standard practice for the use of shrinkage- compensating concrete

Keywords: admixtures; aggregates; calcium aluminate; concrete construc-tion; concrete finishing fresh concrete; concretes; curing; drying shrink-age; ettringite; expansive cement concr

223-1

Shrinkage-compensating concrete is used extensively in various types of

con-struction to minimize cracking caused by drying shrinkage Although its

characteristics are in most respects similar to those of portland cement

con-crete, the materials, selecting of proportions, placement, and curing must be

such that sufficient expansion is obtained to compensate for subsequent

dry-ing shrinkage This standard practice sets forth the criteria and practices

necessary to ensure that expansion occurs at the time and in the amount

required In addition to a discussion of the basic principles, methods and

details are given covering structural design, concrete mix proportioning,

placement, finishing, and curing

Keywords: admixtures; aggregates; calcium aluminate; concrete

construc-tion; concrete finishing (fresh concrete); concretes; curing; drying

shrink-age; ettringite; expansive cement concretes; expansive cements; expansive

cement K; expansive cement M; expansive cement S; formwork

(construc-tion); grouts; joints (junctions); mix proportioning; placing; reinforced

con-crete; restraints; shrinkage-compensating concretes; structural design.

CONTENTS Chapter 1—Introduction, p 223-2

1.1—Background1.2—Purpose of shrinkage-compensating concrete1.3—Scope and limits

1.4—Definitions1.5—General considerations1.6—Preconstruction meeting

Chapter 2—Materials, p 223-3

2.1—Shrinkage-compensating cements2.2—Aggregates

2.3—Water2.4—Admixtures2.5—Concrete

Chapter 3—Structural design considerations, p 223-6

3.1—General3.2—Restraint3.3—Reinforced structural slabs3.4—Reinforced slabs on grade3.5—Post-tensioned structural concrete3.6—Post-tensioned slabs on grade3.7—Walls

3.8—Toppings3.9—Formwork

Standard Practice for the Use of

Shrinkage-Compensating Concrete

ACI 223-98

Reported by ACI Committee 223

Gerald H Anderson Terry J Fricks Robert S Lauderbach Carl Bimel Herbert G Gelhardt III Harry L Patterson Richard L Boone James R Golden William F Perenchio

David A Crocker Robert J Gulyas Edward H Rubin Boris Dragunsky Patrick J Harrison Edward D Russell Leo J Elsbernd George C Hoff Joe V Williams, Jr.

Edward K Rice Secretary

Henry G Russell Chairman

ACI 223-98 became effective February 1, 1998.

Copyright  1998 American Concrete Institute.

All rights reserved including rights of reproduction and use in any form or by any means, including the making of copies by any photo process, or by electronic or mechanical device, printed, written, or oral, or recording for sound or visual reproduc- tion or for use in any knowledge or retrieval system or device, unless permission in writ- ing is obtained from the copyright proprietors.

ACI Committee Reports, Guides, Standard Practices,

and Commentaries are intended for guidance in planning,

designing, executing, and inspecting construction This

document is intended for the use of individuals who are

competent to evaluate the significance and limitations of

its content and recommendations and who will accept

re-sponsibility for the application of the material it contains

The American Concrete Institute disclaims any and all

re-sponsibility for the stated principles The Institute shall

not be liable for any loss or damage arising therefrom

Reference to this document shall not be made in

con-tract documents If items found in this document are

de-sired by the Architect/Engineer to be a part of the contract

documents, they shall be restated in mandatory language

for incorporation by the Architect/Engineer

Chapter 4—Concrete mix proportioning, p 223-17

4.1—General

4.2—Concrete proportions

4.3—Admixtures

4.4—Consistency

4.5—Mix proportioning procedures

Chapter 5—Placing, finishing, and curing, p 223-20

Appendix B—Tensile strain capacity, p 223-24

Appendix C—Application of

shrinkage-compensating concrete in post-tensioned

In an earlier state of knowledge report (ACI 223, 1970)

and in the recommended practice (ACI 223, 1977), ACI

Committee 223 described research investigations, structures,

design procedures, and field practices involving expansive

cements used in the production of shrinkage-compensating

concrete The state of knowledge report outlined the theory

and results of studies into the general behavior of such

con-cretes, including the effects of concrete materials,

admix-tures, mixing, and curing conditions, as well as the success of

early field applications

The recommended practice summarized the theoretical

as-pects of the previous report and recommended design

consid-erations, mix proportioning techniques, and placing, finishing,

and curing practices for the utilization of

shrinkage-compen-sating cements and concretes This standard practice applies

the present state of knowledge to the design and field practices

for shrinkage-compensating concrete in such areas as

rein-forced and post-tensioned structural slabs and slabs-on-grade,

walls, toppings, and environmental structures

1.2—Purpose of shrinkage-compensating concrete

Shrinkage-compensating concrete is used to minimize

cracking caused by drying shrinkage in concrete slabs,

pave-ments, and structures Drying shrinkage is the contraction

caused by moisture loss from the hardened concrete It does

not include plastic volume changes that occur before setting

when surface evaporation exceeds the concrete bleeding rate,

and length or volume changes induced by temperature, tural loads, or chemical reactions

struc-The amount of drying shrinkage that occurs in concrete pends on the characteristics of the materials, mix propor-tions, placing methods, curing, and restraint When apavement, floor slab, or structural member is restrained bysubgrade friction, reinforcement, or other portions of thestructure during drying shrinkage, tensile stresses develop.While portland cement concretes normally possess tensilestrengths in the range of 300 to 800 psi (2.1 to 5.5 MPa), dry-ing shrinkage stresses are often large enough to exceed thetensile strength of the concrete, resulting in cracking Signif-icant early age drying shrinkage stresses can occur beforethese tensile strengths are developed Furthermore, because

de-of the probable existence de-of additional stresses imposed byloads, temperature changes, settlement, etc., the inherent ten-sile strength of the concrete cannot be relied on to resistshrinkage stresses The frequency and size of cracks that de-velop in many structures depend on the amount of shrinkageand restraint

