-0. RILEM committee
documentation, September.
Smaoui, N., Bissonnette, B., Brub, M.A., Fournier, B. and Durand, B. 2006. Mechanical
Properties of ASR-Affected Concrete Containing Fine or Coarse Reactive Aggregates. Journal
of ASTM International, 3(2).
Smaoui, N., Brub, M.A., Fournier, B., Bissonnette, B. and Durand, B. 2004. Evaluation of
the Expansion Attained to Date by ASR-Affected Concrete Part I: Experimental Study.
Canadian Journal of Civil Engineering, 31: 826-845.
Stark, D. 1991. Handbook for the Identification of Alkali-Silica Reactivity in Highway
Structures. SHRP-C-315, TRB National Research Council, 49 p.
Swamy, R.N. and Al-Asali, M.M. 1986. Influence of Alkali-Silica Reaction on the Engineering
Properties of Concrete. (Ed. V.H. Dodson), ASTM STP 930, American Society for Testing and
Materials, 69-86.
174
AAR Facts Book
Swamy, R.N. and Al-Asali, M.M. 1988. Engineering Properties of Concrete Affected by
Alkali-Silica Reaction. ACI Materials Journal, 85: 367-374.
Swamy, R.N. 1995. Effects of Alkali-Aggregate reactivity on Material Stability and Structural
Integrity. Proceedings of the CANMET/ACI International Workshop on Alkali-Aggregate
Reactions in Concrete, Dartmouth, Nova Scotia, 1-4 October, 293-309.
Tajari, M., Shekarchi, M. and Sadri, A. 2011. Use of Impact-echo Technique for Detection of
Distributed Damage in Concrete due to ASR. Material Evaluation, 69(7): 881-890.
Thomas, M.D.A., Fournier, B., Folliard, K.J. and Resendez, B. 2011. Alkali-Silica Reactivity
Field Identification Handbook, FHWA-HIF-12-022, FHWA, 80 p.
Thomas, M.D.A., Fournier, B., Folliard, K.J. and Resendez, B. 2012a. Alkali-Silica Reactivity
Surveying and Tracking Guidelines, July, 32 p.
Thomas, M.D.A., Folliard, K.J., Fournier, B., Drimalas, T. and Rivard, P. 2012b. Study of
remedial actions on highway structures affected by ASR. 14th International Conference on
AAR in Concrete, Austin, Texas, 20-25 May, electronic.
Thompson, G.A., Charlwood, R.G., Steele, R.R. and Coulson, D.M. 1995. Rehabilitation
Program Mactaquac Generating Station, NB. Proceedings of the CANMET/ACI
International Workshop on Alkali-Aggregate Reactions in Concrete, Dartmouth, Nova Scotia, 1
-4 October, 355-368.
Villeneuve, V., Fournier, B. and Duchesne, J. 2012. Determination of the damage in concrete
affected by ASR the Damage rating Index (DRI). 14th International Conference on AAR in
Concrete, Austin, Texas, 20-25 May, electronic.
Walker, H.N., Lane, D.S. and Stutzman, P.E. 2006. Petrographic Methods of Examining
Hardened Concrete: A Petrographic Manual (Revised 2004). FHWA-HRT-04-150, Federal
Highway Administration.
175
AAR Facts Book
176
AAR Facts Book
Chapter 8 Mitigation Methods
8 Mitigation Methods for ASR-Affected Structures

8.1 INTRODUCTION
This chapter describes various mitigation measures that are available for ASR-affected
structures, identifies options that have been used the most, discusses those whose effectiveness
has been proven in the laboratory and field, and describes those that remain experimental in
nature due to lack of data/information proving their merit in real-world applications.
The main objective of this chapter is to provide guidance on means of extending the service life
of ASR-affected structures. The term mitigation is used in lieu of repair because the
methods described herein are generally not able to, nor are they intended to, repair or restore the
original properties or integrity to the ASR-affected structure. Rather, the intention is to reduce
future expansion of the structure or to lessen the detrimental impact of future expansion.
The majority of the work to date on treating existing structures has focused on ASR-affected
structures, as opposed to ACR-affected structures, and there are by far many more ASRaffected structures worldwide. As such, the focus of this portion of this protocol is aimed at
ASR-affected structures. However, some of the mitigation measures, particularly those aimed at
drying the concrete, would be helpful whether it is ASR or ACR that is impacting the structure.
8.2 OVERVIEW OF MITIGATION METHODS

Figure 8.1 summarizes the various mitigation options that have been applied to field structures
affected by ASR. This section will briefly discuss each of the options shown in Figure 8.1 and
will then focus on those that have the greatest potential for effectively treating ASR-affected
structures. For each of these options, the merits will be discussed, as well as inherent
shortcomings, both in terms of general applicability to field structures and specific application
to certain structures.
Improved drainage
Application of coatings/sealers
Application of cladding
Crack filling
Application of lithium compounds
Application of restraint (FRP, etc.)
Saw cutting/slot cutting
Figure 8.1. Mitigation Methods for ASR-Affected Concrete Structures

177
AAR Facts Book
The first four methods highlighted in Figure 8.1 are all aimed at reducing the relative humidity
in concrete, and for convenience, these methods will be discussed together in section 8.3.
Section 8.4 will describe various methods of applying lithium-based compounds to ASRaffected structures. Lastly, methods aimed at restraining ASR-induced expansion and relieving
ASR-induced stress are briefly discussed in sections 8.5 and 8.6, respectively.
The Federal Highway Administration (FHWA) established the ASR Development and
Deployment Program (under FHWA Contract DTFH61-06-D-00035) in 2006 to provide
technical assistance to States and practitioners to address ASR issues. As part of these efforts,
field trials were initiated that included the implementation of various mitigation measures in
highway infrastructure, including bridges, barrier walls, and pavements. Figure 8.1 provides an
overview of the FHWA-funded field trials, including information on the various mitigation
measures (Ahlstrom 2012). Discussion on several of these field trials is included in this chapter,
but it should be kept in mind that the key findings from these studies may not be known for
several years, and certainly beyond the conclusion of the FHWA project in 2013.
Table 8.1. FHWA-Funded Field Trials (adapted from Ahlstrom 2012)
State
Structure
Mitigation Technique
Alabama
Historic bridge
arch
Silane sealer
Arkansas
Pavement
Silane sealer
Delaware
Pavement
Topical application of lithium nitrate
Delaware
Pavement
Monitoring an asphalt overlay of pavement with

lithium nitrate
Hawaii
Aggregates
Testing aggregates and development of field

exposure site
Massachusetts
Median barrier
Silane sealers; topical application of lithium nitrate
Massachusetts
Aggregates
Testing aggregates and development of field

exposure site
Maine
Bridge abutments
and piers
FRP wrap; silane sealer; electrochemical application

of lithium nitrate
Texas
Bridge Columns
Electrochemical application of lithium nitrate;

vacuum impregnation of lithium; silane sealers
Texas
Precast Bridge
Girders
Aggregate testing and investigation of specific

mixture designs
Vermont
Bridge barrier
walls
Silane sealers
178
AAR Facts Book
8.3 REDUCING INTERNAL RELATIVE HUMIDITY

