1

I i i fInvestigation of C t A tConcrete Aggregate f Alk li C b tfor Alkali-Carbonate

ReactionReaction

AAR Reactions2

AAR Reactions are chemical reactions between theAAR Reactions are chemical reactions between the cement and the aggregate causing the concrete to deteriorate.

There are two AAR reactions based on the reactive mineral type:Alkali Silica Reaction (ASR) with unstable forms ofAlkali-Silica Reaction (ASR) with unstable forms of silica (most common)Alkali-Carbonate Reaction (ACR) with certainAlkali Carbonate Reaction (ACR) with certain dolomite textures (least common)

Both are destructive, are not mutually exclusive, and under the right conditions cyclic.

This presentation is focusing on ACR.

Why investigate for a Why investigate for a “uncommon” reaction?“uncommon” reaction?

3

“uncommon” reaction?“uncommon” reaction?ACR cannot be mitigated for. The addition of fly ash, g y ,GGBFS, or silica fume do nothing to stop the reaction. It requires only a little alkali to start the reactionO th ti i t t d it d t t til llOnce the reaction is started, it does not stop until all aggregates have reacted destroying the concrete.It is the hardest reaction to identify when investigatingIt is the hardest reaction to identify when investigating an affected structure because its byproduct, brucite is very difficult to find even by microscopic examination. It does not have clear cut evidence like ASRIt does not have clear cut evidence like ASR. If reactive minerals for both ASR and ACR are present, both reactions are occurring and eventually the g ypresence of the silica gel will obscure any evidence of ACR.

Chickamauga LockChickamauga Lock

4

1937-39-Spillway deck poured 1999

Chickamauga LockChickamauga Lock

p y p1941-Lock and Dam completed1946-Cracking is a problem1963 petrographic-no silica gel or brucite found

1999

1963 petrographic-no silica gel or brucite found1970 petrographic-”AAR reaction definitelyoccurring”, no byproduct1993 94 t hi1993-94 petrographics-positive identification of silica gelg1996 petrographic-silica gel and brucite identified1997 petrographics-only1997 petrographics only found silica gel

Pre-1957

Chickamauga Chickamauga 5

PetrographicsPetrographicsReview of the 1930’sReview of the 1930 s petrographic examination of the aggregate showed 4-25% chert AND 5-32% dolomitechert AND 5-32% dolomite rhombs with 10-41%insoluble residue matrix. The right mineral compositions for bothmineral compositions for both ASR AND ACR. From the lack of by product for the first 25 years of expansion, it is logical to assume the ACR started first.

Definition6

The Alkali-Carbonate Reaction is a chemical reaction between hydroxyl y y

ions associated with the alkalis sodium and potassium in the cement pand certain dolomitic textures in the aggregate resulting in expansion and gg g g peventually cracking of the hardened

concrete.

Conditions Required Conditions Required 7

for ACRfor ACRAlkali Content in the concrete

The industry limit for Equivalent Alkalies in cement is 0.6% which is NOT low enough to stop ACR. Research has indicated the cement alkali content would need tohas indicated the cement alkali content would need to be lower than 0.4% and still be combined with other preventative measures to suppress the reaction.Alkalis may come from other sources than just cementAlkalis may come from other sources than just cement such as cementitious materials, aggregates (feldspars), local water, soil, or deicing agents (Potassium Acetate)(Potassium Acetate).A free alkali content of 2.0 kg/m3 or 0.12 lbs/ft3 in the concrete is normally sufficient to allow a reaction.ACR i li ti hi h it t t itACR is a cyclic reaction, which means once it starts it can generate it’s own (Na, K, Li) OH to continue the reaction.

Reactive AggregateReactive Aggregate8

Dolomite / Dolomitic Limestone –10 to 85% dolomite and 5-50% insoluble residue.Specifically dolomite rhombohedra crystals in a fine grained matrix of calcitecrystals in a fine grained matrix of calcite and insoluble residue (clay and silica). A pore structure that allows alkali solutionA pore structure that allows alkali solution to penetrate the rock

ACR Reactive TextureACR Reactive Texture9

Doloomitte

RRhom Matrix

Insoluble Residue m

bsPlain Light 63X Cross Nichols Light 63X

MatrixResidue

ACR ReactionACR Reaction10

1) CaMgCO3 + 2(Na,K)OH Mg(OH)2 + CaCO3 + (Na,K)2CO3

Dolomite + Alkali Brucite + Calcite + Alkali H d id C b tHydroxide Carbonate

2) (Na,K)2CO3 + Ca(OH)2 2(Na,K)OH + CaCO3

Alkali + Cement Alkali + CalciteAlkali + Cement Alkali + CalciteCarbonate Hydration Hydroxide

(Portlandite)

3) Repeat first reaction

LiOH also causes expansion but not at the rate that NaOH and KOH does.

