After much debate, a tableshowing maximum chlo-ride ion contents for dif-ferent types of concretemembers first appeared

15 years ago in ACI 318-83, “BuildingCode Requirements for ReinforcedConcrete.” That original table hasn’tchanged and will remain unchangeduntil at least 2001 since ACI Commit-tee 318 recently decided not to adoptproposed revisions that would have re-duced most of the maximum values.But the debate continues over propervalues for the limits.

ACI 222R-96, “Corrosion of Metalsin Concrete,” recommends lowermaximum values, basing its case pri-marily on a comparison of the chlo-ride-content limits currently in ACI318-95 with a value called the chloridecorrosion threshold.

Corrosion thresholdThe corrosion threshold is the

amount of chloride ion that must bepresent in concrete, along with mois-ture and oxygen, for corrosion to be-gin. When tested by water-soluble testmethods (see related article on variouschloride test procedures), the thresh-old is generally taken as 0.15% chlo-ride ion by weight of cement. For atypical mixture containing 600 poundsof cement per cubic yard, this thresh-old can be expressed as 0.9 pounds ofchloride ion per cubic yard. Whentested by the acid-soluble test method,the amount is taken as 0.20% chlorideion by weight of cement, or for thesame cement content, 1.2 pounds ofchloride ion per cubic yard.

To understand the rationale for the

proposed revisions, let’s first look atthe limits in the current code require-ments. Table 1 gives the current limitsin Table 4.4.1 of 318-95. These are thevalues many of you must meet whensubmitting a set of mixture propor-tions for a project that specifies chlo-ride limits. We’ve added a column toshow the limits as a percentage of thecorrosion threshold.

Those who advocate changing theselimits base their arguments on whatthey believe to be non-conservativevalues for all types of members exceptprestressed concrete. For instance, ifreinforced concrete will be dry in ser-vice, the allowable chloride ion per-centage exceeds the corrosion thresh-old by a factor greater than six.Proponents for the changes argue thatsuch a high limit doesn’t make sensebecause even reinforced concrete that

will be used under dry service condi-tions may remain moist long afterplacement if it contains lightweight ag-gregate or the surface is sealed.

Table 2 shows limits proposed byACI Committee 222 but not adopted.We have also placed percentage of cor-rosion threshold values in the table.

These proposed limits are more con-servative in that none of them exceeds100% of the corrosion threshold foracid- or water-soluble chloride.

Basis for the decisionThroughout the debate, opponents

of the more stringent chloride limitshave argued that there has been littleevidence of corrosion problems trace-able to the chloride limits in placesince 1983.

Some Committee 318 members wereconcerned that the proposed limits

Chloride limits in the ACI 318 building code requirementsACI committee decides not to make changes … yet

Type of member Maximum water-soluble Percent of water-solublechloride ion (Cl-) content chloride corrosion in concrete, percent by threshold (0.15%weight of cement by weight of cement)

Prestressed concrete 0.06 40

Reinforced concrete exposed to chloride in service 0.15 100

Reinforced concrete that will be dry or protected from moisture in service 1.00 666

Other reinforced concrete construction 0.30 200

Table 1. Water-soluble chloride-ion limits in ACI 318-95

would eliminate the use of certainlimestone aggregates in MidwesternNorth America. These aggregates testhigh in chlorides using either the acid-or water-soluble ASTM test methods,yet the high chloride content doesn’tappear to affect the corrosion perfor-mance of structures made with con-crete containing the aggregates. To ad-dress this objection, ACI Committee222 has adopted a test method that us-es the Soxhlet test method (describedin the related article). Concrete madewith these aggregates doesn’t exceedthe proposed lower chloride limitswhen tested by the Soxhlet procedure.

Another committee concern wasthat the Soxhlet method hasn’t yetbeen adopted by ASTM. This point iscorrect, but change proponents arguethat the ACI version of the testmethod (ACI 222.1) can be used untilan ASTM version is available.