The availability of moisture is critical for ASR to cause significant expansion and cracking. In
the laboratory, Pedneault (1996) showed that below a relative humidity (RH) of 80 percent,
ASR-induced expansion is significantly reduced or suppressed, as shown in Figure 8.2. The
data in Figure 8.2 are for concrete prisms composed of various reactive aggregates and stored at
different relative humidities. Although the effects of moisture on expansion did vary from
aggregate to aggregate, reducing the relative humidity to 80 percent greatly reduced the
observed expansions after two years of testing for all aggregates in the study.
Figure 8.2. Influence of Relative Humidity on Expansion due to ASR (after Pedneault 1998)
The effects of moisture availability on ASR are also quite evident in the field. It is quite
common for part of a structure that is directly exposed to water (e.g., rain water) to exhibit signs
of ASR-induced expansion and cracking while part of the same structure that is protected from
direct contact with water will show very little symptoms of ASR. Figure 8.3 illustrates the
importance of moisture availability on ASR-induced cracking the bent cap, from a bridge in
Houston, Texas, only exhibits visual cracking on the exterior faces that are subject to direct
rainfall. The portions of the bent cap well below the bridge deck, and thus protected from direct
rainfall, show only negligible signs of distress. This is typically the case with precast bridge
girders only the exterior girders tend to show signs of ASR-induced expansion and cracking,
whereas interior girders that are shielded from rain often show negligible symptoms of ASR.
179
AAR Facts Book
Figure 8.3. Photograph of bridge in Houston, Texas, illustrating effects of moisture availability
on ASR-induced expansion and cracking. Only the exposed face of the bent cap (left side of
photo) shows visual cracking, whereas the portions of the bent cap protected by the bridge deck
show very little cracking.
Figure 8.4 shows a photograph of a bridge structure included as part of the FHWA ASR
Development and Deployment Program. Prior to applying various sealers/coatings to the
bridge, cores were extracted from Location A (under the bridge deck, sheltered from rain) and
Location B (exposed directly to precipitation). Petrographic evaluations of the cores showed
that the concrete mixture proportions were identical, meaning the same concrete was used in
each location of the bridge. Using the Damage Rating Index (DRI), a quantitative petrographic
technique described in chapter 7, it was found that the DRI of a core extracted from Location A
was 528, whereas the DRI of a core extracted from Location B was only 133 (note the higher
the DRI, the greater the extent of ASR). Thus, the effects of exposure conditions on ASR are
evident both to the naked eye, as shown in Figure 8.3, and under a petrographers microscope,
as evidenced by the elevated DRI values for cores extracted from portions of the bridge directly
exposed to precipitation in Figure 8.4.
A great deal can be learned from the examples shown in Figures 8.3 and 8.4, where there is a
clear link between moisture availability and ASR-induced expansion and cracking. Examples of
this type are ubiquitous in highway infrastructure, where the influence of moisture availability
on ASR is evident not only in bridges, but also in pavements, where joints tend to deteriorate
first due to the increased availability of water at or near joints. These observations can be
synthesized and applied to mitigation measures for ASR-affected concrete whenever possible,
one should try to limit the availability of moisture. This can involve improving drainage for a
given structure, for example, by diverting drainage from a bridge deck away from an ASRaffected column or cap. Likewise, improving the drainage conditions for an ASR-affected
pavement, when possible, can help to reduce future expansion and cracking.
180
AAR Facts Book
Figure 8.4. Photograph of bridge in Bangor, Maine. Location A is sheltered by the bridge deck,
whereas Location B is exposed directly to precipitation. Sampling of concrete cores from each
location showed strong correlation between moisture availability and ASR-induced expansion
and cracking.
It is not always a feasible option to simply improve drainage for a given ASR-affected
structure, and other, more advanced or aggressive measures may be warranted. For example, the
application of exterior cladding that prevents the ingress of additional moisture may be
beneficial, but it should be noted that the moisture already present within the concrete may be
sufficient for ASR to remain active, and this fact must be considered when contemplating a
cladding as a mitigation measure. The use of ventilated cladding slowed down the progression
of ASR in a race course grandstand by minimizing access to external moisture while still
allowing the concrete to dry out (Hobbs 1989).
For instances where it is not possible to adequately dry concrete out through improved drainage
or where it is not feasible (technically or economically) to apply ventilated cladding, the use of
coatings or sealers to reduce the internal relative humidity should be considered. A coating or
penetrating sealer that will trigger this reduction in internal relative humidity must provide the
following characteristics (after CSA 2000):
1.
2.
3.
4.
5.
6.
be resistant to water absorption;

penetrate to a measurable depth;
resist deterioration from ultraviolet (UV) radiation;
possess long-term stability in an alkaline environment;
be of long-term stability in an alkaline environment; and
allow vapor transmission
There are a range of products, some proprietary in nature, that satisfy the above characteristics,
but silicone-based products have been used most commonly as mitigation measures for not
181
AAR Facts Book
only ASR, but also to help reduce the ingress of water (to enhance frost resistance) and
external chlorides (to reduce the rate of corrosion of reinforcing steel). Silicones exhibit several
characteristics that make them well-suited for such applications, including (after Mayer 1998):
strong repellence to liquid water,

chemical inertness in most cases,
stability in a wide range of temperatures, and
excellent insulative properties.
Members of the silicone family that have been used the most for such applications are silanes,
siloxanes, and silicone resins, with silanes being the most commonly used, as well as the most
commonly evaluated with regard to ASR. Because of the relevance of silanes to ASR research
and implementation, discussion follows on the mechanisms by which silanes are able to reduce
the internal relative humidity.
When silanes are applied to the concrete surface, a chemical reaction known as crosslinking or
polycondensation occurs within the concrete substrate or at the surface to form a silicone resin
network, which is responsible for water repellency, permeability to water vapor, and durability
(Mayer 1998). Silanes achieve their unique properties by chemically bonding a water-repellent
hydrocarbon molecule to the substrate, as illustrated in Figure 8.5. As a result, the critical
surface tension of the concrete substrate is reduced, and if the surface tension of the substrate
falls below that for water, it will be water repellent (McGettigan 1992).
Figure 8.5. Mechanism by which Silanes React with Concrete (adapted from Dow Corning
2005)
There are a variety of silane products available, varying primarily based on the concentration of
silane in the specific formulation (ranging from 20 percent to close to 100 percent) and based
on the type of carrier with which the silane is combined (either water-based or solvent-based).
More stringent restrictions regarding VOC emissions have resulted in more water-based silanes
182
AAR Facts Book
or solvent-based silanes with higher silane contents (and thus lower solvent content and reduced
VOCs).
Silanes are almost always applied topically (see Figure 8.6), at a coverage rate typically in the
range of 2.4 4.8 m2/L (or 100 200 ft2/gal). The depth of penetration will vary from product
to product and based on environmental conditions. Lute (2008) reported depths of penetration
of silanes into concrete mixtures typically used in highway applications to be in the range of 2
to 6 mm. This depth of penetration is sufficient to form a functional barrier, preventing water
from entering but allowing moisture vapor to escape.
Figure 8.6. Photograph Showing Topical Application of 40%-Silane Solution (solvent-based) to

