7/30/2019 PL973[1]

1/8

Portland Cement Association

Contents

Volume 18/Number 3

December 1997

5420 Old Orchard Road

Skokie, Illinois 60077-1083

Phone: (847) 966-6200

Fax: (847) 966-8389

Web Site:

www.portcement.org

New Literature

Effect of Ettringite on FrostResistance

Origins of Chloride Limitsfor Reinforced Concrete

Effect of Ettringite on Frost Resistance

by Rachel J. Detwiler and Laura J. Powers-Couche*

In investigations of frost damagedconcrete, white deposits are com-

monly found in air voids. These de-posits are often identified asettringite. Questions arise as towhat role ettringite has on freeze-thaw deterioration of concrete. Alaboratory study was developed toexamine this issue. The followinghypotheses were suggested:

1. Gypsum-bearing deicing salt dif-fuses into the concrete. Gypsum,a common component of deicingsalts, reacts with the hydrationproducts of tricalcium aluminate(C

3

A) in the cement paste to formettringite that fills the air voidsuntil they are unable to protectthe concrete from frost damage.

2. Excess gypsum already presentin the cement reacts with the C

3A

hydration products, filling the airvoids with ettringite and com-promising the air-void system asdescribed above.

3. Deterioration is due to frost dam-age of inadequately air-entrainedconcrete. As the paste breaks upmicroscopically, ions go into so-lution and recrystallize asettringite in voids or cracks. Theettringite seen in the air voids isthus a consequence, not a cause,of frost damage.

Ettringite Formation

The mineral ettringite has limitednatural occurrence in rocks; how-ever, ettringite has widespreadoccurrence as a primary mineralin fresh concrete and as a second-ary mineral in mature concreteexposed to moisture. Mineralo-gists define primary minerals asthe minerals or phases initiallyformed in a system. Secondaryminerals are deposited or formedat a later time, usually as a resultof alteration of existing minerals.As a secondary mineral, ettringiteis generally deposited in openspaces such as air voids andcracks (Fig. 1). When it appears in

Figure 1. White ettringite deposits invoids of field concrete (20x).(S67533)

* Principal Engineer and Senior

Microscopist, respectively, Construction

Technology Laboratories, Skokie, Illinois

60077, USA.

R e t u r n To I n d e x

7/30/2019 PL973[1]

2/8

Concrete Technology Today

2R e t u r n To I n d e x

hardened concrete, its presence mayor may not indicate distress.

Gypsum is added to portland ce-ment clinker during finish grindingto provide sulfate to react with C

3A

during hydration in order to controlsetting. This reaction formsettringite.

If the supply of sulfate ions is notadequate to maintain the stability ofthe ettringite as the C

3A continues to

hydrate, the ettringite converts tomonosulfoaluminate.

The initial formation of primaryettringite is innocuous because ittakes place while the cement paste isstill plastic. Monosulfoaluminatetakes up less space than ettringite.Provided no additional sulfate be-comes available, it is stable.

Secondary ettringite crystals areneedle-like and are usually less than

0.1 mm long. Masses of needles ap-pear as white deposits. Secondarydeposits such as ettringite require ahand lens or stereomicroscope to beseen. Ettringite can be confused withother minerals, such as thaumasite,that have a tendency to needle-likegrowth. Confirmation of ettringiterequires expertise with mineralogy,the petrographic microscope, scan-ning electron microscope, or X-raydiffraction analysis.

It is generally agreed that the mor-phology (size and shape of grains or

crystals), associations (the mineralstypically found together), amount,and location of deposits within amass of concrete distinguish theettringite as harmful or harmless.

Materials and Mix Designs

The basic mix design and air-voidcharacteristics are shown in Tables 1and 2. Three target air contents forthe fresh concretes were initiallyused: 2 0.5%, 4 0.5%, and 6 0.5%. The low air contents were in-

tended to represent cases of deterio-rated pavements that are oftenfound to have air contents of 1% to4% with air-void spacing factors ex-ceeding the recommended maxi-mum of 0.20 mm (0.008 in.). Afterapproximately 300 cycles of testingwith no apparent signs of deteriora-tion, the 6% air specimens werefound to be too durable and tests ofthese samples were discontinued.

The aggregates were siliceoussand and gravel (19 mm max. size).