Some ACI 318 members were con-cerned that new, lower limits wouldimpose an economic burden on con-crete producers since they would notbe allowed to use as much calcium

chloride accelerator as in the past.Change proponents, however, argue

that while the use of calcium chloridecould be limited by adoption of lowervalues for the maximum chloride-ionpercentages, many specifications a l-r e a d y ban any admixture with inten-tionally added chlorides when chloride

content is an issue. In other words, en-gineers are currently writing specifica-tion provisions that are far more strictthan ACI 318 code requirements. ACI318 does not allow the use of chlorideor admixtures containing chlorides forprestressed concrete (section 3.6.3),but it has not gone this far for non-

Category of Maximum chloride ion (Cl-) content in concrete,reinforced concrete percent by weight of cement

Acid-soluble Water-soluble Water-soluble(ASTM C 1152) (ASTM C 1218) (ACI 222.1)

Prestressed 0.08 0.06 0.06(pretensioned or post-tensioned) (40%) (40%) (40%)

Non-prestressed, 0.10 0.08 0.08

Wet conditions (50%) (53%) (53%)

Non-prestressed, 0.20 0.15 0.15

Dry conditions (100%) (100%) (100%)

Table 2. Chloride limits for new constructionproposed by ACI Committee 222

How do the different test methods for chlorides differ?Here’s a brief description of the three methods now be-

ing used.■ To test for water-soluble chlorides (ASTM C 1218), a

sample of concrete or aggregate is crushed into a pow-der, and about 10 grams of the powder is mixed withabout 50 milliliters of water. The mixture is boiled for fiveminutes and then allowed to stand for 24 hours before theliquid is filtered and tested for chloride content.

This test method does not detect chlorides that arebound into an aggregate or into hydrated cement. There-fore, the test should theoretically detect only those chlo-rides that are available to cause corrosion. The amountof chloride detected by the water-soluble test is usuallyabout 50% to 75% of the total amount of chloride inconcrete, depending on the nature of the ingredients andthe cement.

Unfortunately, the water-soluble test does detect somechlorides that are not available for corrosion if the aggre-gate particle is intact, but become available as a result ofthe crushing process required for the test. A further draw-back to the method is that test results can be highly vari-able, depending on particle size of the powder, extractiontime and temperature.

■ The test for acid-soluble chlorides (ASTM C 1152) treatscrushed material with nitric acid rather than water. Almostall of the chlorides present in the concrete are detected,whether they are bound or are available to cause corro-sion. Because the test method finds all chlorides, limitsbased upon acid-soluble chloride content are higher thanthose established for water-soluble chloride content.Either of these tests may find chlorides that are present in

some limestone aggregates in Midwestern North America.Based upon service records, we know that the chloridesinside these aggregates don’t contribute to corrosion.

■ A new test, the Soxhlet method (ACI 222.1), has been de-veloped to test these aggregates or to test concrete madewith these aggregates. This test uses chunks of concreteor aggregate instead of crushed and powdered material.The chunks or pieces are treated with boiling water, whichis then tested for chloride content. So far, the test appearsto allow the use of harmless aggregates that would be re-jected by the other test methods. ACI has developed aprovisional standard for this test, and ASTM is currentlyworking on a version of the test method. This test isn’t in-tended to replace the water-soluble chloride test method,but is a fallback test for use when aggregate is suspectedto contain bound chlorides.

Testing for chlorides in hardened concrete and concrete ingredients

prestressed concrete.Change proponents also argue that

ACI 318 doesn’t apply to soil-support-ed slabs, in which chlorides are nowtypically used as accelerators, unlessthe slab transmits vertical loads fromother portions of the structure to thesoil.

Two facts are certain in this chlo-ride-limit argument. First, ACI 318will revisit this issue in the near futureas it prepares for the 2001 edition of itscode requirements. Second, the Soxh-let method will ultimately be acceptedfor use, paving the way for a perfor-mance focus on chloride limits. ■

—TERRY HOLLAND

Say you’re supplying concrete for a parking structure for which ACI 318 per-mits a maximum of 0.15% water-soluble chloride by weight of cement.

How do you demonstrate that your concrete meets the requirement? When the amount of chloride ion present in each ingredient is known, you can

calculate the total amount present in the concrete. However, this is normally notdone because as much as a quarter to a half of these chlorides combine chemical-ly with the hydrating cement paste and aren’t available for corrosion.

Instead, producers send a sample of hardened concrete proposed for a pro-ject to an outside lab for testing at a cost of up to $50 per test. Producers mustindicate the test procedure to be used, depending on which chloride-contentmeasure (water-soluble, acid-soluble or Soxhlet) is specified. They should alsotell the lab how the results are to be reported, since labs may give results inparts per million, pounds per cubic yard or percent by weight of cement. Youcan generally convert the values to a percent by weight of cement if the cementcontent is known. Then simply attach the lab report to your mix submittal.

How to submit chloride-test results

Publication #J980869Copyright© 1998, The Aberdeen GroupAll rights reserved