BruciteBrucite11

Photomicrograph of crack, partially filled with brucite, from ACR g p , p y ,deterioration of impure dolomitic limestone aggregate in a pre-stressed concrete. Crack runs through fossiliferous carbonate particle into cement paste on right. The brownish-

ll d it i l ftyellow deposit in left side of crack is brucite. Blue sections of crack are filledof crack are filled with blue epoxy for clarity. Field width approximately 4.8 mm.

Used courtesy of Steven J. Stokowski, Jr. and Janet Sarson of Stone Products Consultants

approximately 4.8 mm. Plane polarized light.

Early vs. Late Early vs. Late 12

ExpandersExpandersEarly Expanders - dolomite rhombohedra crystalsEarly Expanders dolomite rhombohedra crystals floating in a fine grained matrix of calcite and insoluble residue (clay and silica). The calcite to dolomite ratio is approximately 1 to 1 with 5 to 25% insoluble residue. The

ti ll h i ith i th fi t 5reaction usually shows expansion with in the first 5 years of the project.

Late Expanders – has more or larger dolomite crystals so the crystals interlock, decreasing the calcite/clay/silica matrix and slowing down the reaction. Matrix also is not as fi i d i h E l E d Th d l ifine grained as in the Early Expanders. The dolomite content is 75 to >90% of the soluble content and the rock contains 10 to 25% calcite and 21 to 49% insoluble material The Calcite:Dolomite ratio is between 0 14:1 andmaterial. The Calcite:Dolomite ratio is between 0.14:1 and 0.36:1. These can take 10 to 20 years before expansion shows in the project.

Potential Early vs. Later ExpanderPotential Early vs. Later Expander13

ThinThin sections

were stainedstained,

the calcite

accepts the stain while thewhile the dolomite does not,

so theso the pink is calcite.63X 63X

AlkaliAlkali--Silica and AlkaliSilica and Alkali--Carbonate Reactive AggregateCarbonate Reactive Aggregate

14

gg ggg g

te

sD

olom

itR

hom

bsD R

ble

aU

nsta

bSi

lica

Plain Light 63X Cross Nichols Light 63X

Late ExpandersLate Expanders15

pp

It is much more difficult to distinguish between Late Expanders and non-reactive Dolomite rocks. There has even been some dolomite rock with nohas even been some dolomite rock with no limestone and very little clay/silica that have expanded. All occurrences of carbonate rock that have been found to be reactive have portions of the dolomite rhomb texture. However, not all dolomitic ,limestones with this texture expand with sufficient force to cause distress.There have been many structures built out of goodThere have been many structures built out of good dolomite aggregate.

Determining if your Determining if your aggregate is potentially aggregate is potentially

16

gg g p ygg g p yreactivereactive

Sampling the aggregateSampling the aggregateProcure a good map of the quarry to be able to judge the overall quality of the quarry andjudge the overall quality of the quarry and determine is selective quarrying is required.Best if possible to test a core or cores from theBest if possible to test a core or cores from the quarry. Ledge rocks are difficult to collect and to ensure full coverage of the quarry.g q yTake the petrographic samples and the samples for the ASTM C 586 tests from the same sample/location. This way there is good correlation between the two tests.

ACR TestsACR Tests17

ACR TestsACR Tests

1) ASTM C 295 P t hi E i ti1) ASTM C 295 Petrographic Examination2) ASTM C 586 Potential Alkali Reactivity of Carbonate Rocks for Concrete Aggregate (Rock Cylinder Method)3) ASTM C 1105 Length Change of Concrete Due to Alkali-Carbonate Rock Reaction

ACR Tests on Cores or ACR Tests on Cores or Ledge RocksLedge Rocks

18

Ledge RocksLedge Rocks

ASTM C 295 Petrographic – determines the amount of reactive minerals and textures in aggregate.ASTM C 586 Potential Alkali Reactivity of C b t R k 11/32 i h di tCarbonate Rocks – a 11/32 inch diameter by 1-3/8 inch long rock cylinder is cored out of the ledge rock and soaked in NaOHout of the ledge rock and soaked in NaOH at room temperature for 1-12 months.