ASR-Affected Highway Barrier in Massachusetts (part of FHWA Project DTFH61-02-C
00097)
There have been several studies that have confirmed the benefits of applying silanes to field
structures to reduce future ASR-induced expansion. The research by Brub et al. (2002) was
particularly encouraging as it showed that applying silane to highway barriers heavily damaged
by ASR resulted in a dramatic reduction in cracking (Figure 8.7), as well as future expansion
(Figure 8.8).
183
AAR Facts Book
Control
Silane-treated
Figure 8.7. Reduction in Cracking of Highway Barriers in Canada, after Application of Silane
(Brub et al. 2002)
Figure 8.8. Reduction in Expansion of Highway Barriers in Canada, after Application of Silane
(Brub et al. 2002)
184
AAR Facts Book
Badly cracked concrete piers supporting the Hanshin Expressway in Japan were repaired at an
age of 7 years by first filling the cracks with an epoxy resin injected under pressure and then
either coating with an epoxy resin or impregnating with silane followed by a cosmetic coating
of a polymer cement paste (Kojima et al. 1992), as shown in Figure 8.9. Of all the treatments
shown in Figure 8.9, the columns treated with silane performed the best, and those columns
treated with non-breathable coatings, such as epoxy and acrylic, performed the worst as
moisture vapor was not allowed to escape the concrete.
Figure 8.9. Reduction in Expansion of Bridge Columns (Hanshin Expressway, Kobe, Japan)
Triggered by Use of Silane Coating (Kojima et al. 1992)
Several FHWA-funded field trials over the past few years (under FHWA Project DTFH61-06-D
-00035) have included the topical application of silane-based products, as detailed in Table 8.1.
Silanes were topically applied to highway barriers, pavements, and bridges. Most of the ASRaffected elements treated by silane-based products were treated quite recently (2010-2012), and
as such, it is too premature to make conclusions regarding the efficacy of the various coatings
and sealers in reducing the relative humidity inside concrete and reducing future ASR-induced
expansion.
Fortunately, highway barriers treated under a previous FHWA project (DTFH61-02-C-00097)
focusing on implementation of lithium technologies have been included in the latest FHWA
project (DTFH61-06-D-00035), and as such, the barriers were included in the ongoing
monitoring efforts and have been monitored for a total of about five years. Figure 8.10 shows
the significant reduction in cracking of highway barriers in Massachusetts, approximately three
years after the topical application of various silane-based products. Figure 8.11 shows the
expansion data for the barriers, confirming that all of the silane products studied reduced
expansion, compared to the control. In fact, the barriers that were most affected by silane
185
AAR Facts Book
applications underwent shrinkage, which is consistent with the work done in Canada on
highway barriers (see Figures 8.7 and 8.8).
Silane
Control
Figure 8.10. Reduction in Cracking of Highway Barriers in Massachusetts,
after Application of Silane (FHWA Project DTFH61-06-D-00035)
Figure 8.11. Effects of Silane Treatment on the Expansion of Highway Barriers in

Leominster, MA (note Silanes A and B are water-based. Silane C is solvent-based)
Based on the experience with highway barriers, silanes can be quite effective in reducing
internal relative humidity, expansion, and cracking. However, this is somewhat of an ideal
situation for applying silanes the barriers are relatively thin and are treated on both sides,
thereby allowing for the relative humidity to be reduced from both sides of the wall. There are
186
AAR Facts Book
certainly other applications where silanes or products that work in a similar manner may not be
effective. For pavements, slabs on grade, wingwalls, or other applications where moisture is
available from below (or beneath), silanes will likely not be as effective (or effective at all) as
their benefits are only realized from the treated surface. To quantify the effects of treating such
elements with silane, FHWA-funded field trials have in recent years included the treatment of
wingwalls in Maine and Rhode Island and pavements in Arkansas; it is too early to determine
what benefits, if any, will be derived from such treatments.
Lastly, it should be noted that the application of silanes will not be effective in concrete with
large crack widths. For these larger cracks, flexible caulking or similar products should be used
to seal the larger cracks. In a FHWA field trial in Wetumpka, Alabama, flexible caulking was
applied to selected cracks in a bridge that were deemed to be too wide to be effectively treated
by silanes. There have been other recent developments, including the use of high-build paints
or elastomeric coatings that may show promise in bridging larger cracks and avoiding the need
for caulking of individual cracks. The need to seal larger cracks becomes critical when
reinforced concrete is exposed to external chlorides or in regions exposed to cycles of freezing
and thawing. These elastomeric coatings have recently been applied to bridge structures in
Maine and Vermont and highway barriers in Massachusetts as part of ongoing FHWA field
trials; again, it is too premature to make any conclusions about the efficacy of such applications,
but monitoring is ongoing.
8.4 APPLICATION OF LITHIUM COMPOUNDS

There have been several laboratory-based publications related to using lithium compounds to
treat concrete already suffering from ASR-induced expansion. Research by Stark et al. (1993),
Stokes et al. (2000), and Barborak et al. (2004) have shown that lithium compounds can reduce
future expansion of small, ASR-affected concrete specimens in accelerated laboratory tests.
Figure 8.12 shows that immersing concrete prisms in lithium nitrate reduced the future
expansion of these prisms when subjected to storage conditions similar to ASTM C 1293
(Stokes et al. 2000). Although one cycle of lithium application slightly reduced expansion,
multiple cyclical applications of lithium nitrate significantly reduced long-term expansion. It
should also be noted that prisms that were treated by multiple cycles of immersion in water, as
opposed to lithium nitrate, exhibited reduced expansion as well, presumably due to the leaching
of alkalis from the prisms to the water during the immersion period.
187
AAR Facts Book
Figure 8.12. Effects of Submerging Concrete Prisms in Lithium Nitrate (or water) on
Expansion of Prisms in ASTM C 1293 Environment (after Stokes et al. 2000)
Although the mechanisms by which immersing concrete or mortar specimens in lithium