Shrinkage-compensating concrete is proportioned so theconcrete will increase in volume after setting and during ear-

ly age hardening When properly restrained by ment or other means, expansion will induce tension in thereinforcement and compression in the concrete On subse-quent drying, the shrinkage, instead of causing a tensilestress that might result in cracking, merely reduces or re-lieves the expansive strains caused by the initial expansion ofthe shrinkage-compensating concrete

reinforce-1.3—Scope and limits

This standard practice is directed mainly toward the use ofshrinkage-compensating concrete in structures (reinforcedand post-tensioned slabs, both on grade and elevated) andpavements Recommendations are included for proportioning,mixing, placing, finishing, curing, and testing based on datapresented in the committee’s previous reports, and on the ex-perience of producers, users, consultants, and contractors.Shrinkage-compensating concrete can be produced usingexpansive cements or expansive components The scope ofthis standard practice is limited to shrinkage-compensatingconcrete made with expansive cements

The recommendations of this standard practice are not plicable to self-stressing expansive cement concretes propor-tioned to produce a prestressed concrete structure for load-bearing purposes Procedures for proportioning, handling,and curing of self-stressing concretes are often radically dif-ferent from procedures for shrinkage-compensating con-cretes used to compensate for normal drying shrinkage

ap-1.4—Definitions

The following terms relating to shrinkage-compensatingconcrete are used in this standard practice:

Expansive cement (general)—A cement that when mixed

with water forms a paste that, after setting, tends toincrease in volume to a significantly greater degree thanportland cement paste; the cement is used to compensate

for volume decrease due to shrinkage, or to induce

ten-sile stress in reinforcement

Expansive cement K—A mix of portland cement, anhydrous

tetracalcium trialuminate sulfate C4A3 (where C = CaO,

A = Al2O3, and = SO3), calcium sulfate (CaSO4), and

lime (CaO) The C4A3 is a constituent of a separately

burned clinker interground with portland cement, ground

separately and blended with portland cement, or

alterna-tively, formed simultaneously with portland cement clinker

compounds during the burning process

Expansive cement M—Interground or blended mixes of

portland cement, calcium-aluminate cement (CA and

C12A7), and calcium sulfate suitably proportioned The

expansive cement M produced in the United States is not

to be confused with the stressing cement (SC) produced

in the former Soviet Republics also from portland

cement, calcium aluminate cement, and gypsum The SC

product is proportioned so that quick-setting,

fast-han-dling, and high early strength are obtained and, therefore, it

is not used in conventional concrete

Expansive cement S—Portland cement containing a large

computed tricalcium aluminate (C3A) content and more

calcium sulfate than usually found in portland cement

Shrinkage-compensating cement—An expansive cement so

proportioned that when combined with suitable amounts

of aggregate and water forms a shrinkage-compensating

concrete or mortar

Shrinkage-compensating concrete—A concrete that, when

properly restrained by reinforcement or other means,

expands an amount equal to, or slightly greater than, the

anticipated drying shrinkage Subsequent drying

shrink-age will reduce these expansive strains but, ideally, a

residual expansion will remain in the concrete, thereby

eliminating shrinkage cracking

Ettringite—A mineral, high-sulfate calcium sulfoaluminate

(3CaO • Al2O3 • 3CaSO4 - 30-32H2O) also written as

Ca6[Al(OH)6]2• 24H2O[(SO4)3• 11/2H2O]; occurring in

nature or formed by sulfate attack on mortar and

con-crete; the product of the principal expansion-producing

reaction in expansive cements; designated as “cement

bacillus” in older literature

Further explanation and definitions can be obtained by

reference to the previous ACI Committee 223 “State of

Knowledge” report (ACI 223, 1970), and “Cement and

Con-crete Terminology,” ACI 116R

1.5—General considerations

The same basic materials and methods necessary to

pro-duce high quality portland cement concrete are required to

produce satisfactory results in the use of

shrinkage-compen-sating concrete The performance of the expansive cement in

minimizing cracking in concrete depends in large measure

SS

S

on early curing In some instances special procedures arenecessary to ensure adequate hydration at the proper time.Consequently, it is essential that early and thorough curingand adequate protection of the concrete be provided Similar-

ly, the mix proportions must ensure adequate expansion tooffset subsequent drying shrinkage Details of the essentialrequirements necessary for successful application are dealtwith in the following chapters The physical characteristics

of the cured shrinkage-compensating concrete are usuallysimilar to other types of concrete The durability of shrink-age-compensating concrete should be judged on the samebasis as portland cement concrete

1.6—Preconstruction meeting

The owner’s representative should be responsible, in eration with the Architect/Engineer and general contractor, forsetting up a preconstruction meeting after all necessary expan-sion tests have been completed, but not less than 1 week be-fore the first concrete is to be placed The purpose of themeeting is to review, discuss, and agree to the proper proce-dures for placing, finishing, and curing the concrete in order

coop-to meet the specifications under the anticipated field tions Responsible representatives of all contractors and ma-terial suppliers, including the manufacturer of the expansivecement, the ready-mix producer, and testing laboratoryshould attend and actively participate in this meeting

CHAPTER 2—MATERIALS 2.1—Shrinkage-compensating cements

2.1.1 Types—The three different shrinkage-compensating

cements described in ASTM C 845 are designated as K, S,and M The expansion of each of these cements when mixedwith sufficient water is due principally to the formation ofettringite

2.1.2 Composition—Seventy-five to 90 percent of

shrink-age-compensating cements consist of the constituents ofconventional portland cement, with added sources of alumi-nate and calcium sulfate For this reason, the oxide analysis

on mill test reports does not differ substantially from thatspecified for portland cement in ASTM C 150, except for thelarger amounts of sulfate (typically 4 to 7 percent total SO3)and usually, but not always, a higher percentage of aluminate(typically 5 to 9 percent total Al2O3) The free lime (CaO)content may also be somewhat higher