Petrographic TestPetrographic Test19

g pg p

1) X-ray diffraction will determine amount of dolomite, calcite, and insoluble reside in the sample but cannot determine the typethe sample but cannot determine the type or texture.2) Microscopic Thin Sections (25-30µm)will determine type and texture of thewill determine type and texture of the dolomite, calcite, and insoluble residue in the aggregate.3) Ultra-Thin Thin-Sections (10-15 µm) to accurately discern the reactive texture since most of the dolomite rhombs are 25µm or less.

ACR Petrographic ACR Petrographic Prediction DiagramsPrediction Diagrams

20

Prediction DiagramsPrediction Diagrams2 diagrams use chemical composition of the rock to2 diagrams use chemical composition of the rock to predict potential ACR expansion.

Canadian Standards

CaO: MgO RATIO VERSES INSOLUBLE RESIDUE PERCENTAGE FOR A SUITE OF ALKALI-CARBONATE REACTIVE AND NON-REACTIVE CARBONATE ROCKS (CSA A23.2-26A & C.A. Rogers, “Evaluation of the Potential for Expansion and Cracking of Concrete Potentially Reactive Carbonate Rock Progress

R t #2 “St t id S f R ti C b t

Potentially Alkali Carbonate Reactive Rocks

Canadian Standards Association CSA A23.2-26A Virginia Council of Highway

Investigation and Research

200

100

50

p gCaused by the Alkali-Carbonate Reaction”, Figure 8)

AGGREGATES CONSIDERED NON-EXPANSIVE

Report #2 “Statewide Survey for Reactive Carbonate Aggregates” by W. Cullen Sherwood and Howard H. Newlon, Jr.; Virginia Council of Highway Investigation and Research; Feb. 1964.

CaO

: M

gO R

ATI

O

20

10

5

AGGREGATES CONSIDERED POTENTIALLY EXPANSIVE

2

21

2

2

0 5

2

1

55

50 45 40 35 30 25 20 15 10

INSOLUBLE RESIDUE IN PERCENT

AGGREGATES CONSIDERED NON-EXPANSIVE

Known ACR expanders have been plotted in this area.

1

1

2

1

Known ACR ExpandersKnown ACR Expanders

21

CaO:MgO RATIO VERSES INSOLUBLE RESIDUE PERCENTAGE FOR A SUITE OF ALKALI-CARBONATE REACTIVE AND NON-REACTIVE CARBONATE ROCKS (CSA A23.2-26A & C.A. Rogers, “Evaluation of the Potential for Expansion and Cracking

of Concrete Caused by the Alkali-Carbonate Reaction”, Figure 8) C S S

200

100

Known Alkali-Carbonate Reactors – ASTM STP 169D-pg. 414, Table 1K - Kingston, Ontario; I - Iowa, Illinois, & Indiana; VE - Virginia-Early; G - Gull River; Virginia-Late

AGGREGATESCONSIDERED

STP 169B Dolar-Mantuani, pg.733 STP 169D Milanesi, pg. 413 STP 169D Hilton, pg. 416 Very reactive 0.275/14wks

TIO

50

CONSIDERED NON-EXPANSIVE

AGGREGATES

D-M Very slow reaction took 40 wks Progress Report #7b, Hilton, 1968 Very reactive-1.21-1.05/21wks Medium reactive-0.97/21wks Slow reaction – 0.12/21 wks Table IV, pg. 25

CaO

: M

gO R

AT

20

10

AGGREGATES CONSIDERED POTENTIALLY

EXPANSIVE

Highway Research Record # 45, Lemish and Moore, 1964 Table 2, pg. 61 Very reactive-1.2/21wks Medium reactive-0.8/21wks Slow reactive-0.5-0.2/21wks

VE15

2

55

K1 K2I1 I2

I3 I4

VE2

VE4G1 G2G3 G4

VE3

0 5

2

1

50 45 40 35 30 25 20 15 10

INSOLUBLE RESIDUE IN PERCENT

AGGREGATES CONSIDERED NON-EXPANSIVE G3 G4

VL

Limestone QuarryLimestone Quarry

22

CaO : MgO RATIO VERSES INSOLUBLE RESIDUE PERCENTAGE FOR A SUITE OF ALKALI-CARBONATE REACTIVE AND NON-REACTIVE CARBONATE ROCKS (CSA A23.2-26A & C.A. Rogers, “Evaluation of the Potential for Expansion and Cracking of Concrete Caused by the Alkali-