compounds may reduce future expansion are not fully understood, it is generally believed that it
changes the nature and behavior of the gel from expansive to essentially non-expansive.
Because of the positive results in laboratory-based work and driven by the need for viable
mitigation measures for ASR-affected structures and pavements, there has been considerable
interest in treating ASR-affected field structures, especially in recent years under FHWAfunded research. A detailed review of past field trials using lithium compounds can be found in
Folliard et al. (2006), and several field trials are still being monitored under current FHWA
projects. The most common method of applying lithium compounds in field trials has been via
topical application, primarily for pavements (Figure 8.13), highway barriers (Figure 8.14A),
and bridge decks. There have also been a handful of field trials where lithium was applied either
by vacuum (Figure 8.14B) or through electrochemical means, both aimed at increasing the
depth of penetration of lithium.
188
AAR Facts Book
Figure 8.13. Photograph Showing Topical Application of 30%-LiNO3 Solution to
Concrete Pavement in Idaho
Figure 8.14. Photographs Showing Topical Application (A) and Vacuum Application (B) of
30%-LiNO3 Solution to Highway Barrier in Massachusetts
Although lithium compounds have been found to be effective in laboratory-based research,
which has focused on treating small specimens affected by ASR, there is, unfortunately, very
little, if any, documentation that lithium is effective in reducing ASR-induced expansion in
actual structures in the field. Part of this is due to the general lack of monitoring of field trials in
which lithium compounds have been applied to structures (or pavements). Recognizing the need
to obtain such critical data, field trials were initiated under FHWA funding (under FHWA
Project DTFH61-02-C-00097) in Idaho, Massachusetts, and Texas. The primary focus of these
studies was primarily on the topical application of lithium nitrate, but the applications of lithium
nitrate by vaccum and by electrochemical methods were also evaluated.
189
AAR Facts Book
Table 8.2 summarizes the lithium penetration data for field trials in three states (after Folliard
2008). When evaluating lithium profile data, it is important to view it in context of how much
lithium is needed to control expansion. Consider the following example: Assuming a plain
concrete mixture contains 3 kg/m3 Na2Oe, a standard dose of 30% lithium nitrate solution (4.6
L of solution per 1 kg of Na2Oe) would amount to 13.8 L/m3 of lithium nitrate solution. This
dosage of LiNO3 solution contains 504 g of Li, and assuming that half of this lithium gets bound
in early hydration products (as well documented in literature), approximately 250 g of lithium
per m3 will remain in the concrete to combat ASR. If one assumes a concrete density of 2350
kg/m3 the concentration of lithium ions in pore solution would be approximately 100 ppm, and
this concentration would be enough to control expansion when used as an admixture. However,
it should be noted that the standard dose of 30% lithium nitrate solution (4.6 L of solution per
1 kg of Na2Oe) has been reported to be sufficient for controlling expansion in about half of the
aggregates recently tested in North America (Tremblay et al. 2008). Some aggregates,
especially greywackes, still expand considerably at over twice the standard dose. For the
purpose of this discussion, it is assumed that the aggregate in the concrete being treated
responds favorably to lithium (4.6 L of solution per 1 kg of Na2Oe) and that 100 ppm would be
required in pore solution for suppression of excessive expansion.
Table 8.2. Summary of Lithium Penetration Data for Highway Applications of Lithium Nitrate
Following the above discussion and assuming a requisite lithium threshold of 100 ppm, the data
in Table 8.2 was not very positive with regard to topical or vacuum applications of lithium
nitrate. In fact, depths of penetration for topical applications in an Idaho pavement were found
to be only a few millimeters, with dosages of lithium necessary to suppress expansion measured
only down to the first 2 to 3 mm, even after three treatments in heavily cracked pavements.
190
AAR Facts Book
These penetration results are consistent with laboratory evaluations performed under the same
project, and when taken as a whole, it appears that due to an inherent lack of penetration, the
topical application of lithium compounds shows little, if any, promise of mitigating ASR in
structures and pavements.
Figure 8.15 shows that none of the topical lithium applications reduced expansion in the
highway barriers treated in Leominster, MA, unless the lithium treatment was followed up with
a subsequent application of 40% silane. Given that multiple treatments of lithium nitrate did not
reduce expansions, it is logical to conclude that the reduction in expansion can be attributed
solely to the benefits of silane in reducing the internal relative humidity of the concrete. Similar
results were obtained for the barriers in which a vacuum was applied to try to drive lithium
nitrate in further. Reductions in expansion were only observed when the vacuum treatment was
followed up with an application of 40% silane.
Figure 8.15. Effects of Topical Lithium Nitrate Application on Expansion. Note that only the
barriers treated with silane reduced long-term expansion.
Because of the documented lack of penetration in field and laboratory trials in which lithium
compounds have been applied topically, recent focus has shifted towards more aggressive
means of driving lithium into ASR-affected concrete, specifically through vacuum
impregnation and electrochemical methods. Unfortunately, in research performed under FHWA
Project DTFH61-02-C-00097, vacuum impregnation was not found to be effective in the
laboratory or in field structures in Texas and Massachusetts. For example, for ASR-affected
bridge columns in which lithium nitrate was applied via vacuum, the depths of lithium
191
AAR Facts Book
penetration were found only to be present in the outer 9 to 12 mm, drawing into question
whether such an elaborative and expensive vacuuming technique is justified. Substantially
higher depths of penetration were observed in the same study when lithium nitrate was
electrochemically driven into bridge columns, with dosages sufficient to reduce ASR measured
all the way down to the reinforcing steel (50 mm from outer surface). Information on the
specific details of the electrochemical method used for this bridge structure can be found in East
(2007). However, one major concern with this technique is that the electrochemical process,
itself, tends to drive alkalis already present in the concrete towards the rebar, which may be a
significant obstacle to this technology. It appears that as a whole, this technique is quite
powerful in driving external lithium into the concrete, but the rearrangement of internal alkalis
and accumulation of sodium and potassium (which in turns leads to an augmented pH near the
rebar) is a serious concern that deserves further attention. More work is in progress to evaluate
this treatment technique and to quantify the benefits (and downsides) of this approach.
Despite the general lack of penetration observed in laboratory and field structures in which
lithium was applied topically or by vacuum, it is hoped that data will be generated from other
field trials, thereby increasing the state of knowledge and expanding the database of depth of
penetration data. Lastly, the success in driving lithium all the way to the reinforcing steel is
encouraging, but the adverse effects of pushing sodium and potassium to the vicinity around the
steel deserve further attention. Given that lithium compounds have clearly been shown to be
effective in reducing future expansion in ASR-affected concrete in the laboratory, and given
that the options for treating the cause of ASR in the field are limited, it is hoped that additional
lithium-based field trials will be conducted and monitored, thereby helping to quantify the
effects, if any, of lithium application on remaining service life.
8.5 APPLICATION OF EXTERNAL RESTRAINT

Numerous studies and field trials have shown that physical restraint or confinement (e.g.,
encapsulation of the affected member by a surrounding non-reactive concrete, applied stress or
reinforcement) can significantly reduce deleterious expansion due to ASR in the direction of
restraint (Fournier et al. 2004). Because of the unique nature of this mitigation approach and the
fact that the structural response is impacted, it is imperative that a structural engineer play the
leading role in specifically designing the methodology for a given ASR-affected structure.
Post tensioning in one or two dimensions, or by encasement in conventional reinforced
concrete, is currently used as a means to restore the integrity of the structure; however, it should
generally be restricted to relatively small masses of structural concrete because of the huge
forces that may result from the expansive process due to ASR (Rotter 1995; CSA 2000). Posttensioned tendons or cables are considered to be an effective solution for thin arch dams
192
AAR Facts Book
(Singhal and Nuss 1991) or structural members of bridge/highway structures; however, they
may be less attractive for large concrete structures because of the necessity of periodic
destressing (Rotter 1995). Methods to restrain expansion and movement in mass concrete
foundations such as tower bases have also included rock anchors and/or encapsulation (Brub
et al. 1989).
Strengthening by introducing reinforcement with straps, steel plates, and tensioning through
bolts was also found to be effective in providing containment for selected ASR-affected
concrete members (Wood and Angus 1995). Wrapping ASR-affected reinforced concrete
columns with composite materials (FRP) has also been applied in field structures (Figure 8.16)
and more recently in Vermont as part of ongoing FHWA field trials (under FHWA Project
DTFH61-06-D-00035).
Figure 8.16. Photograph Showing Topical Bridge Column Wrapped with FRP (Carse 1996)
8.6 STRESS RELIEF