The three types of expansive cements differ from each

oth-er in the form of the aluminate compounds from which theexpansive ettringite is developed, as shown in Table 1.The kind of aluminate used influences the rate and amount

of ettringite formation at early ages and thus, the total sion Total potential expansion is governed by the type andamount of aluminates and calcium sulfate and the rate atwhich they form ettringite As with other types of portlandcements, the compressive strength is principally due to thehydration of the calcium silicates

expan-2.1.3 Cement proportioning—These cements are

manufac-tured to produce the proper amount of expansion without versely affecting the concrete quality and retaining the normalrange of concrete shrinkage An important requirement is the

ad-selection of material proportions so that the CaSO4 and the

Al2O3 become available for ettringite formation during the

appropriate period after the mixing water is added

Determi-nation of these proportions is based on the results of

labora-tory tests, outlined in Section 2.1.8, conducted under

standard conditions similar to those used for other portland

cements

2.1.4 Hydration process—Two basic factors essential to the

development of expansion are the appropriate amount of

solu-ble sulfates and the availability of sufficient water for

hydra-tion Ettringite begins to form almost immediately when the

water is introduced, and its formation is accelerated by mixing

To be effective, however, a major part of the ettringite must

form after attainment of a certain degree of strength; otherwise

the expansive force will dissipate in deformation of the plastic

or semiplastic concrete For this reason, mixing more than

re-quired to ensure a uniform mix is detrimental since the

ettringite formed during the prolonged mixing will reduce

the amount available later for expansion With proper curing,

ettringite formation continues during and after hardening,

until either the SO3 or Al2O3 is exhausted

2.1.5 Heat of hydration—The heat of hydration or

temper-ature rise depends on the characteristics and type of the

port-land cement portion In general, the heat of hydration falls

within the range of the variation of the heat of hydration of

the particular portland cement used

2.1.6 Fineness—The surface area determined by air

perme-ability methods (Blaine fineness measured by ASTM C 204)

is not directly comparable to the surface area of portland

ce-ments Shrinkage-compensating cement contains significantly

more calcium sulfate than portland cement Because the

calci-um sulfate grinds more readily than clinker, it contributes a

greater part of the total Blaine fineness value obtained

The specific surface has a major influence on the

expan-sion as well as the early strength of concrete As the surface

area increases above the optimum for a given

shrinkage-compensating cement with a specific calcium sulfate content,

the formation of ettringite is accelerated in the plastic

con-crete Thus, less expansion will be obtained in the hardened

concrete Shrinkage-compensating cement, like portland

ce-ment, produces a higher early strength if it has a higher face area

sur-2.1.7 Handling and storage—These cements are affected

adversely by exposure to atmospheric levels of CO2 andmoisture in a manner similar to portland cements Addition-ally, such exposure can reduce the expansion potential ofthese cements If there is any question as to the expansive po-tential because of method or length of storage and exposure,the cement should be laboratory tested before use

2.1.8 Testing—The expansion characteristics of

shrink-age-compensating cements are determined by measuring thelength changes of restrained 2 x 2 x 10 in (50 x 50 x 254mm) standard sand mortar prisms according to ASTM C

806 These tests measure the expansive potential of the ment and should be used to assess compliance with specifi-cations for the cement Levels of expansion will be differentwhen job materials are used in the concrete mix

ce-2.2—Aggregates

Concrete aggregates that are satisfactory for portland ment concretes can also be used for shrinkage-compensatingcement concretes Good results can be obtained with normal-weight, lightweight, and high-density aggregates meetingthe appropriate ASTM requirements The aggregate typeused, however, has a significant influence on the expansioncharacteristics and drying shrinkage For example, results oflaboratory tests have shown that after a year, a shrinkage-compensating concrete containing river gravel retained a re-sidual expansion of 0.03 percent, whereas concrete madewith the same shrinkage-compensating cement but contain-ing sandstone aggregate had 0.02 percent net shrinkage(Klieger, 1971)

ce-Aggregates containing gypsum or other sulfates may crease expansions or cause delayed expansion or subsequentdisruption of the concrete Significant amounts of chlorides

in-in aggregates, such as found in-in beach sands, tend to decreaseexpansion and increase drying shrinkage For these reasons,

it is recommended that job aggregates be used in the tory trial mix proportioning tests

labora-2.3—Water

Mixing water should be of the same quality as used in land cement concrete (PCA, 1988) If the use of mixer washwater or water containing sulfates or chlorides is contemplat-

port-ed, the water should be used in trial mixes to disclose ble adverse effects on the desired expansion levels ofshrinkage-compensating concrete

possi-2.4—Admixtures

The effect of air-entraining admixtures, water-reducingadmixtures, retarding admixtures, and accelerating admix-tures on the expansion of a specific type or brand of shrink-age-compensating cement may be either beneficial or detri-mental The cement and admixture producers should be con-sulted as to past experience and compatibility of a specifictype or brand of admixture with the cement that is to be used.Data obtained from laboratory testing and field experienceshow that the performance of admixtures is greatly influenced

Table 1—Types of shrinkage-compensating

cements and their constituents

Expansive

cement

Principal constituents

Reactive aluminates available for ettringite formation

K

(a) Portland Cement (b) Calcium sulfate (c) Portland-like cement containing C4A3S

C4A3

M

(a) Portland cement (b) Calcium sulfate (c) Calcium-aluminate cement (CA and C12A7)

CA and C12A7

S

(a) Portland cement high

in C3A (b) Calcium sulfate

C3A S

by the composition of the cement, the ambient temperature,

and the mixing time

In all cases, admixtures should be tested in trial mixes

with job materials and proportions under simulated ambient

conditions Such tests should evaluate the admixture’s

influ-ence on expansion, water requirement, air content,

consis-tency, rate of slump loss, bleeding, rate of hardening,

strength, and drying shrinkage In general:

a Air-entraining admixtures are as effective with

shrink-age-compensating concrete as with portland cement

con-crete in improving freezing and thawing resistance and

scaling resistance in the presence of deicing chemicals

b Some water-reducing and water-reducing and retarding

admixtures may be incompatible with

shrinkage-compensat-ing concrete due to acceleration of the ettrshrinkage-compensat-ingite reaction This