40 3

34

4b 5

7

19

22 9

200

100

p g yCarbonate Reaction”, Figure 8)

AGGREGATESCONSIDERED

LEGEND Sample Numbers

37

38

1

2

16 218b10

Bench 1

Bench 2TIO

50

CONSIDERED NON-EXPANSIVE

AGGREGATES

25

4, 31, a1

37

15

1

30

8Bench 2

CaO

: M

gO R

AT

20

10

AGGREGATES CONSIDERED POTENTIALLY

EXPANSIVE

Aggregates with Late Expander Chemistry

b1

3635

32 33

30

4a 17 8a20

26 5

2

55

y

39

0 5

2

1 50 45 40 35 30 25 20 15 10

INSOLUBLE RESIDUE IN PERCENT

AGGREGATES CONSIDERED NON-EXPANSIVE (several known ACR expanders have been plotted in this area)

Dolomite QuarryDolomite Quarry23

CaO:MgO RATIO VERSES INSOLUBLE RESIDUE PERCENTAGE FOR A SUITE OF ALKALI-CARBONATE REACTIVE AND NON-REACTIVE CARBONATE ROCKS (CSA A23.2-26A & C.A. Rogers, “Evaluation of the Potential for Expansion and Cracking of Concrete Caused by the Alkali-

LEGENDBench 1Bench 2

Bench 3Benches 1 & 2

200

100

Potential for Expansion and Cracking of Concrete Caused by the AlkaliCarbonate Reaction”, Figure 8) FLORIDA ROCK QUARRY

AGGREGATESCONSIDERED

Sample Numbers

2 9 11 22

1, 12, 13 19

16

Bench 3Bench 4

TIO

50

CONSIDERED NON-EXPANSIVE

7, 24

5, 8, 17, 23

2, 9, 11, 22

3, c2

16

CaO

: M

gO R

AT

20

10

AGGREGATES CONSIDERED POTENTIALLY

EXPANSIVE

b2

a2

c14

5

55 Sample with Late Expander

a1

1

b1

0

AGGREGATES CONSIDERED NON-EXPANSIVE (several known ACR expanders have plotted in this area) 15 10 14 18 25

5

2

50 45 40 35 30 25 20 15 10

INSOLUBLE RESIDUE IN PERCENT

Chemistry

Known ACR ExpanderKnown ACR Expander24

pp

Potentially Reactive Carbonate Rock Progress

Potentially Alkali Carbonate Reactive Rocks

Known Alkali-Carbonate Reactors – ASTM STP 169D – pg. 414, Table 1 Kingston, Ontario – Early ExpanderIowa, Illinois, & Indiana – Early Expander Virginia – Early Expander G ll Ri O i L E d Potentially Reactive Carbonate Rock Progress

Report #2 “Statewide Survey for Reactive Carbonate Aggregates” by W. Cullen Sherwood and Howard H. Newlon, Jr.; Virginia Council of Highway Investigation and Research; Feb. 1964.

Gull River Ontario – Late ExpanderVirginia – Later Expander

STP 169D Milanesi’s Expander – pg. 413 STP 169B Dolar-Manuani, pg. 733

Very reactive 0 275/14 wksSTP 169D Hilton pg. 416

Very Reactive-1.21-1.05/21 wks

Very reactive 0.275/14 wks Very slow reaction, took 40 wks

Progress Report 7b, Hilton, 1968

Medium Reactive-0.97/21 wksSlow Reaction – 0.12/21 wks Table IV Pg 25

2

2

Table IV, Pg. 25

Highway Research Record #45 Lemish and Moore, 1964 Table 2, pg. 61 Very Reactive-1.2/21 wks

Known ACR expanders have 1

2

1

12

Medium Reactive-0.8/21 wks Slow Reactive-0.5-0.2/21wks

been plotted in this area. 1

16

Limestone QuarryLimestone Quarry25

Potentially Reactive Carbonate Rock Progress Report #2 “Statewide Survey for Reactive Carbonate Aggregates” by W. Cullen Sherwood and Howard H Newlon Jr ; Virginia Council of Highway

Potentially Alkali Carbonate Reactive Rocks# - core sample number that has potential for expansion or plots differently than the majority of passing samples

core sample that does not exceed limit of expansion H. Newlon, Jr.; Virginia Council of Highway Investigation and Research; Feb. 1964.