For certain applications, such as a pavement suffering from ASR-induced expansion, a viable
option to extend the service life is to remove sections of concrete near the joints by saw cutting.
Removing these sections is helpful in eliminating joint-related failures and minimizing ride
quality issues. The sections that have been removed can be replaced by sound concrete, with
careful attention paid to restoring the intended joint details (opening, dowel bars, etc.). This
approach has been done on a much larger scale for concrete dams, where large slots have been
cut to accommodate future expansion. It should be noted that this approach (saw cutting/joint
cutting) only relieves stresses but does nothing to address the root cause of the expansion. It is
common for this method to be performed repeatedly as expansion continues and negates the
benefits achieved from the previous concrete removal.
193
AAR Facts Book
8.7 SUMMARY
This section described a variety of methods that have been applied to field structures suffering
from ASR-affected expansion and cracking. Some methods, such as the application of silanes,
have shown significant promise, especially when applied to elements such as small columns
and highway barriers, whereas other methods, such as the topical application of compounds,
have shown little or no promise in reducing ASR-induced expansion and cracking. It is hoped
that ongoing FHWA field trials (under FHWA Project DTFH61-06-D-00035) will better
quantify the potential benefits of these and other preventive measures and will help to
contribute to improved management of ASR-affected structures in the future.
194
AAR Facts Book
REFERENCES
Ahlstrom, G. 2012. The United States Federal Highway Administrations Alkali-Silica
Reactivity Development and Deployment Program. Proceedings of the 14th International
Conference on Alkali-Aggregate Reactivity (ICAAR), Austin, TX, USA, 20-25 May.
Barborak, R., Folliard, K.J. and Thomas, M.D.A. 2004. Using Lithium Compounds to Treat
Hardened Concrete Suffering from ASR: Preliminary Laboratory Results. Proceedings of the
12th International Conference on Alkali-Aggregate Reactivity (ICAAR), Beijing, China, 483
489.
Brub, M.A., Fournier, B. and Frenette, J. 1989. Dtrioration de fondation de pylnes
d'ancrage de lignes de transport d'lectricit par des ractions alcalis-granulats, performance
mcanique et rparation du bton. Canadian Journal of Civil Engineering, 16(6): 945-959.
Brub, M.A., Chouinard, D., Pigeon, M., Frenette, J., Rivest, M. and Vzina, D. 2002.
Effectiveness of Sealers in Counteracting Alkali-Silica Reaction in Highway Median Barriers
Exposed to Wetting and Drying, Freezing and Thawing, and Deicing Salts. Canadian Journal
of Civil Engineering, 29(2): 329-337.
Canadian Standards Association (CSA). 2000. Guide to the Evaluation and Management of
Concrete Structures Affected by Alkali-Aggregate Reactions. CSA A864-00, 2000b, Ontario,
Canada.
Carse, A. 1996. The Asset Management of Alkali-Silica Rection in a Long Bridge Structure.
Proceedings of the 11th International Conference on Alkali-Aggregate Reaction in Concrete
(Eds. M.A. Brub, B. Fournier and B. Durand), Quebec City, Canada, 1025-1032.
Dow Corning. 2005. "A Guide to Silane Solutions." Dow Corning Corporation, http://
www.dowcorning.com/content/publishedlit/SILANE-GUIDE.pdf, (July 24, 2008).
East, B. 2007. Laboratory and Field Investigations on the use of Lithium Nitrate to Prevent or
Mitigate Alkali-Silica Reaction. M.S. Thesis, The University of Texas at Austin.
Folliard, K.J., Thomas, M.D.A., Ideker, J.H., East, B.L. and Fournier, B. 2008. Case Studies of
ASR-Affected Structures with Lithium Nitrate. Proceedings of the 13th International
Conference on Alkali-Aggregate Reaction (Eds. M.A.T.M. Broekmans and B.J. Wigum),
Trondheim, Norway, 90-99.
195
AAR Facts Book
Hobbs, D.W. 1988. Alkali-Silica Reaction in Concrete. Thomas Telford, London, 183 p.
Hobbs, D.W. 1989. Countering alkali-silica reaction in concrete Experience in the United
Kingdom. Concrete Construction, C890948, The Aberdeen Group.
Lute, R. 2008. Evaluation of Coatings and Sealers for Mitigation of Alkali-Silica Reaction and/
or Delayed Ettringite Formation. M.S. Thesis, The University of Texas at Austin.
Mayer, H. 1998. "The Chemistry and Properties of Silicone Resins." Pigment & Resin
Technology, 26(6): 364-373.
McGettigan, E. 1992. "Silicon-Based Weatherproofing Materials." Concrete International,
14(6): 52-56.
Pedneault, A. 1996. Development of Testing and Analytical Procedures for the Evaluation of
the Residual Potential of Reaction, Expansion, and Deterioration of Concrete Affected by
ASR. M.Sc. Memoir, Laval University, Qubec City, Canada, 133 p.
Rotter, H.M. 1995. Alkali-Aggregate Reaction: from Basic Principles to Structural Behaviour
A literature Review. Rapport No. EPM/GSC 1995-11, cole Polytechnique de Montral,
Montral, PQ, 166 p.
Singhal, A.C. and Nuss, K.L. 1991. Cable Anchoring of Deteriorated Arch Dam. Journal of
Performance of Constructed Facilities, 5(1): 19-37.
Stark, D., Morgan, B., Okamoto, P. and Diamond, S. 1993. Eliminating or Minimizing AlkaliSilica Reactivity, SHRP-C-343, National Research Council, Washington, D.C.
Stokes, D.B., Thomas, M.D.A. and Shashiprakash, S.G. 2000. Development of a LithiumBased Material for Decreasing ASR-Induced Expansion in Hardened Concrete. Proceedings of
the 11th International Conference on Alkali-Aggregate Reaction in Concrete (Eds. M.A. Brub
et al.), CRIB, Quebec City, Canada, 10791087.
Wood, J.G.M. and Angus, E.C. 1995. Montrose Bridge: Inspection, Assessment and Remedial
Work to a 65 Year Old Bridge with AAR. Structural Faults and Repair 95, 6 p.
196
AAR Facts Book
Chapter 9 Alkali-Carbonate Reaction
9 Alkali-Carbonate Reaction
9.1 INTRODUCTION
As discussed in the introduction, alkali-carbonate reaction (ACR) is a reaction between alkali
hydroxides (NaOH and KOH) in the pore solution and certain carbonate rocks, particularly
calcitic dolostone and dolomitic limestones, present in some aggregates. The reaction is usually
accompanied by dedolomitization and expansion of the affected aggregate particles, leading to
abnormal expansion and cracking of concrete in service.
Cases of ACR-induced damage in concrete are much less widespread than cases of ASR and,
consequently, the carbonate reaction has attracted significantly less attention by researchers.
The precise mechanisms of expansion are equivocal, and it has been conjectured that ACR may
be a form of ASR involving reactive silica (e.g., cryptocrystalline quartz) present in the
dolomitic limestone; however, this theory has not been universally accepted.
There are some distinct features of ACR that differentiate it from ASR including, obviously, the
nature and composition of the rock, the timeframe of the reaction, the influence of aggregate
size, the inability of some standard tests to identify reactive aggregates, the threshold alkali
content required to generate expansion, and the impact of preventive measures such as
pozzolans, slag, and lithium-based compounds. Generally measures that have been found to be
effective in preventing expansion due to ASR do not work with alkali-carbonate rocks and it is
generally recommended that such rocks are avoided for use in concrete.
This chapter presents an overview of alkali-carbonate reaction (ACR) including a discussion of
the chemistry of the reaction, the mechanisms of expansion, contributing factors, and
differences between ACR and ASR. Field symptoms in ACR-affected concrete structures are
similar to those observed in structures affected by ASR, and these are discussed in chapter 3.
Chapter 4 discusses test methods for both types of alkali-aggregate reaction. Recently
developed AASHTO PP65, for identifying reactive aggregates, includes procedures for
detecting and, subsequently, avoiding alkali-carbonate reactive rocks. This practice is presented
in chapter 6.
197
AAR Facts Book
9.2 CHARACTERISTICS OF ALKALI-CARBONATE REACTIVE ROCKS

Alkali-carbonate reactive rocks can generally be described as argillaceous dolomitic limestones
and tend to have a characteristic texture and structure comprising dolomite [CaMg(CO3)2]
rhombohedra (typically 10 to 50 m) dispersed or floating in a fine-grained matrix of fine
calcite, clay, and, usually, silica. Figure 9.1 shows the microstructure of typical reactive rock.
Typically the carbonate-mineral composition of the rock consists of substantial amounts of
calcite (CaCO3) and dolomite, and the (HCl-acid) insoluble residue comprises of a significant
amount of clay (ACI 221 2000; Ozol 2006). A modification of this typical texture has been
observed in late expanding alkali-carbonate rocks which may have interlocking dolomite
rhombs in a coarser-grained matrix.
Figure 9.1. Photograph of Petrographic Thin Section Showing Dolomite Rhombs

(clear equilateral parallelograms) Floating in a Matrix of Fine-Grained
Calcite, Clay and Silica (dark background)
The bulk composition of some early and late expanders is shown in Table 9.1. Rogers (1986)
compared the composition of expansive and non-expansive limestones in Ontario and observed
that the expansive rocks fell within a fairly distinct region when plotted on a graph of
CaO:MgO ratio versus either the alumina (Al2O3) or (HCl-acid) insoluble residue (see Figure
9.2). This observation forms the basis of the chemical screening test (CSA A23.2-26A) for
potential alkali-carbonate reactivity which has been adopted by the Canadian Specification
(CSA A23.2-27A) and the AASHTO PP65 Standard Practice (see chapter 6).
198
AAR Facts Book
Table 9.1. Composition of Alkali-Carbonate Reactive Rocks

(adapted from Walker 1978; ACI 221 2000; Ozol 2006)
Acid-insoluble resi
due (%)
Dolomite (% total
carbonate)
5 to 15
Approx. 50
10 to 20
40 to 60
13 to 29
46 to 73
21 to 49
75 to 87
Kingston, Ontario, Early Expanders