usually has the effect of decreasing expansion of the concrete

c Calcium chloride will reduce expansion and increase

shrinkage of the concrete

d Fly ash and other pozzolans may affect expansions and

also influence strength development and other physical

properties of the concrete

Since the methods of mixing and placing can influence

ad-mixture performance, laboratory results may not always

cor-relate with job results

Further details on the use and influence of admixtures are

given in Chapter 4

2.5—Concrete

2.5.1 Strength—The tensile, flexural, and compressive

strength development after expansion has been completed is

similar to that of portland cement concrete under both

moist-and steam-curing conditions

The water requirement is greater than that of portland

ce-ment concrete for a given consistency Compressive

strengths, however, are at least comparable to portland

ce-ment concrete manufactured from the same clinker and

hav-ing the identical cement content and aggregate proportions

since the extra water is required for hydration of the

expan-sive material As with portland cement concrete, the lower

the water-cementitious material ratio, the greater the

com-pressive strength

2.5.2 Modulus of elasticity—The modulus of elasticity of

shrinkage-compensating concrete is generally comparable

to that of portland cement concrete

2.5.3 Volume change—After expansion, the

drying-shrinkage characteristics of a drying-shrinkage-compensating

con-crete are similar to those of portland cement concon-crete The

drying shrinkage of shrinkage-compensating concrete is

af-fected by the same factors as portland cement concrete

These include water content of the concrete mix, type of

ag-gregate used, cement content, etc The water content

influ-ences both the expansion during curing and subsequent

shortening due to drying shrinkage Fig 2.5.3 illustrates the

typical length change characteristics of

shrinkage-compen-sating and portland cement concrete prism specimens tested

in accordance with ASTM C 878

The minimum recommended amount of concrete sion is 0.03 percent, when measured in accordance withASTM C 878 This is lower than the minimum expansion of0.04 percent specified for a mortar when measured in accor-dance with ASTM C 806 ASTM C 806 uses a larger diam-eter threaded rod, a higher cement content, and a smallercross-sectional area of prism than does ASTM C 878 Theexpansion of a portland cement concrete rarely exceeds 0.01percent when tested using the same test methods

expan-Shrinkage-compensating concrete of relatively high unitwater content may develop some tensile stress at later ages,

as shown in Fig 2.5.3, instead of remaining in compression

2.5.4 Creep—Data available on the creep characteristics

of shrinkage-compensating concrete indicate that creep ficients are within the same range as those of portland ce-ment concrete of comparable quality

coef-2.5.5 Poisson’s ratio—There has been no observed

differ-ence between Poisson’s ratio in shrinkage-compensatingconcrete and portland cement concrete

2.5.6 Coefficient of thermal expansion—Tests have shown

that the coefficient of thermal expansion is similar to that ofcorresponding portland cement concrete

2.5.7 Durability—When properly designed and

adequate-ly cured, shrinkage-compensating concrete made with pansive cements K, S, or M is equally resistant to freezingand thawing, and resistance to scaling in the presence of de-icer chemicals, as portland cement concrete of the same wa-ter-cement ratio The effects of air content and aggregates areessentially the same Recommendations of ACI 201.2Rshould be followed Before being exposed to extended freez-ing in a severe exposure, it is desirable that the concrete at-tain a specified compressive strength of 4000 psi (27.6 MPa).For moderate exposure conditions, a specified strength of

ex-3000 psi (20.7 MPa) should be attained A period of dryingfollowing curing is advisable

Shrinkage-compensating concrete, when properly tioned and cured, has an abrasion resistance from 30 to 40 per-cent higher than portland cement concrete of comparable mixproportions (ACI 223, 1970; Nagataki and Yoneyama, 1973;Klieger and Greening, 1969)

propor-The type of shrinkage-compensating cement and larly, the composition of the portland cement portion can

particu-Fig 2.5.3—Typical length change characteristics of age-compensating and portland cement concretes.

shrink-have a significant effect on the durability of the concrete to

sulfate exposure Shrinkage-compensating cement made

with a Type I portland cement may be undersulfated with

re-spect to the aluminate available and therefore susceptible to

further expansion and possible disruption after hardening

when exposed to an external source of additional sulfates On

the other hand, shrinkage-compensating cements made with

Type II or Type V portland cement clinker, and adequately

sulfated, produce concrete having sulfate resistance equal to

or greater than portland cement made of the same type

clin-ker (Mehta and Polivka, 1975)

2.5.8 Testing—Compressive, flexural, and tensile

strengths should be determined in the same manner and

us-ing the same ASTM methods as for portland cement

con-crete In a shrinkage-compensating concrete, the amount of

expansion is as important as strength Consequently, the

performance of a shrinkage-compensating concrete should

be tested in accordance with ASTM C 878 to determine the

quantity of shrinkage-compensating cement required to

achieve the desired concrete expansion When other methods

(Gaskill and Jacobs, 1980; Liljestrom and Polivka, 1973;

Williams and Liljestrom, 1973) are used, particularly to

de-termine field expansions, they should be correlated with

ex-pansions determined by ASTM C 878 in the laboratory at the

same ages

CHAPTER 3—STRUCTURAL DESIGN

CONSIDERATIONS 3.1—General

The design of reinforced concrete structural elements

us-ing shrinkage-compensatus-ing concrete shall conform to the

requirements of applicable ACI standards At the same time,

adequate concrete expansion should be provided to

compen-sate for subsequent drying shrinkage to minimize cracking

Since the final net result of expansion and shrinkage is

essen-tially zero, no structural consideration need normally be

giv-en to the stresses developed in the concrete during this

process Provision for dead and live loads required by

build-ing codes and specifications will result in at least the same

structural integrity with shrinkage-compensating concretes

as with portland cement concretes However, provisions

shall be made for initial expansive movements

3.2—Restraint

3.2.1 Types of restraint—A resilient type of restraint, such

as that provided by internal reinforcement, shall be provided

to develop shrinkage compensation Other types of restraint,

such as adjacent structural elements, subgrade friction, and

integral abutments are largely indeterminate and may

pro-vide either too much or too little restraint Subgrade frictional

coefficients in the range of 1 to 21/2 have been found

satisfac-tory Values of the coefficient of friction for different bases

and sub-bases are given in ACI 360 High restraint will

in-duce a high compressive stress in the concrete but provide

lit-tle shrinkage compensation Wherever possible, the design

shall, therefore, specify the reinforcement recommended in

Section 3.2.2 Alternatively, the design shall be performed

using the procedures of Section 3.2.3 or other criteria that dress the issues of shrinkage-compensation

ad-3.2.2 Minimum reinforcement and location—Established

engineering design practices for structural elements will mally provide a sufficient amount of steel In some non-load-bearing members, slabs on grade, and lightly reinforcedstructural members, the usual amount of steel reinforcementmay be less than the minimum amount necessary for shrink-age-compensating concretes For such designs, a minimumratio of reinforcement area to gross concrete area of 0.0015shall be used in each direction that shrinkage compensation

nor-is desired Thnor-is minimum nor-is approximately that

recommend-ed by ACI 318 for temperature and shrinkage stresses ever, when procedures outlined in Section 3.2.3 arefollowed, a reinforcement ratio less than the above minimummay be used