or # - Bench 1

or # - Bench 2

Sample with Late Expander Chemistryp p y

2

2

2

Known ACR expanders have been plotted in this area

1

2

1

1

8

21 8b

20 36

been plotted in this area.1

31a14

1530

32 334a17

b1

2635

39

Dolomite QuarryDolomite Quarry26

Potentially Reactive Carbonate Rock Progress Report #2 “Statewide Survey for Reactive Carbonate Aggregates” by W. Cullen Sherwood and Howard H Newlon Jr ; Virginia Council of Highway

Potentially Alkali Carbonate Reactive Rocks# - core sample number that has potential for expansion or plots differently than the majority of passing samples

core sample that does not exceed limit of expansion H. Newlon, Jr.; Virginia Council of Highway Investigation and Research; Feb. 1964. Bench 1

Bench 2Benches 1 & 2Bench 3Bench 4Bench 4

2

2

2

Known ACR expanders have been plotted in this area

10

c1 3a1

b1a2

b2

c2

4

2

2

1 11

been plotted in this area.

251131812

4

16 19 21 20

6

1

Petrographic Exam Petrographic Exam ––27

Thin SectionsThin Sections

When having the thin sections prepared, get half of the slide stained p p , gwith red, this allows for very easy identification of the calcite and dolomite.Thin sections should be prepared forThin sections should be prepared for all samples in the questionable zones of the ACR Prediction Diagramsof the ACR Prediction Diagrams

Limestone Quarry #36Limestone Quarry #36ACR & ASR R ti T tACR & ASR R ti T t

28

ACR & ASR Reactive TextureACR & ASR Reactive TexturePPL XN PPL 25X 25X

Limestone Quarry #10 &14 Limestone Quarry #10 &14 Less or NonLess or Non Reactive TexturesReactive Textures

29

Less or NonLess or Non--Reactive TexturesReactive Textures

#10 63X#10 63X PPL

Slight gtexture along a stylolite

# 14 63X

stylolite

63X PPL

NoNo ACR

texture

Dolomite Quarry #4 Dolomite Quarry #4 Reactive TextureReactive Texture

30

Reactive TextureReactive Texture

PPL XN 63XPPL 63X

XN 63X

Dolomite Quarry #12&13 Dolomite Quarry #12&13 NonNon--reactive Texturesreactive Textures 31

These had the

chemistry of late

expanders;expanders; from the

texture, it is stillstill

possible but unlikely, the texture

appears stable.

XN 63X

PPL 25X

stable.

Dolomite Quarry #18Dolomite Quarry #18--32

Late Expander TextureLate Expander TextureUltra-thinUltra thin

Thin section, it has been stained thestained, the slight pink tint is the

calcite disseminated

among theamong the dolomite

XN 63X

ASTM C 586 Rock ASTM C 586 Rock 33

Cylinder MethodCylinder Method586 samples should be chosen from the core after586 samples should be chosen from the core after petrographic results (at least the x-ray) is completed. Corresponding expansion from a 586 co p eted Co espo d g e pa s o o a 586with a reactive petrographic result will provide the best analysis of a quarry. Sampling from the ledge wall or the shotrock before the petrographics have been completed

fdoes not allow for good correlation between the tests. No matter how hard one attempts to match the shotrock/ledge rocks samples to the cores itthe shotrock/ledge rocks samples to the cores, it is never a sure match.

ASTM C 586 ASTM C 586 34

Sample PreparationSample PreparationEnsure test 0.8

0.9

cylinders have good smooth endpoints or

0.5

0.60.7

0.8

h endpoints or the data will be erratic and

0 1

0.2

0.3

0.4

ge in

Len

gth

wrong.

Th l-0.2

-0.1

00.1

Perc

ent C

han

These samples had their point reground after

0 6

-0.5-0.4

-0.3P

68 days in NaOH, early

data was lost.

-0.7

-0.6

0 50 100 150 200 250

Number of Days in NaOH

Aggregate TestsAggregate Tests35

gg ggg g

ASTM C 1105 Determination of Length Change of Concrete g gdue to ACR – concrete mortar bar is stored in high humidity at 73o F for one year. Job concrete mix is used so the alkali content is what would be in the job.ASTM C 1293 D t i ti f L th Ch f C tASTM C 1293 Determination of Length Change of Concrete due to ASR – concrete mortar bar with high alkali cement is stored in high humidity at 100o F for one year. Extra alkali is added to force reaction May be modified by decreasingis added to force reaction. May be modified by decreasing alkali, adding supplemental cementitious materials, or using the job mix.Despite the titles ASTM C 1105 and ASTM C 1293 cannotDespite the titles ASTM C 1105 and ASTM C 1293 cannot distinguish between ACR and ASR if the mineralogy is present for both reactions.If both mineralogies are present the theory is to run theIf both mineralogies are present, the theory is to run the test with sufficient cementitious materials to suppress the Alkali-Silica Reaction and thereby determine if the aggregate is also Alkali-Carbonate Reactive.