Three different cements, madefrom the same raw materials, wereused for the concretes: two ver-sions of a Type I cement with 12%C

3A, one having 2.03% sulfate and

the other having 3.14% sulfate, anda Type II cement with 5% C

3A and

2.72% sulfate. The higher sulfate,

higher C3A cements would be ex-pected to create more ettringitethan the lower sulfate, lower C

3A

cement.A 3-day moist cure followed by 25

days in laboratory air (23C [73F], 50%RH) was used to reflect field curing.

Exposure Conditions

The 76x76x286 mm (3x3x11-1/4 in.)prism specimens were tested accordingto ASTM C 666, Procedure A, Freez-ing and Thawing in Water, with the

following modifications:

1. Instead of water, one of two salt so-lutions was used. One consisted of3% sodium chloride (NaCl) bymass, the other, a 3% NaCl withadded gypsum. The level of gyp-sum added was 5% by mass ofNaCl to simulate the worst casethat would still meet ASTM D 632,which states that road salt must beat least 95% NaCl.

2. At the end of each week, the speci-mens were removed from the solu-

tions and allowed to dry over theweekend in laboratory air (simu-lated field exposure conditions).

Since the purpose of the test programwas not only to determine whether thepresence of ettringite in the entrainedair voids accelerates frost damage, but

also to establish the failure mecha-nism (that is, to determine whetherettringite deposition or frost dam-age comes first), it was necessary tomonitor the changes in the speci-mens in several ways through thecourse of the program. Before the

specimens were returned to their re-spective solutions at the beginningof each week, the mass, length, andfundamental transverse frequencywere measured and recorded. Sac-rificial companion specimens wereperiodically sampled for petro-graphic examination.

This test program demonstratedthe need to monitor freeze-thawsamples for changes in at leastthree properties: relative dynamicmodulus (RDM), mass, and lengthbecause each property reflects a

different mechanism of frost dam-age and they do not necessarilycorrelate with each other. For ex-ample a concrete may have nochange in length or RDM, but mayhave severe mass loss (scaling). Pe-trography is also helpful when cor-related with these properties.

Specimen Target air Air content, Air content, Spacing factor,

content, % fresh, %* hard, %** mm (in.)**

Type I 2 2.1 2.4 0.38 (0.015)

4 4.0 3.9 0.23 (0.009)

6 6.0 8.2 0.07 (0.003)

Type I with added 2 2.0 1.8 0.99 (0.039)

gypsum 4 3.8 3.8 0.36 (0.014)

Type II 2 2.5 3.2 0.56 (0.022)

4 4.5 3.5 0.20 (0.008)

Table 2. Air Void Characteristics for Laboratory Concrete Specimens

* ASTM C 231, "Standard Test Method for Air Content of Freshly Mixed Concrete by the Pressure

Method"

** ASTM C 457, "Standard Test Method for Microscopical Determination of Parameters of the Air-

Void System in Hardened Concrete"

Table 1. Concrete Mix Design

Cement 314 kg/m3 (530 lb/yd3)

Fine agg. 873 kg/m3 (1472 lb/yd3)

Coarse agg. 1067 kg/m3 (1798 lb/yd3)

Water Not to exceed 154 kg/m3

(259 lb/yd3)

Slump 50 12 mm (2 1/2 in.)

Water reducer Not to exceed 1 L/m3

(5 oz/cwt)

7/30/2019 PL973[1]

3/8

3

December 1997

R e t u r n To I n d e x

Microscopy

Proper specimen preparation tech-niques are essential in order to ob-tain valid microscopic observations.The specimens were stabilized withepoxy before drying in order to pre-vent further cracking from takingplace during preparation for micro-

scopic examination. Thus any cracksobserved were generated by the du-rability testing, not by the specimenpreparation techniques.

Some of the solid materials in ce-ment paste and concrete, particu-larly calcium hydroxide, ettringite,gypsum, and salts, are at least some-what soluble in water. For this rea-son only nonaqueous lubricantswere used for cutting and polishingof the specimens to a uniform thick-ness before identifying crystalphases using optical properties.

Specimens for air-void systemanalysis (ASTM C 457) and micro-scopical examination were cut trans-versely from the sacrificial speci-mens using an oil-lubricated saw.One side of each transverse slice waslapped for modified point-count lin-ear traverse analysis.

Microscopical examination (ASTMC 856) of each specimen was per-formed using a polarized-light mi-croscope. Observations were madeat four magnifications, 100x, 200x,400x, and 1000x.

Selected thin sections were laterexamined by scanning electron mi-croscopy using backscattered elec-tron imaging and X-ray microanaly-sis to supplement the informationobtained by optical microscopy.