(Swenson, 1957; Swenson and Gillott,
1960; Dolar-Mantuani, 1964)
Iowa, Illinois & Indiana Early Expanders
(Hadley, 1961, 1964)
Virginia Early Expanders
(Newlon and Sherwood, 1962; Sherwood
and Newlon, 1964)
Gull River, Ontario, Late Expanders
(Dolar-Mantuani, 1964)
Virginia Late Expanders
33
(Newlon, Ozol, and Sherwood, 1972)
199
> 90
AAR Facts Book
Figure 9.2. Using Chemical Composition as a Basis for Determining Potential Alkali-Carbonate
Reactivity of Quarried Carbonates (from CSA A23.2-26A)
Observation of the characteristic texture and composition is considered to be a reliable
indication of potential alkali-carbonate reactivity (Ozol 2006), and rocks that display the
characteristic texture or composition should be tested further to evaluate their reactivity in
concrete.
9.3 CHEMISTRY OF ALKALI-CARBONATE REACTION

Dolomitic limestones with the characteristic microscopic texture described in the previous
section will react in an alkaline environment and undergo a process termed dedolomitization8,
which can be written as follows:
CaMg(CO3)2 + 2ROH CaCO3 + Mg(OH)2 + R2CO3
dolomite + alkali hydroxide calcite + brucite + alkali carbonate
Eqn 9.1
ACI 221.1R-98 states there appear to be no known exceptions to the observation that all limestones that pos
sess the characteristic texture and composition will react, or dedolomitize, in an alkaline environment. Also, all
limestones that dedolomitize in an alkaline environment possess the characteristic texture and composition.
200
AAR Facts Book
where R represents K or Na. In some cases the magnesium may further react to form
magnesium silicate.
The alkali carbonate produced in the dedolomitization reaction may react with calcium
hydroxide in the cement paste as follows:
R2CO3 + Ca(OH)2 CaCO3 + 2ROH
Eqn. 9.2
thereby regenerating alkalis for further reaction. This reaction may result in carbonation
halos forming around reacted aggregate particles.
9.4 MECHANISMS OF EXPANSION DUE TO ACR

The dedolomitization reaction shown in Equation 9.1 is accompanied by a reduction in solid
volume. In other words the molar volume of CaCO3 plus Mg(OH)2 is less than the molar
volume CaCO3MgCO3 and, since Na2CO3 is expected to pass in to solution, no volume
expansion is expected (Gillott 1964). Hence, dedolomitization does not provide a simple
mechanical explanation for expansion. Several theories have been proposed to explain the
expansion mechanism, and these include:
Swelling of clay matrix (Swenson and Gillott 1964; Gillott 1964; Gillott and Swenson
1969). Dedolomitization of the reactive material exposes previously trapped clay
minerals (illite and chlorite) to alkali solution. The previously unhydrated clay minerals
are in an active state. Exchange sites on the surfaces of the clay, which may be vacant
or occupied by Ca2+ or Mg2+ ions, absorb Na+ (and K+) ions from solution (cation
exchange) and a double layer develops accompanied by water uptake, which leads to
swelling of the clay minerals and, consequently, an increase in the solid volume of the
aggregate. The expansion essentially results from the wetting (and cation exchange) of
previously unhydrated clay minerals that was made possible by the dedolomitization
reaction which opened access channels for the moisture.
Growth and rearrangement of the products of dedolomitization (i.e., brucite and calcite)
through a topochemical reaction (Tang et al. 1987). Clay is not embedded in dolomite
rhombs but forms a network between the fine calcite grains surrounding the rhombs.
The clay network provides channels for alkali hydroxides in the pore solution to
attack the dololomite rhombs resulting in dedolomitization. After reaction, brucite is
observed to form around the outside of the original rhomb in rings approximately 2 m
thick, the original structure of the rhomb being maintained by a framework of calcite
which largely remains in place after the migration of the magnesium. The original
rhomb with the peripheral brucite layer requires more space resulting in the generation
of expansive forces. A schematic illustrating the phenomenon is presented in Figure 9.3.
201
AAR Facts Book
Alkali-silica reaction of cryptocrystalline quartz in matrix surrounding the dolomite

rhombs (Katayama 1992; 2010). Alkali-silica gel has been observed in mortars and
concretes containing alkali-carbonate reactive aggregates, and it has been postulated that
reaction of cryptocrystalline quartz contained within the carbonate is responsible for the
expansion of the mortar and concrete, and that the accompanying dedolomitization is
harmless.
Figure 9.3. Schematic Showing Topochemical Formation of Brucite (and Calcite) Ring
around Dolomite Rhomb (modified from Tang et al. 1987)
Although the clay-swelling mechanism appears to have lost support, there is still debate among
researchers whether the expansion is due to the dedolomitization reaction or the alkali-silica
reaction.
9.5 FACTORS CONTRIBUTING TO EXPANSION DUE TO ALKALI-CARBONATE

REACTION
Regardless of the mechanism(s) causing expansion there are a number of contributing factors
that increase the expansion of concrete containing reactive argillaceous dolomitic limestones.
The petrological features of the rock that appear to maximize reactivity/expansion are (Ozol
2006):
5 to 25% clay or insoluble residue,
Ratio of calcite:dolomite of approximately 1:1,
Increase in dolomite content up to point at which interlocking of the dolomite rhombs
becomes a restraining factor, and
Small size of the discrete floating rhombs.
202
AAR Facts Book
Other factors that increase expansion are (Ozol 2006):

Increasing coarse aggregate size,
Moisture availability,
Increased alkali content of concrete (or pH of solution),
Increased proportion of reactive stone in the coarse aggregate, and
Lower concrete strength.
Clearly if the expansion truly does result from alkali-silica reaction the quantity and form of the
reactive silica will be a major determining factor. Grattan-Bellew et al. (2010) have
demonstrated that there is a correlation between the amount of quartz in the insoluble residue
and expansion of concrete prisms produced with three different reactive limestones.
9.6 DIFFERENCES BETWEEN ACR AND ASR

As discussed above, the debate between the cause(s) of expansion in concrete containing
reactive argillaceous dolomitic limestones is ongoing and, at the time of writing, there appears
to be no overall consensus. However, regardless of whether ACR or ASR produces the
expansion, there are some features of reactive aggregates that undergo the dedolomitization
reaction that set them apart from aggregates that are unequivocally alkali-silica reaction; these
are summarized as follows:
Expansion in the laboratory or damage in the field occurs in a much shorter timeframe
with ACR.
Expansion increases with an increase in the size of the reactive aggregate particles.
Many tests used for the detection of alkali-silica reaction are not capable of detecting
ACR rocks; these include the mortar bar test (ASTM C 227), the accelerated mortar bar
test (ASTM C 1260), and the quick-chemical test (ASTM C 289).
Expansion occurs in concrete with alkali contents below the threshold level generally
considered necessary for ASR.
The use of pozzolans and slag at levels normally sufficient for controlling ASR is not
effective with ACR rocks, even when combined with low-alkali cement.
Lithium-based compounds are not effective in controlling expansion with ACR rocks.
Indeed, concrete microbars with ACR aggregate and immersed in LiOH solution have
been observed to expand, albeit at 150C (Tang et al. 2000).
203
AAR Facts Book
Little (or no) gel is observed after expansion has occurred. Even when alkali-silica gel
has been observed it is often difficult to reconcile the high amount of expansion with the
meagre content of gel.
Proponents of the ASR mechanism have provided partial explanation for some of the apparent
anomalies between the behaviour of these rocks and that of established alkali-silica reactive
rocks (Katayama 2010; Grattan-Bellew et al. 2010). However, it is clear that rocks that cause
expansion and undergo dedolomitization need to be treated as a separate category with regards
to AAR, whatever the predominant expansion mechanism (ACR or ASR).
9.7 PREVENTION OF EXPANSION DUE TO ALKALI-CARBONATE REACTION