How-In structural members, the reinforcement location will bedetermined from design requirements This may result in anover-concentration of reinforcement in one section—partic-ularly in flat plate construction Experience has shown thatwarping caused by concentrated reinforcement is not a prob-lem in structural slabs because dead weight tends to counter-act the warping deflection However, when the location ofthe reinforcement is not determined by structural consider-ations, it shall be positioned to minimize warping For exam-ple, in slabs on grade, where most of the drying occurs in thetop portion, the reinforcement should be placed in the upperhalf of the slab (preferably 1/3 of the depth from the top),while still allowing for adequate cover

3.2.3 Estimation of maximum expansions—When

struc-tural design considerations result in a reinforcement ratiogreater than the recommended minimum, such as bridge

Fig 3.2.3—Estimation of member expansion from prism data.

decks (Gruner and Plain, 1993), or when it is desired to use

less than the minimum reinforcement of Section 3.2.2, the

level of expansion in structural members should be

estimat-ed from Fig 3.2.3 This graph shows the relationship

be-tween member expansion, prism expansion, and percentage

reinforcement when the member and prism are made from

the same concrete and are mixed and cured under identical

conditions The prism expansion test is defined in ASTM C

878 The figure is based on published data (Russell, 1973)

but modified to allow use of the ASTM C 878 test Fig 3.2.3

indicates that for a given prism expansion, a higher amount

of reinforcement will reduce expansion

Fig 3.2.3 may also be used to estimate the required

expan-sion of control prisms to obtain a given expanexpan-sion in a

struc-tural member without external restraint To provide

satisfactory shrinkage compensation, the required expansion

in the reinforced structural member is recommended to be

greater than, or at least equal to, the anticipated shrinkage

Consider a concrete member where the anticipated

shrink-age is 0.025 percent The required expansion for complete

shrinkage compensation is also 0.025 percent If the member

contains 0.5 percent reinforcement, then a restrained prism

expansion of 0.04 percent is required for complete

compen-sation On the other hand, if the restrained prism expansion

is 0.05 percent, then a reinforcement percentage up to 0.75

percent may be used and shrinkage compensation can still be

achieved

For a concrete member containing less than the minimum

reinforcement specified in Section 3.2.3, the same procedure

may be used (Gulyas and Garrett, 1981) Consider an

antic-ipated shrinkage of 0.025 percent in a slab containing 0.1

percent reinforcement Using Fig 3.2.3, the restrained prism

expansion required to offset shrinkage is 0.025 percent A

member expansion in excess of the shrinkage will increase

the likelihood that complete shrinkage compensation will be

obtained

Concrete member expansion is reduced as the amount of

re-inforcement is increased Shrinkage is also reduced, but to a

lesser extent Consequently, to achieve complete shrinkage

compensation, the expansive potential should be higher for

more heavily reinforced members Increased expansion can be

obtained with a higher cement content However, expansion as

measured using ASTM C 878 shall not be greater than 0.1

percent and, in general, should not be less than 0.03 percent

When the amount of reinforcement in a member varies

from area to area, an average expansion shall be used The

lightly reinforced areas will then be overcompensated and

the heavily reinforced areas undercompensated In

deter-mining the anticipated shrinkage, the effects of member

thickness and concrete materials on shrinkage should be

considered An example showing the influence of member

thickness on required levels of expansion is given in

Appen-dix A

3.2.4 Reinforcing steel—Reinforcement should be either

welded wire fabric or deformed bars meeting the

require-ments of ACI 318 Plain bar reinforcement shall not be used

be-cause adequate bond cannot be developed To ensure accurate

positioning, deformed bars placed on chairs or tied to otherfixed rods, concrete supports, or portions of the structureshould be used Where wire fabric is used in lieu of deformedbars, it should be in flat sheets or mats rather than rolls Theuse of rolled wire fabric is not recommended However, ifrolled wire fabric is used, it should be unrolled on a hard flatsurface to remove all curvature before being placed in finalposition The wire fabric may be sandwiched between twolayers of plastic concrete or supported on chairs or blocks.Hooking or pulling the wire fabric off the form or subgradeshould not be permitted Working the wire fabric in from thetop may be permitted if it can be demonstrated that the rein-forcement will be at the correct depth from the top surfacethroughout the slab

3.3—Reinforced structural slabs

3.3.1 Structural design—To provide proper safety factors,

the design shall be based on the strength design provisions ofACI 318 This procedure will avoid consideration of theamount of stress in the reinforcement caused by the expan-sion of the concrete since in the strength analysis, the previ-ous state of prestress does not influence the capacity of thesection In structural members, however, where it is antici-pated that there will be high concrete expansion combinedwith loading at an early age, it is desirable to check that thenet steel stresses caused by the expansion and loading condi-tions do not exceed permissible values

The magnitude of concrete stresses induced by tension inthe reinforcement may be determined as follows:

Consider a reinforced concrete member that expands an

amount εc.

If the areas of concrete and steel are A c and A s, respectively,then

Tensile force in steel = εc E s A s

Compressive force in concrete = εc E s A s

L = length of concrete that is going to expand or shrink.