Concretes with 2% Air

All2% air specimens were destroyedor had significant deteriorationwithin 360 cycles, exhibiting similarperformance in salt solutions with or

without gypsum. Figures 2, 3, and 4show the length change, masschange, and relative dynamic modu-lus for the specimens. Non-air-en-trained concretes used in the normalASTM C 666 Method A test wouldhave failed much sooner; however,these specimens lasted longer be-cause of the weekend drying period.

The concrete was severely scaled,exposing the coarse aggregate. Be-low the scaled surface were a fewshort microcracks oriented

subparallel to the surface and pass-ing around aggregates. No second-ary deposits were observed in themicrocracks. Scattered air voidswith diameters up to 1 mm werepresent. The air voids were empty;the walls were smooth, shiny, andfree of deposits. Thusthe concrete

had suffered severe damage withoutthe formation of secondary ettringitedeposits.

Concretes with 4% Air

The concretes with 4% air performedmuch better than the concretes with2% air, as shown in Figures 2through 4. In the Type I cementspecimens with a nominal 4.0% aircontent exposed to the salt + gyp-sum solution, a small amount ofettringite was observed in the air

voids located within approximately1 mm of the exposed surfaces of thetest specimens after 500 freeze/thawcycles. Ettringite occurred as clus-ters or rosettes of small needles, oras poorly-defined small masses pro-jecting into the air voids. Generally,only one small cluster or rosette waspresent in an air void and not everyvoid contained ettringite. The ma-jority of air voids remained empty.At the same time the specimens ex-posed to NaCl solution did not haveany detectable ettringite in the voids

and no microcracking was present.After 616 to 740 cycles, both sets ofspecimens from all the 4% air con-cretes had some ettringite and cal-cium hydroxide deposits in thevoids within 3 mm of the surface(voids below 3 mm were clean). Thespecimens exposed to salt + gypsumhad greater deposition of bothettringite and calcium hydroxide.After 1033 cycles, the amount of ma-terial deposited in the voids of bothsets (both solutions) of Type I speci-mens appeared to be about the same.

Voids containing deposits occurredwithin 3 mm of the exposed surface.A few short surface-parallelmicrocracks were observed at maxi-mum depths of 0.5 mm below theexposed surface. No secondary de-posits were observed in themicrocracks. At this stage thesespecimens were subjected to con-tinuous freeze/thaw cycling withoutthe weekend drying period.

After a total of 1291 cycles, the TypeI cement specimens showed no sig-

nificant length change or loss ofmodulus. However, they did showsome surface scaling and shallow sur-face-parallel microcracks. Portions ofthese microcracks contained loosemasses of long ettringite needles.Many air voids within 5 mm of thescaled surface contained ettringite or

calcium hydroxide, or both. Air voidswith diameters less than 25 m typi-cally appeared to be filled with sec-ondary deposits. Individual ettringiteneedles were approximately 1 to 2 mwide and 5 to 7 m long. All threemeasures of deterioration show theeffect of the deicer solution composi-tion to be insignificant.

The 4% air concretes made withType I cement with added gypsumdeteriorated faster than the otherspecimens with a nominal 4% aircontent because the spacing factor

was significantly higher at 0.36 mm(0.014 in.) versus 0.20 to 0.23 mm(0.008 to 0.009 in.) for the other 4%air concretes. As with the 2% airconcretes, the 4% air concretes withbetter air void systems performedbetter, as would be expected. As inthe other 4% air concretes, secondarydeposits in air voids were rare belowthe deteriorated surface layer, andcracks or microcracks with orienta-tions other than roughly parallel tothe surface were not present.

Conclusions1. With 2% air contents (no entrained

air), frost damage occurs withoutthe formation of ettringite depos-its. Without air entrainment thedeterioration process is too rapidto allow time for ettringite depo-sition to take place in the voidsbefore the concrete disintegrates.

2. The failure process was muchmore gradual for the 4% air con-cretes. Deposition of ettringite ap-peared to follow, rather than

cause, the formation of cracks.Cracks did not emanate fromettringite-filled air voids andettringite deposits did not spanthe cracks, but lined them. Theorientation of the ettringiteneedles perpendicular to thewalls of the cracks indicates thatthey grew into open (water-filled)space. All of these observationssuggest that the ettringite deposi-tion was neither a primary nor acontributing cause of the cracks.

7/30/2019 PL973[1]

4/8

Concrete Technology Today

4R e t u r n To I n d e x

3. The presence of gypsum in thesalt solution had no significant ef-fect on the durability of the con-cretes. Those with poor air voidsystems (nominal 2% air) per-formed poorly. Those with mar-ginal air void systems (nominal4% air) performed remarkablywell. The best protection againstfrost damage is a proper air voidsystem.