Regardless of the true mechanism of expansion, it is extremely difficult to control expansion of
alkali-carbonate reactive aggregates that have the characteristic composition and texture
described in section 9.2 and undergo the dedolomitization reaction described in section 9.3.
Consequently, it is strongly recommended that such aggregates are not used in the production of
concrete.
The reactive material may be avoided by selective quarrying or, alternatively, the reactive phase
can either be diluted until the reactive material represents less than 20% of the coarse aggregate
(or 15% of the total aggregate if both fine and coarse aggregate contain reactive material), or
crushed to a smaller particle size until deleterious expansion is eliminated. However, avoidance
of the reactive material appears to be the most prudent solution in the authors opinion.
It has been suggested that expansion can be reduced to an acceptable level using very low alkali
cement, e.g. no greater than 0.40% Na2Oe (Newlon et al. 1964). However, expansion has been
observed with cements with alkali contents only marginally above this level such as 0.43%
Na2Oe (Deng Min and Tang Mingshu 1993). Williams and Rogers (1991) observed cracking in
sidewalk slabs after just 24 months of field exposure in Kingston, Ontario, Canada, even though
the alkali content of the concrete was just 1.74 kg/m3 (2.9 lb/yd3) Na2Oe. A section of sidewalk,
curb, and gutter in the same vicinity also exhibited cracking, and it is believed (Williams and
Rogers 1991) that the alkali content of the cement was just 0.31% Na2Oe. Figure 9.4 (from
Shehata et al. 2009) shows the effect of cement alkalis on the expansion of concrete prisms
produced with an argillaceous dolomitic limestone (Pittsburg aggregate) compared with a
siliceous limestone (Spratt aggregate). The results provide a good illustration of the different
behavior of ACR versus ASR reactive limestones.
204
AAR Facts Book
Figure 9.4. Effect of Cement Alkali Content on Expansion of Concrete Prisms with
Argillaceous Dolomitic Limestone versus Siliceous (Spratt) Limestone (Shehata et al. 2009)
Swenson and Gillott (1964) reported that a variety of pozzolans found to be effective with the
alkali-silica reaction when used at replacement levels of 25% were not effective in mitigating
ACR. Subsequent research has reinforced this finding. Ground granulated blast-furnace slag
used at replacements of 25 and 50% with high alkali cement of 1.04% Na2Oe was also not
effective in controlling expansion of laboratory-exposed concrete prisms and field-exposed
concrete slabs produced with the Pittsburg Quarry dolomitic limestone from Kingston, Ontario
(Rogers and Hooton 1992). Thomas and Innis (1998) tested the same aggregate with slag in
concrete prisms (CSA A23.2-14A; ASTM C 1293) and found that neither 65% slag combined
with high-alkali cement (1.25% Na2Oe) nor 50% slag combined with low-alkali cement (0.50%
Na2Oe) was sufficient to control expansion; the data are shown in Figure 9.5. The data in Figure
9.5 serve again to accentuate the different behavior of ACR aggregate (Pittsburg) and ASR
aggregate (Spratt). Tang et al. (1994) have shown that very high levels of fly ash (70%) or slag
(90%) may be effective in controlling expansion when combined with low-alkali cement
(0.43% Na2Oe). Perry and Gillott (1985) showed that 20% silica fume was not effective in
suppressing the expansion due to ACR.
205
AAR Facts Book
Figure 9.5. Effect of Slag Content and Cement Alkalis on Expansion of Concrete with
Argillaceous Dolomitic Limestone versus Siliceous (Spratt) Limestone (Thomas and Innis
1998)
Testing has also shown that incorporating lithium-compounds (e.g., LiCl, LiOH, and Li2CO3) in
concrete was not effective in controlling ACR expansion (Swenson and Gillott 1964; Wang et
al. 1994). Such compounds have been shown to be effective with many (but not all) alkali-silica
reactive aggregates, including siliceous limestones (Feng et al. 2010).
9.8 SUMMARY OF ALKALI-CARBONATE REACTION