This relationship is shown graphically in Fig 3.3.1 where

E s is taken as 29 × 106 psi (200 GPa)

As an example, a concrete member containing 0.15 percentsteel that expands 0.10 percent has an induced compressivestress of 43.5 psi (300 kPa), whereas a member that only ex-pands 0.02 percent but contains 2 percent steel has a compres-sive stress of 116 psi (800 kPa) (providing the expansivepotential of the concrete is not exceeded)

The induced compressive stress is a function of the amount

of reinforcement as well as the expansion of the concrete Theinduced compressive stress causes an elastic shortening of theconcrete that for practical reinforcement ratios (up to 1/2 per-cent) is small compared to the errors in predicting the shrink-age As the amount of reinforcement in a member increases,the compressive stress developed in the concrete also increas-

es The expansive strains of a highly reinforced member are

∆

usually low, requiring only a small amount of shrinkage for

the member to return to its original length and then develop

shrinkage strains and concrete tension Strain in concrete and

steel is the most important element to consider when

attempt-ing to counteract shrinkage Concrete strain should stay

ex-panded since shrinkage strains indicate the concrete is going

into tension The range and median of expansions generally

obtained with shrinkage-compensating concretes are shown

in Fig 3.3.1

The Architect/Engineer should specify the minimum level

of expansion for each project The required level of

expan-sion as measured using ASTM C 878 under laboratory

con-ditions should be calculated using the procedure outlined in

Appendix A or by other methods In most situations, the

spec-ified level of expansion should not be less than 0.03 percent

3.3.2 Deflection—The deflection analysis to satisfy load

performance criteria shall be made in the same manner as for

portland cement concrete Any residual compressive stress

caused by expansion will improve the service load

perfor-mance since a higher load is required to produce first

crack-ing Residual compressive stress, however, shall not be taken

into account when calculating deflections

3.3.3 Crack spacing—In two independent investigations,

(Pfeifer, 1973; Cusick and Kesler, 1976) where

shrinkage-compensating concrete was compared with portland cement

concrete, it was observed that the number of cracks were less

in the shrinkage-compensating concrete This occurred for

reinforced concrete specimens loaded in flexure (Pfeifer,

1973) and in direct tension (Cusick and Kesler, 1976) Since

the comparisons in each case were made on specimens with

the same length, the data can be interpreted to mean that

few-er cracks will occur in shrinkage-compensating reinforced

concrete members This has been confirmed in field

observa-tions (Randall, 1980; Rosenlund, 1980)

3.3.4 Cracking moment—The use of

shrinkage-compen-sating concrete does not affect the flexural strength of

rein-forced concrete members However, it does influence the

moment at which flexural cracking occurs Tests (Pfeifer,

1973; Russell, 1980) have shown that members made with

shrinkage-compensating concrete crack at a 15 to 59 percent

higher bending moment than corresponding members made

with portland cement concrete The increase occurs because

shrinkage-compensating concrete members have an induced

concrete compressive stress caused by the expansion This is

equivalent to a mild prestressing Therefore, a

shrinkage-compensating concrete can resist a higher applied moment

before cracking occurs

At later ages, drying shrinkage reduces the concrete

com-pressive stress in shrinkage-compensating concrete By

con-trast, restrained drying shrinkage in a portland cement

concrete will always cause tensile stress Consequently, even

at later ages, shrinkage-compensating concrete can resist a

higher applied moment than portland cement concrete before

cracking occurs

3.4—Reinforced slabs on grade

Because shrinkage-compensating concrete expandsshortly after setting, reinforced slabs on grade with shrink-age-compensating concrete will initially behave differentlyfrom portland cement concrete slabs Certain differentiatingitems are discussed in the following sections

3.4.1 Tensile strains—Portland cement concrete generally

is assumed to possess limited tensile strain capacity fore, reinforcement in slabs on grade is primarily used to con-trol crack widths caused by bending and drying shrinkage.Research (Pfeifer, 1973; Pfeifer and Perenchio, 1973; See-ber et al., 1973; Spellman et al., 1973; Kesler, 1976; Russell,1980) has shown that portland cement concrete has approxi-mately 0.02 percent tensile strain capacity before cracking.Shrinkage-compensating concrete has a higher tensile straincapacity than comparably reinforced portland cement con-crete when the former is allowed to expand and elongate thereinforcement A more detailed discussion of tensile strains

There-is given in Appendix B

3.4.2 Warping—Because of the subgrade restraint and the

internal top steel restraint against the expansion of the crete, differential expansive strains between top and bottomcan be expected During drying shrinkage, expansive strainsare relieved more quickly at the top drying surface than at thesubgrade Research (Keeton, 1979) has shown that the net ef-fect indicates residual restrained expansive strains to begreater at the top surface than at the bottom, so reversed curl-ing conditions develop Warping stresses tend to be counter-balanced by the dead weight of the slab itself

con-Fig 3.3.1—Calculated compressive stresses induced by expansion.

The function of the top reinforcement is to balance the

straint of the subgrade, in addition to providing resilient

re-straint against the expansion

If the subgrade restraint is too low in comparison to

re-straint from heavy top reinforcement, curling may occur

This condition can develop with a heavily reinforced slab on

polyethylene An increase in subgrade friction with sand

placed on top of the polyethylene or placing reinforcement in

the top half to one-third of depth will reduce the curling

If the internal restraint is moved to the bottom third of the

slab or below, warping stresses are not offset by the top

re-inforcement, and cracking may occur upon drying

3.4.3 Isolation joints—Joints used to accommodate

verti-cal movement or horizontal movement shall be provided at

junctions with walls, columns, machine bases, footings, or

other points of external restraint, for example, pipes, sumps,

and stairways In addition to their normal action, these joints

shall be used to accommodate the initial expansion of the

concrete Details of isolation joints are shown in Fig

3.4.3(a) through 3.4.3(g)

Thickness of compressible material shall be estimated

from Fig 3.2.3, as described in Appendix A

Rigid exterior restraint shall not be used since it prevents

expansion of the concrete and a small amount of shrinkage

later will result in negative strains and tensile stress in the

concrete In addition, large forces will be imposed on the

re-straining members Laboratory tests (Russell, 1973) have

shown that rigid restraints result in compressive stresses as

high as 170 psi (1.2 MPa)