4. From the above observations, it isclear that the ettringite in thesespecimens did not cause thecracking, nor did it contribute tothe propagation of existingcracks. Rather, it appears to havebeen opportunistic. That is,cracks due to frost damage cre-

ated space for the crystals togrow. Water entered the cracksand dissolved ettringite from thehydrated cement paste. On dry-ing of the concrete, the ettringiteprecipitated in the open space leftby the cracks. Damage pro-gressed through repeated cyclesof freezing and thawing, withprogressive formation ofettringite deposits during thedrying periods.

5. The air void system played a farmore important role in frost resis-tance than the chemistry of ce-ment and deicer solutions or thepresence of ettringite and othersecondary deposits.

Acknowledgment

The research reported in this paper(PCA R&D Serial No. 2128c) wasconducted at Construction Technol-ogy Laboratories, Inc. with the spon-sorship of the Portland Cement As-sociation (PCA Project Index No. 93-05). The contents of this paper reflect

the views of the authors, who are re-sponsible for the facts and accuracyof the data presented. The contentsdo not necessarily reflect the viewsof the Portland Cement Association.

Figure 2. Length change, masschange, and relative dynamicmodulus for the Type I cementconcretes with 2% and 4% air andexposed to 3% NaCl solution orsalt-gypsum solution.

Figure 3. Length change, masschange, and relative dynamicmodulus for the Type I+gypsumcement concretes with 2% and4% air and exposed to 3% NaClsolution.

Figure 4. Length change, masschange, and relative dynamicmodulus for the Type II cementconcretes with 2% and 4% air andexposed to 3% NaCl solution orsalt-gypsum solution.

BB

BBBB

B

BB

BBBBBB

CCCCCC

C

CC

CCCCCCCEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE

EEEEEEEEEEEEEE22222222222222222222222222222222222222

222222222222222

-0.10

0.10

0.30

0.50

0.70

Lengthchange,

%

B 2% air, NaCl Solution

C 2% air, NaCl+Gypsum Soln.

E 4% air, NaCl Soln.

2 4% air, NaCl+Gypsum Soln.

BBB

BBB

BB

B

BBB

BBB

CC

C

CCC

CC

C

CCC

CCCC

EEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEEE

EEEEEEEE222222222222

22222222222222222222222222222222222222222

-40

-30

-20

-10

0

10

Masschange,

%

B

BB

B

BBBBBB

CC

C

CCC

CCCC EEEEE

EEEEEEEE

EEEEEEEEEEEEE

EEEEEEEEEEEEEEEEEEEEEEEEEEE 22

2222

2222222

222222222222

2222222222222222222222222222

40

60

80

100

120

1 10 100 1000 2000Relativedynamicmodulus,

%

Number of cycles

360

1291

360

1291

360

1291

B

BBBBBBBE

EEEEE

EEEEEEEEE

EE

-0.10

0.10

0.30

0.50

0.70

Lengthchange,

%

B 2% air, NaCl Soln.

E 4% air, NaCl Soln.

B

B

BB

B

BB

BB

E

EEEE

E

EEE

EE

EE

EEEE

-40

-30

-20

-10

0

10

Masschange,

%

B

BB

BBB

BB

EEEEEEEEE

E

E

40

60

80

100

120

1 10 100 1000 2000Relativedynamicmodulus,

%

Number of cycles

172

616

616

616

172

172

BBBBBBBBBBBBB

CCCCCCCCCCCCCC

EEEE

EEEEEEEEEEE2222

2222222222222

-0.10

0.10

0.30

0.50

0.70

Lengthchange,

%

B 2% air, NaCl Soln.

C 2% air, NaCl+Gypsum Soln.

E 4% air, NaCl Soln.

2 4% air, NaCl+Gypsum Soln.

B

BBBBBB

BBBBBB

B

C

CCCCCCCC

CCCCC

E

EEEE

EEEE

EEE

EEEE

2222

22222

22222222

-40

-30

-20

-10

0

10

Masschange,

%

BB

BBB

BBBBBB

BB

CCC

CCCC

CCCC

CC

EEEE

E

EEE

EEEEEEEE

E

2

2222

2222222222

22

40

60

80

100

120

1 10 100 1000 20Relativedynam

icmodulus,

%

Number of cycles

315

616

315

616

616

315

7/30/2019 PL973[1]

5/8

5

December 1997

R e t u r n To I n d e x

Origins of Chloride Limits forReinforced Concrete

by David Whiting*

The 1977 ACI 201 Limits

In 1977, ACI Committee 201 pub-lished limits on chloride content inconcrete (see Table 1). The first limitshown in the table, 0.06%, is for pre-stressed concrete, and was justifiedbecause of concern that only a smallamount of corrosion on highlystressed strands might potentiallylead to fracture of tendons.