The alkali-carbonate reaction was first documented in 1957 (Swenson 1957). Since this time,
damage due to ACR has been observed in a limited number of states in the USA and a few
other countries. ACR is much less widespread than ASR. Rocks that are alkali-carbonate
reactive have a characteristic composition and microscopic texture, and expansion of concrete
containing these rocks appears to be accompanied by dedolomitization of the rock. It is not
been established unequivocally whether expansion results directly from the dedolomitization
reaction or whether other mechanisms create the expansive forces. One school of thought is that
reactive silica is present in these rock types and that the expansion is due to alkali-silica
reaction. Regardless of whether the expansion is due to ACR or ASR, reactive rocks that have
the characteristic composition and texture and that undergo dedolomitization clearly behave
differently compared with reactive aggregates that are universally accepted as being alkali-silica
reactive. From a practical standpoint the main differences are that these rocks cause expansion
of concrete at very low alkali contents and even when relatively high replacement levels of
206
AAR Facts Book
pozzolans or slag are used. As such, it is very important that appropriate test procedures are
adopted to identify these rock types and that they are excluded from use for the production of
concrete.
207
AAR Facts Book
REFERENCES
ACI 221. 1998. Report on alkali-aggregate reactivity. ACI 221.1R-98, American Concrete
Institute, Farmington Hills, MI. 30 p.
PA, 6 p.
CSA. 2009. " Potential Expansivity of Aggregates (Procedure for Length Change Due to Alkali
-Aggregate Reaction in Concrete Prisms at 38 C. CSA A23.2-14A. A23.2-09 - Test methods
and standard practices for concrete, Canadian Standards Association, Mississauga, Ontario,
Canada.
Deng, M. and Tang, M. 1993. Measures to inhibit alkali-dolomite reaction. Cement and
Dolar-Mantuani, L. 1964. Expansion of Gull River Carbonate Rocks in Sodium Hydroxide.
Highway Research Record No. 45, Highway Research Board, Washington, D.C., 178-195.
208
AAR Facts Book
Feng, X., Thomas, M.D.A., Bremner, T.W., Folliard, K.J. and Fournier, B. 2010. Summary of
research on the effect of LiNO3 on alkalisilica reaction in new concrete. Cement and
Gillott, J.E. 1964. Mechanism and Kinetic of Expansion in the Alkali-Carbonate Rock
reaction. Canadian Journal of Earth Science, 1: 121-145.
Gillott, J.E. and Swenson, E.G. 1969. Mechanism of the Alkali-Carbonate Rock reaction.
Journal of Engineering Geology, 2: 7-23.
Grattan-Bellew, P.E., Mitchell, L.D., Margeson, J. and Deng, M. 2010. Is alkalicarbonate
reaction just a variant of alkalisilica reaction ACR=ASR? Cement and Concrete Research,
40: 556562.
Proceedings of the Highway Research Board, 40: 462-474.
Hadley, D.W. 1964. Alkali-Reactive Carbonate Rocks in IndianaA Pilot Regional
Investigation. Highway Research Record No. 45, Highway Research Board, Washington,
D.C., 196-221.
Katayama, T. 1992. A critical review of carbonate rock reactions is their reactivity useful or
harmful? Proceedings of the 9th International Conference on Alkali-Aggregate Reaction in
Concrete (ICAAR) (Ed. A.B. Poole), London, U.K., 508517.
Katayama, T. 2010. The so-called alkali-carbonate reaction (ACR) Its mineralogical and
geochemical details, with special reference to ASR. Cement and Concrete Research, 40: 643
675.
Newlon, H.H., Jr. and Sherwood, W.C. 1962. An Occurrence of Alkali-Reactive Carbonate
Rock in Virginia. Bulletin 355, Highway Research Board, Washington, D.C., 27-44.
Newlon, H.H., Jr. and Sherwood, W.C. 1964. Methods for Reducing Expansion of Concrete
Caused by Alkali-Carbonate Rock Reactions. Highway Research Record No. 45, Highway
Research Board, Washington, D.C., 134-150.
Newlon, H.H., Jr., Ozol, M.A. and Sherwood, W.C. 1972. An Evaluation of Several Methods
for Detecting Alkali-Carbonate Reaction. Progress Report No. 5, 71-R33, Virginia Highway
Research Council, Charlottesville, VA, May, 73 pg.
Ozol, M.A. 2006. Alkali-carbonate rock reaction. Significance of Tests and Properties of
Concrete and Concrete Making Materials (Ed. J.F. Lamond and J.H. Pielert), ASTM STP 169D,
Chapter 35, American Society for Testing and Materials, West Conshohocken, PA, 410-424.
209
AAR Facts Book
Perry, C. and Gillott, J.E. 1985. The Feasibility of Using Silica Fume to Control Concrete
Expansion Due to Alkali-Aggregate Reaction. Durability of Building Materials, 3: 133-146
Rogers, C.A. 1986. Evaluation of the potential for expansion and cracking of concrete caused
by the alkali-carbonate reaction. Cement, Concrete and Aggregates, 8(1): 13-23.
Rogers, C.A. and Hooton, R.D. 1992. Comparison between laboratory and field expansion of
alkali-carbonate reactive aggregates. Proceedings of the 9th International Conference on Alkali
-Aggregate Reaction in Concrete, Vol. 2, The Concrete Society, Slough, 877-884.
Shehata, M., Jagdat, S., Lachemi, M. and Rogers, C. 2009. Do supplementary cementing
materials control alkali-carbonate reaction? Proceedings of the 17th Annual Symposium (Eds.
D. Fowler and J. Allen), International Center for Aggregates Research, The University of
Texas, Austin, May.
Sherwood, W.C. and Newlon, H.H., Jr. 1964. A Survey for Reactive Carbonate Aggregates in
Virginia. Highway Research Record No. 45, Highway Research Board, Washington, D.C.,
222.
Swenson, E.G. 1957. A reactive aggregate undetected by ASTM tests. Proceedings of
American Society for Testing and Materials, 57: 48-51.
Swenson, E.G. and Gillott, J.E. 1960. Characteristics of Kingston Carbonate Rock Reaction:
Concrete Quality Control, Aggregate Characteristics and the Cement Aggregate Reaction.
Highway Research Bulletin No. 275, Highway Research Board, Washington, D.C., 18-31.
Swenson, E.G. and Gillott, J.E. 1964. Alkali-carbonate rock reaction. Highway Research
Record, 45: 21-40.
Tang, M., Liu, Z. and Han, S. 1987. Mechanism of alkali-carbonate reaction. Proceedings of
the 7th International Conference on Concrete Alkali-Aggregate Reactions (Ed. P.E. GrattanBellew), Noyes Publications, New Jersey, 275-279.
Tang, M., Deng, M. and Xu, Z. 2000. Comparison between alkali-silica reaction and alkalicarbonate reaction. Proc. 11th International Conference on Alkali-Aggregate Reaction in
Concrete (Ed. M-A. Berube et al.), Quebec City, 109-118.
Walker, H.N. 1978. Chemical Reactions of Carbonate Aggregates in Cement Paste. Special
Technical Publication 169-B, ASTM, Philadelphia, PA, 722-743.
210
AAR Facts Book
Wang, H., Tysl, S. and Gillott, J.E. 1994. Practical Implications of Lithium Based Chemicals
and Admixtures in Controlling Alkali-Aggregate Reactions. Special Publication 148,
American Concrete Institute, 353-366.
Williams, D.A. and Rogers, C.A. 1991. "Field Trip Guide to Alkali-Carbonate Reactions in
Kingston, Ontario." MI-145, Ontario Ministry of Transportation, Downsview, Ontario.
211
AAR Facts Book
212
FHWA-HIF-13-019

Alkali-Aggregate Reactivity (Aar)

Uploaded by

Copyright:

Available Formats

Alkali-Aggregate Reactivity (Aar)

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Alkali-Aggregate Reactivity (Aar)

Uploaded by

Copyright:

Available Formats

ALKALI-AGGREGATE REACTIVITY (AAR)

2. Government Accession No.

3. Recipients Catalog No.

4. Title and Subtitle

6. Performing Organization Code

8. Performing Organization Report No.

9. Performing Organization Name and Address

10. Work Unit No.

12. Sponsoring Agency Name and Address

13. Type of Report and Period Covered

11. Contract or Grant No.

14. Sponsoring Agency Code

15. Supplementary Notes

17. Key Words

9. Security Classif. (of this

20. Security Classif. (of this

18. Distribution Statement

SI* (MODERN METRIC) CONVERSION FACTORS

When You Know

TEMPERATURE (exact degrees)

FORCE and PRESSURE or STRESS

APPROXIMATE CONVERSIONS FROM SI UNITS

When You Know

TEMPERATURE (exact degrees)

FORCE and PRESSURE or STRESS

1.2 HISTORY OF AAR.............................................................................................................2

1.3 AAR TODAY .......................................................................................................................4

2 Fundamentals of Alkali-Silica Reaction ................................................................................9

2.1 INTRODUCTION ...............................................................................................................9

2.2 CHEMISTRY OF THE REACTION ..............................................................................10

2.3 MECHANISM OF EXPANSION ....................................................................................13

2.4 REACTIVE SILICA .........................................................................................................15

2.4.1 Types of Reactive Silica ...............................................................................................15

2.4.2 Effect of Aggregate Size ...............................................................................................18

2.4.3 Pessimum Effect ........................................................................................................... 20

2.5 SOURCES OF ALKALI ...................................................................................................22

2.5.1 Pore Solution Composition ...........................................................................................22

2.5.2 Threshold Alkali Contents ............................................................................................25

2.5.3 Alkali Recycling ........................................................................................................... 27

2.6 ROLE OF MOISTURE.....................................................................................................29

3 Symptoms of Alkali-Silica Reaction (ASR).........................................................................35

3.1 INTRODUCTION .............................................................................................................35

3.2 CRACKING .......................................................................................................................35

3.3 EXPANSION CAUSING MOVEMENTS AND DEFORMATIONS...........................40

3.4 POP-OUTS .........................................................................................................................43

3.5 SURFACE DEPOSITS (GEL EXUDATION VS. EFFLORESCENCE).....................45

4.1 INTRODUCTION .............................................................................................................47

4.2 ASR TEST METHODS FOR EVALUATING AGGREGATE REACTIVTY ...........47

Aggregate Combinations (Mortar-Bar Method) .................................................................... 51

Length Change of Concrete Due to Alkali-Silica Reaction................................................... 54

4.3 ASR TEST METHODS FOR EVALUATING PREVENTIVE MEASURES ............ 58

Ground Blast-Furnace Slag in Preventing Excessive Expansion of Concrete Due to the Al

Length Change of Concrete Due to Alkali-Silica Reaction................................................... 63

4.3.4 Other ASR Test Methods.............................................................................................. 64

4.4 ACR TEST METHODS FOR EVALUATING AGGREGATE REACTIVITY......... 66

5 Prevention of Alkali-Silica Reaction.................................................................................... 75

5.1 INTRODUCTION ............................................................................................................. 75

5.2 PREVENTIVE MEASURES OPTIONS...................................................................... 75

5.3 USE OF NON-REACTIVE AGGREGATES ................................................................. 76

5.4 LIMITING THE ALKALI CONTENT OF THE CONCRETE................................... 76