Preparations for Preparations for 36

ASTM C 1293/1105ASTM C 1293/1105

Because 1293 is run at 100oF and 1105 is tested at 73oF, 1293 is the harsher test; therefore running ASTM C 1293 should give a g gbetter estimate of the potential reactivity of an aggregate whether it is ASR or ACR. Federal Highway Administration recommendsFederal Highway Administration recommends the modified ASTM C 1293 be run for 2 years; therefore, these tests need to be planned in advance.Criteria for the two year test is a combination of 1105 and 1293 - <0.015/3mon; <0.025/6mon; <0.03/1yr; <0.04/2yr.<0.03/1yr; <0.04/2yr.

ACR Mitigation for ACR Mitigation for 37

New StructuresNew Structures

Avoid use of aggregate classified as potentially reactive by selective quarrying.Specify the use of low alkali cement andSpecify the use of low alkali cement and cementitious materials in the hope of keeping the alkali content of the concrete low enough the reaction will not start (< 2 kg/m3)the reaction will not start (< 2 kg/m3).Dilute so the amount of potentially reactive rock is less than 20% of the coarse or fine aggregate or 15% of the total if reactive material is present in both.Specify the minimum aggregate size that isSpecify the minimum aggregate size that is feasible for the project.

What to do if the What to do if the 38

aggregate is reactive.aggregate is reactive.

Hope it is a late expander andThat you have already retiredThat you have already retired.

Th l thi th t b d thThe only thing that can be done then is slot cutting to release the

i i th t iexpansion pressure in the concrete in the hopes of preventing cracking.

ReferencesReferences

39

ReferencesReferencesASTM STP 169B (1978), C (1994), and D (2006)E l ti f th P t ti l f E i d C ki f C t C d b Alk li C b tEvaluation of the Potential for Expansion and Cracking of Concrete Caused by Alkali-Carbonate Reaction, C. A. Rogers, 1986Virginia Council of Highway Investigation and Research Potentially Reactive Carbonate Rocks Progress Report #2 Statewide Survey for Reactive Carbonate Aggregates, W. C. Sherwood, H. Newlon, Jr., 1964Ontario Division of Mines Industrial Mineral Report 42 L Dolar Mantuani 1975Ontario Division of Mines, Industrial Mineral Report 42, L. Dolar-Mantuani, 1975Handbook of Concrete Aggregate, L. Dolar-Mantuani, 1983ACI 221.1R-98CSA A23.1-00/A23.2-00 Concrete Materials and Methods of Concrete Construction / Methods of Test for Concrete, 2001ASTM C 33 02ASTM C 33-02Highway Research Record #45, Highway Research Board, 1964Virginia Highway Research Council Potentially Reactive Carbonate Rocks Progress Report #7b The Effects of Textural and External Restraints on the Expansion of Reactive Carbonate Aggregates, M. H. Hilton, 1968U S B f R l ti P t h f th Chi k Li t W Y H ll d 1936U.S. Bureau of Reclamation, Petrography of the Chickamauga Limestone, W. Y. Holland, 1936Principles of Engineering Geology and Geotechnics, D. P. Krynine, W. R. Judd, 1957Federal Highway Administration FHWA-RD-03-047, Guidelines for the Use of Lithium to Mitigate or Prevent Alkali-Silica Reaction, 2003Federal Highway Administration FHWA HRT-04-150 Petrographic Methods of Examining hardened C t A P t hi M l H N W lk D S L P E St t 200Concrete: A Petrographic Manual, H. N. Walker, D. S. Lane, P. E. Stutzman, 200Use of the Accelerated Mortar Bar Test for Evaluating the Efficacy of Mineral Admixtures for Controlling Expansion due to Alkali-Silica Reaction, M. D. A. Thomas, F. A. Innis, 1999ACI Materials Journal 95-M71, Effect of Slag on Expansion Due to Alkali-Aggregate Reaction in Concrete, M. D. A. Thomas, F. A. Innis, 1998