Some studies showed that as littleas 0.15% water-soluble chloride byweight of cement could trigger corro-sion in reinforced concrete that ismoist and where oxygen is abun-

dant.** The limit of 0.10% for con-crete that may eventually be exposedto external chloride appears to be anarbitrary reduction factor to the0.15% limit, recognizing that if chlo-rides will eventually penetrate intothe concrete, it is better to have lesschloride in the concrete initially. Notethat the chloride equivalent of 2% cal-cium chloride is approximately 1%chloride ion by mass of cement, sig-nificantly above the 0.15% limit origi-nally published by ACI 201.

Deliberations Prior toACI 318-83

Until 1982, little documentation wasavailable regarding the adoption ofthe ACI 201-77 limits into the Build-ing Code. A modified set of limitswas developed by a task group ofACI 201 and forwarded to ACI 318(see Table 1). The most significantchange is the relaxation of the limitfor concrete not exposed to externalsources of chlorideincreased to0.30%because the original intent of

ACI 201 appeared to be to limit thechloride to no more than the thresh-old amount for concrete that may beexposed to moisture in service.

In December 1982, ACI 318 pub-lished limits and encouraged formaldiscussion.

ACI 318-83 Limits

The ACI 318 limits, adopted in 1983and currently still in force, are also

given in Table 1. The only significantchange from the ACI 201 TaskGroups recommendations was forconcrete that will remain dry in ser-vice, where the no limit of ACI201-77 and the 1982 Task Group hasbeen replaced by a limit of 1.00%.This allows the use of up to 2.0%flake calcium chloride as an accelera-tor assuming chloride is not presentin other concrete ingredients.

Two years later (1985), the Britishcode stated calcium chloride andchloride-based admixtures should

never be added to reinforced con-crete, prestressed concrete, and con-crete containing embedded metal,thus formally banning the use of cal-cium chloride accelerators in theUnited Kingdom.

At this point, the rationale for thelimit on concrete exposed to chloridewas again questioned, the argumentbeing that one has no control overhow much chloride will eventually

Figure 1. Steel corrosion inducedby chlorides. (S67482)

The subject of chloride limits in con-crete is currently being debated

within the concrete industry. Thepurpose of chloride limits is controlof steel corrosion in reinforced andprestressed concrete (see Fig. 1). Theorigins of these limits provide an ap-preciation of the need for limits, andthe potential need for changes. Cur-rently, the American Concrete Insti-tute (ACI) committees most involvedwith chloride limits are: ACI 318,Standard Building Code; ACI 222,Corrosion of Metals in Concrete; andACI 201, Durability of Concrete.

A distinction should be made be-

tween the limit, which refers to theamount of chloride permitted in newconcrete by code authorities (a crite-rion set by man) and the so-calledthreshold limit for corrosion,which is the actual amount of chlo-ride needed to initiate corrosion undera given set of conditions (determinedby nature). The two are not necessar-ily equal. The limit must be set; thethreshold must be investigated. Ra-tional limits should reflect thresh-olds. Significant research into thethreshold for chloride-induced cor-

rosion has been carried out over thepast few decades. Throughout thisdocument, the commonly acceptedreference for chloride limits, percentof chloride ion by mass of cement,will be used unless noted otherwise.

A Historical Perspective

Early studies (from 1923) indicatedlittle progressive corrosion when cal-cium chloride was added to rein-forced concrete. By the late 1950s and1960s, reports of corrosion of pre-

stressing steel in concrete containingchloride salts were published. In1972, the British banned the use ofcalcium chloride in all prestressedconcrete and in certain reinforcedconcrete. In 1974, the ACI BuildingCode Commentary suggested a limit of400 to 500 ppm for prestressed con-crete, but did not include this limit inthe code body. At the same time, ACICommittee 201 began a series of dis-cussions on chloride limits.

* Senior Principal Engineer, Materials Research

and Consulting, Construction Technology

Laboratories, Inc., Skokie, Illinois.