Stresses of this magnitude could produce sufficient force

to damage the restraining structure Footings, pits, walls,

drains, and similar items should be protected by isolation

joints to prevent damage during the expansion stage and to

allow the necessary expansive strain to develop

Compress-ible filler strips or joint materials shall be used to control this

behavior

Isolation joints shall be composed of a material that is

compressible enough to deform under the expansive action

of the concrete If too stiff, some rigid asphaltic isolation

materials may act as external restraint and restrict the

expan-sion of the concrete A material with a maximum

compres-sion of 25 psi (170 kPa) at 50 percent deformation according

to ASTM D 1621 or D 3575 should be used Joint materials

meeting ASTM D 994, D 1751, and D 1752 may be too stiff

to allow adequate expansion

Column box-outs may be reduced or eliminated if a

com-pressible material is provided A comcom-pressible bond breaker

wrapped around the column or compressible cardboard

forms brought to floor level have been satisfactory in

permit-ting vertical movement At the same time, the reinforcement

or mesh should be increased locally in the column area

where high stresses are likely to develop This will restrict

the width of any cracks that occur

3.4.4 Construction joints—With the use of

shrinkage-compensating concrete, slab placement patterns of

approxi-mately 20 to 30 ft (6 to 9 m) used with portland cement

con-crete may be enlarged Slabs located inside enclosed

structures, or where temperature changes are small, may beplaced in areas as large as 16,000 ft2 (1500 m2) withoutjoints For areas where temperature changes are larger orwhere slabs are not under enclosed structures, slab place-ments are normally reduced to 7000 to 12,000 ft2 (650 to

1100 m2) The area shall not be larger than a work crew canplace and finish in a day

Building slab sections should be placed in shapes as square

as possible For pavements, which are thicker and moreheavily reinforced than building slabs, successful installa-tions have been made with length-to-width ratios as high as5:1 (Keeton, 1979; Randall, 1980; Williams, 1973) In theseinstallations, joints shall be designed for the anticipated ex-pansion and also become a form of contraction joint Exam-ples of joint details for slab on grade are shown in Fig.3.4.4(a) through 3.4.4(f)

Provision should be made to accommodate differentialmovement between adjacent slabs in the direction parallel tothe joint between the two slabs Differential movement may

be caused by expansion of the shrinkage-compensating crete, differential shrinkage of the adjacent slabs, and ther-mal expansion from heat of hydration of the new slabs Ifprovision is not made for this movement, cracking perpen-dicular to the joint may occur in one or both slabs Two com-mercially available details that allow for movement paralleland perpendicular to the joint and provide vertical load trans-fers are shown in Fig 3.4.4(f) Both details have a patent orpatent pending Other details may be developed to performthe same functions

con-Supporting data should be available showing that the loadtransfer devices are specifically designed for use in concrete,and that the systems will provide essentially immediate ver-tical load transfer with essentially no horizontal restraint.Where applicable, the system should be designed to elimi-nate or minimize potential problems due to corrosion, abra-sion, or repeated loads

Unless specifically required for unusual conditions, the loadtransfer device should not undergo more than 0.01 in (0.25mm) of vertical deformation under the service vertical load

In some cases, these details are different from those usedwith portland cement concrete

Construction joints typically should be designed and tailed as contraction joints to accommodate temperaturemovements, allowing the opportunity for the joint to open,relieving the tensile stress acting on the slab When loadtransfer is required, slip dowels at the joint should be usedrather than deformed bars Tongue and grooved joints may

de-be used when large temperature contractions are not presentand high load transfer is not required

Bonded joints with deformed reinforcement (bars ormesh) passing through the joint may be used, provided thatonly two slabs are locked together in each direction Move-ments from temperature, expansion, and shrinkage strainsmust then be accommodated at the perimeter edges of thetwo slabs

3.4.5 Contraction (control) joints—These joints are

sawed, formed, or otherwise placed in slabs between other

joints Their primary purpose is to induce controlled drying

shrinkage cracking along the weakened planes (joints) With

shrinkage-compensating concrete, larger distances may be

used between contraction joints For exposed areas, a

maxi-mum spacing of 100 ft (30.5 m) between joints is

recom-mended Where the area is protected from extreme

fluctuations in temperature and moisture, joint spacings of

150 to 200 ft (45.7 to 61 m) have been used Contraction

joints may be made in the same way as for portland cement

concrete Normally, contraction joints are eliminated with

shrinkage-compensating concrete except in high stress areas

The larger joint spacing with a shrinkage-compensating

concrete will produce larger movement at the joint This

shall be taken into account when designing load transfer and

joint sealing details

3.4.6 Expansion joints—The location and design of

ex-pansion joints for control of thermal movements are not

changed with the use of shrinkage-compensating concrete

However, joints for thermal movements shall be designed toensure that adequate expansion can take place during the ex-pansion phase An expansion of 0.06 percent is equivalent to

a 100 F temperature change In the event of high load fer, slip plates or dowel bars should be provided, as shown in

trans-Fig 3.4.4(f)

3.4.7 Details—Suggested details of isolation joints,

con-struction joints, contraction joints, door openings, and wallfootings are shown in Fig 3.4.3(a) through 3.4.3(g), and

3.4.4(a) through 3.4.4(f) Additional details using the samebasic principles shall be developed by the Architect/Engi-neer as required

3.4.8 Placing sequence—For a slab on grade, placement

sequence shall allow the expansive strains to occur against afree and unrestrained edge The opposite end of a slab whencast against a rigid element shall be free to move A formededge should have the brace stakes or pins loosened after the

Fig 3.4.3(b)—Circular box-out for deep footing.

Fig 3.4.3(c)—Circular box-out for deep footing.

Fig 3.4.3(a)—Re-entrant corners (pits, trenches, floor

lay-out, truck dock, etc.).