** Where did 0.15% water-soluble chloride come

from? Apparently, it is from FHWA studies by

Clear (1973) showing that a threshold limit of

0.20% total chloride could induce corrosion of

reinforcing steel in bridge decks. Water-

soluble chlorides are only a portion of the

total: work at the FHWA demonstrated that the

conversion factor could range from 0.35 to

0.90 depending upon the particular

constituents and history of the concrete.

Arbitrarily, 0.75 was chosen: 0.75 x 0.20%

total chlorides = 0.15% water-soluble

chlorides.

7/30/2019 PL973[1]

6/8

Concrete Technology Today

6R e t u r n To I n d e x

ACI 201.2R-77 ACI 201 Task Group ACI 318-83 andrecommendations (1982) ACI 318-95

Guide to Durable Concrete Durability of Concrete Building Code Requirements for Structural Concrete

Maximum water-soluble chloride ion content in concrete, percent by weight of cement

Type of member Limit Type of member Limit Type of member Limit

Prestressed concrete 0.06 Prestressed concrete 0.06 Prestressed concrete 0.06

Conventionally rein- 0.10 Reinforced concrete 0.15 Reinforced concrete 0.15

forced concrete in that will be exposed exposed to chloridea moist environment to chlorides in in service

and exposed to servicechloride

Conventionally 0.15 Reinforced concrete 0.30 Other reinforced concrete 0.30

reinforced concrete that will not be constructionin a moist environ- exposed to chloride

ment but not but may be wet inexposed to chloride* service

Above ground building No Reinforced concrete No Reinforced concrete 1.00

construction where limit that will be completely limit that will be dry orthe concrete will stay for dry, or suitably for protected from

dry corro- protected, in service corro- moisture in servicesion** sion

* Includes locations where the concrete will be occasionally wetted such as kitchens, parking garages, waterfrontstructures, and areas with potential moisture condensation.

** If calcium chloride is used as an admixture, a limit of 2% is generally recommended for reasons other thancorrosion. Using 2% of the usual form (the dihydrate, CaCl2

. 2H2O) results in approximately 1% chloride ion. Includes bridge decks, parking garages, marine construction, and certain industrial plants.

Tested on hardened concrete at an age of 28 days: includes chlorides from water, aggregates, cementitiousmaterials, and admixtures.Note: ASTM did not develop a water-soluble chloride content test method until 1992.

Table 1. ACI Chloride Limits

penetrate into concrete, so its initialchloride content should not matter.The ACI Committee 318 0.15 % limiton concrete exposed to chloride wasa consensus value, and was to beused for the most severe conditionsto prolong the life of the structure.

Recent Developments and theACI 222 Limits

In the early 1990s, an entirely differ-ent concern about chloride limitswas raised, which led to reexamina-

tion of the limits.The new concern is the naturallyoccurring presence of chloride incertain limestone-based aggregates.Evidence suggests that this chlorideis part of the crystal structure ofthese aggregates and cannot becomesoluble under normal conditions ofservice. As an example, a high-qual-ity limestone aggregate used inmany structures in the Chicago area,has been found to contain up to0.06% chloride ions when tested by

normal water-soluble procedures.This equates to 0.18% chloride ionsby mass of cement in a typical con-crete mix (about 356 kg of cementper cubic meter [600 lb per cubicyard]), well above the accepted limitfor prestressed concrete and for rein-forced concrete exposed to chlorides.However, no chloride problems canbe attributed to these aggregatesduring their many years of use.Similar situations have been re-ported for aggregates in EasternCanada.

ACI 222 was presented with a testprocedure utilizing a Soxhlet extrac-tion method (ACI 222.1-96) to leachwater-soluble chloride out of con-crete specimens. It did not, appar-ently, extract the chloride present inthe limestone aggregates, so it wasproposed as a test method for ac-cepting these materials. At the sametime, dissatisfaction with the ACI318 limits led to a proposal for a newset of limits from ACI 222, with sub-sequent submittal to ACI 318.

New ACI 222limits (Table 2)were approvedand published.The old ACI 222R-89 limits for pre-stressed and rein-forced concrete

were 0.08% and0.20% acid-solublechloride, respec-tively (irrespectiveof concrete expo-sure conditions).The new limits al-low one toprogress frommore severe to lesssevere test condi-tions when at-tempting to haveconcrete ap-

proved. The acid-soluble test can berun first; this is thesimplest test forchloride in con-crete. Using thewidely accepted0.75 conversionfactor, the limit onprestressed con-crete (0.08% totalor acid-solublechloride) is effec-

tively the same as ACI 318-95 (0.06%

water-soluble). The limit for rein-forced concrete in wet conditions is0.10% acid-soluble, vs 0.30% water-soluble for 318-95, a considerable dif-ference. The 0.10% limit apparentlyderives from a paper published in1987. There is no separate categoryfor concrete exposed to chloride inservice, as there is in ACI 318, be-cause the new ACI 222 limit of 0.10%is stringent enough for both expo-sures (chloride and non-chloride).

Table 2 shows the ACI 222R-96

limit on reinforced concrete in dryconditions to be less than the ACI318-95 limit for concrete in wet con-ditions. Corrosion does not occur indry concrete, but the committees po-sition is that one cannot ensure thatconcrete will stay dry throughout theservice life of the structure. Concretemembers that have been dry formany years may begin to corrode ifsufficient chlorides are present andthey are suddenly exposed to mois-ture (like wet laundry areas).

7/30/2019 PL973[1]

7/8

7

December 1997

R e t u r n To I n d e x

The remaining columns in Table 2allow a producer to resort to othertest procedures if concrete does notmeet the acid-soluble test limits. Thewater-soluble or Soxhlet procedurescould also be used when borderlineamounts of chloride are present.

Canadian LimitsCanadian limits (see Table 3) are verysimilar to the ACI 318-95 limits. Forprestressed concrete, both codes limitchloride ion to 0.06%; for dry rein-forced concrete, 1.00% chloride ion isallowed. The Canadian StandardsAssociation (CSA) limit of 0.15% ap-plies to concrete exposed to a moistenvironment or chlorides or both,whereas ACI is more lenient (0.30 %)for concrete in a moist environmentthat is not exposed to chlorides.

In addition, CSA A23.2-4B, Sam-pling and Determination of Water-Soluble Chloride Ion Content in Hard-ened Grout or Concrete, acknowledgesthat the test method includes chlo-ride ions in the aggregate that mayunduly influence the test results. Themethod suggests determining thechloride content of the aggregateseparately and adjusting the water-soluble chloride ion content of theconcrete accordingly.

Summary

In the United States the chloride lim-its in ACI 318 are referenced as theyare the limits used in building codes.ACI 222 has developed options forcases where significant quantities ofchloride in aggregates are encoun-tered. Canadian standards directlyaddress the issue of chlorides in ag-gregates. Changes to U.S. standardsare currently being evaluated.

Acknowledgment

The above article (PCA R&D Serial

No. 2153a) is an excerpt taken from aresearch report (Serial No. 2153) con-ducted for PCA by ConstructionTechnology Laboratories, Inc. Thefull report, of the same title, coversthe development of foreign chloridelimits in more detail. The contents ofthis paper reflect the views of the au-thor, who is responsible for the factsand accuracy of the data presented.The contents do not necessarily re-flect the views of the Portland Ce-ment Association.

References

1. American Concrete Institute,Corrosion of Metals in Concrete, ACI222R-96, Farmington Hills,Michigan, March 1997.

2. American Concrete Institute, Pro-

visional Standard Test Method forWater-Soluble Chloride Available forCorrosion of Embedded Steel in Mor-tar and Concrete Using the SoxhletExtractor, ACI 222.1-96,Farmington Hills, Michigan, 1997.

3. Clear, K. C., and Hay, R. E.,Time-to-Corrosion of ReinforcingSteel in Concrete Slabs, V. 1: Effectof Mix Design and ConstructionParameters, Report No. FHWA-RD-73-32, Federal Highway

* ACI 222.1-96Note: Normally concrete materials will be tested for chloride content using either the acid-soluble test described in ASTM C 1152 or water-soluble test described in ASTM C 1218. Ifthe materials meet either of these limits given in the above table, they are acceptable. Ifneither limit is met, then materials may be tested using the Soxhlet test method, whichappears to measure only those chlorides that contribute to corrosion. This permits the use ofsome aggregates that would not be allowed by ASTM C 1152 or ASTM C 1218. If thematerials fail the Soxhlet test, then they are not suitable.

* These limits may be exceeded if the Owner is satisfied that no corrosion problems haveoccurred in other concrete structures made with similar materials and exposed to similarconditions.Note: Quarried carbonate coarse aggregates from the Niagara Escarpment of SouthernOntario contain sufficient chloride ions to cause concrete to exceed the above values,but experience shows this chloride to remain within the aggregate (unavailable to

participate in corrosion). Concrete made with these aggregates may be safely used,provided chloride ion contributed by other concrete components does not cause theconcrete to exceed the above limits.Reference: 1994 edition of CSA A23.1 Concrete Materials and Methods of ConcreteConstruction, Section 15.1.6.1.

Table 3. Chloride Limits in Canada

Application

Maximum water-soluble chloride

ion, percent bymass of cement*

Prestressed concrete 0.06

Reinforced concrete exposed to a moistenvironment or chlorides or both

0.15

Reinforced concrete exposed to neither moistenvironment nor chlorides

1.00

Table 2. New Chloride Limits in ACI 222R-96

Category Chloride limit for new construction

Acid-soluble Water-soluble

Test method ASTM C 1152 ASTM C 1218 Soxhlet

Prestressed concrete 0.08 0.06 0.06Reinforced concrete in w etconditions

0.10 0.08 0.08

Reinforced concrete in dryconditions

0.20 0.15 0.15

*

Administration, Washington, DC,April 1973, 103 pages.

4. Hope, B. B., and Ip, A. K. C.,Chloride Corrosion Threshold inConcrete,ACI Materials Journal,Vol. 84, No. 4, July-August 1987,

pages 306-314.5. National Bureau of Standards,

Effect of Calcium Chloride onCorrosion of Steel in Concrete,Engineering and Contracting, Vol.60, July 1923, page 202.

6. Reading T. J., Chloride ContentLimits Recommended by ACICommittee 201, Concrete Con-struction, Vol. 27, No. 10, October1982, page 777.

7/30/2019 PL973[1]

8/8

Concrete Technology Today

8R t T I d

PL973.01B

Sent to you compliments of:

Intended for decision makers associ-ated with design, management, andconstruction of building projects,Concrete Technology Today is publishedtriannually by the Construction Infor-mation Services Department of thePortland Cement Association.

PUBLISHER'S NOTE: Our purpose is to show various waysof using concrete technology to youradvantage and avoiding problems. Ifthere are problems or ideas readerswould like discussed in future issues,please let us know. Items from thisnewsletter may be reprinted in otherpublications subject to prior permis-sion from the Association.

Direct all correspondence toSteve Kosmatka, Editor

Jamie Farny, Assistant EditorConcrete Technology TodayPortland Cement Association5420 Old Orchard RoadSkokie, Illinois 60077-1083Phone: 847/966-6200 Fax: 847/966-8389E-mail: steve_kosmatka@portcement.orgE-mail: jamie_farny@portcement.org

Printed in U.S.A.

This publication is intended SOLELY for use by PROFESSIONAL PERSONNEL who are competent to evaluate the sig-nificance and limitations of the information provided herein, and who will accept total responsibility for the application

of this information. The Portland Cement Association DISCLAIMS any and all RESPONSIBILITY and LIABILITY for theaccuracy of and the application of the information contained in this publication to the full extent permitted by law.

New Literature

The following publications are now

available. In Canada please direct re-quests to the nearest regional officeof the Canadian Portland CementAssociation (Halifax, Montreal,Toronto, and Vancouver).

Effects of Substances onConcrete and Guide toProtective Treatments, IS001

Substances that attack concrete in-clude: salts and alkalies, oils and fats,coal tar distillates, and solvents andalcohols. Durable concrete mixtures

and proper design are emphasizedcover over reinforcement, drainage,surface preparation and cleaning,and coatings. Sources of products arealso included.

Concrete Slab SurfaceDefects: Causes, Prevention,Repair, IS177

Defects that can occur on concreteflatwork include: blisters, cracking,crazing, curling, efflorescence,delaminations, discoloration, dust-ing, low spots, popouts, scaling, andspalls. Common causes and sugges-tions for avoiding each type of defectare given.

Working Safely with Concrete,MS271

Safety on the job continues to be animportant topic, especially for con-

struction workers who are at riskfrom numerous hazards. Protectionfor head and eyes, back, and skin arehandled with common sense sugges-tions and a knowledge of basicprecautions regarding cement andconcrete.

1996/1997 PCA ResearchReports Summary, MS375

This 6-page document provides a listof PCA research reports from 1996and 1997. The 59 reports are catego-rized by research projects with re-spect to their market or technicalarea. Categories include: EngineeredStructures, Residential, Public Works,Product Standards and Technology,Energy and Environment, and Manu-facturing Technology.

Insulating Concrete Forms forResidential Design andConstruction, SP208

This book is the most comprehensiveresource for designing and buildingconcrete homes with Insulating Con-crete Forms (ICFs). Written to meetthe needs of architects, engineers,and building officials, it focuses onthe technical aspects of ICF home de-sign including design principles, de-tails, formulas, and performance.