Fig 3.4.3(d)—Circular box-out for shallow footing.

concrete is set to accommodate the expansive action

(Beck-man and Gulyas, 1986)

The placing sequence shall be organized so that the edges

of slabs are free to move for the maximum time possible prior

to placing adjacent slabs At least 70 percent of the maximum

measured laboratory expansion according to ASTM C 878

should occur prior to placing slabs that are not free to expand

on two opposite ends Three examples of placement patterns

are shown in Fig 3.4.8(a) through 3.4.8(c) Checkerboarded

placements should not be used unless a compressible joint

material is placed between the slabs prior to concrete

place-ment The compressible joint materials as described in 3.4.3

shall be used to accommodate the anticipated movements

Before establishing the pour sequence, it is desirable to

have a series of tests made in accordance with ASTM C 878,

based on the proposed concrete mix design A minimum level

of prism expansion of 0.03 percent is recommended for slabs

on grade It is essential that the tested mix design use materialsidentical to those that will be used in construction and betested at the proposed slump that will be used in the field and

as much as possible under the weather conditions anticipated

in the field

3.4.9 Connections—Connections between prefabricated

shrinkage-compensating concrete members or cast-in-placemembers are designed in the same manner as for portland ce-ment concrete The design shall be checked to ensure that nei-ther the expansive strain nor the shrinkage of the adjacentmember produce any undesirable movement

3.5—Post-tensioned structural concrete

3.5.1 Design requirements—Design of post-tensioned

concrete structures using shrinkage-compensating concreteshould meet requirements of ACI 318, and follow recom-mendations of ACI 423.3R

3.5.2 Length changes—All concrete structures are affected

by shrinkage and creep Post-tensioning introduces elasticshortening and additional creep Shrinkage, creep, elasticshortening, and, in some structures, temperature produce neg-ative volume changes in a structure (ACI SP-27) The framedstructural elements become shorter in length and width As aresult of these movements, the supporting columns must bedesigned for higher moments and shears This is particularlytrue for the columns between the foundation and first framedstructural element In this case, the foundations are a fixeddistance apart and the framed elements shorten and pull thecolumns inward The higher moments will require additionalreinforcement and the higher shear will sometimes result in

a larger section

Fig 3.4.3(g)—Slab perimeter tied to wall.

Fig 3.4.3(e)—Wrapped column with stress bars.

Fig 3.4.3(f)—Slab perimeter not tied to wall.

With shrinkage-compensating concrete, the

Architect/En-gineer shall calculate movements caused by expansion and

subsequent shrinkage An example is given in Appendix C

3.6—Post-tensioned slabs on grade

Recommendations for post-tensioned slabs on grade with

portland cement concrete are contained in the

Post-Tension-ing Institute manual (1980)

Because shrinkage-compensating concrete expands during

early hydration, certain modifications can be made in

post-tensioned slabs on grade because of the initial expansion

3.6.1 Restraint—Shrinkage-compensating concrete

ex-pands more in an unrestrained condition than restrained This

tendency for expansion can be restrained externally in

post-tensioned concrete slabs Compression is developed rather

than tension during the initial phases

Comparative stress levels (Russell, 1973) are shown in

Fig 3.6.1 for slabs with different amounts of reinforcement

for externally restrained slabs If adequate restraint is applied

externally, the level may approach 70 to 100 psi compression

within 5 to 10 days after casting But, it is dissipated quickly

unless the post-tensioning cables are stressed before the

con-crete shrinks The data shown in Fig 3.6.1 indicate that theshrinkage-compensating concretes developed higher tensilestresses than the portland cement concrete after an age ofabout 28 days

3.6.2 Subgrade restraint—Restraint of the expansion can

be obtained by the frictional forces imposed by the subgrade.Coefficients ranging from 2.0 for rough-textured sub-basematerials to 0.8 for vapor barrier substrates are commonlyfound to cause tensile stresses in portland cement concretesupon shrinkage For shrinkage-compensating concrete, thesame subgrade friction working against the expansion willproduce compression in the concrete, thereby offsetting thetensile stresses

Because of the compression on the concrete, there is noneed to post-tension the slab at an early age to induce earlycompression This eliminates the phased post-tensioning or

Fig 3.4.4(b)—Doweled contraction joint.

Fig 3.4.4(c)—Control joints at columns.

Fig 3.4.4(d)—Door opening.

Fig 3.4.4(a)—Keyed contraction joint.

partial post-tensioning required in many slab on grade

appli-cations to prevent cracking with portland cement concrete

3.6.3 Placing sequence—Unlike conventionally

rein-forced slabs, post-tensioned shrinkage-compensating

con-crete slabs require restraint to provide early compression in

the concrete The formed edge should be reasonably stiff to

provide adequate restraint against the expansion Placement

of subsequent concrete after form removal will also provide

some restraint to induce compression

3.6.4 Design implications—Research (Nagataki and

Yoneyama, 1973) has shown that compression developed

against external or subgrade restraint can be effectively

uti-lized to reduce the loss of post-tensioning force due to

sub-grade friction with the use of shrinkage-compensating

post-tensioned concrete To be effective, the mechanical

pre-stressing force must be introduced before the

shrinkage-compensating concrete starts to shrink, generally within 7

days The compression developed by the external restraint

can be utilized in reducing the total mechanical force applied

to the slab by using less stressing steel (Nagataki and

Yon-eyama, 1973)

3.6.5 Details—Stiffened bulkheads should be used at

con-struction joints unless metal forms are used with frequently

pinned stakes

Pour strips used to relieve shrinkage in portland cement

concrete slabs should not be used Construction joints are

suggested

Typical details are shown in Fig 3.6.5 These details can be

modified by the Architect/Engineer for specific applications

To accommodate movements parallel to the construction

joint, the details shown in Fig 3.4.4(f) may be used

3.7—Walls

3.7.1 Placing sequences—The sequence of placing

shrinkage-compensating concrete in walls is very

impor-tant The sequence of placing should allow one edge of the

wall in each direction to remain free to expand The top ofthe wall is free to expand in the vertical direction At leastone vertical construction joint shall remain free to allow ex-pansion in the horizontal direction Free edges are needed

so the concrete can expand without rigid external restraint.The method of checkerboard placement of walls (cast a wallsection, skip a wall section, cast a wall section) that has beenused for portland cement concrete is not recommended forshrinkage-compensating concrete unless provision is made

to allow expansion of the wall The checkerboard methodwhere concrete is cast directly against adjacent concreteleaves no space for the concrete to increase in length Theconcrete would build up compressive stresses, which would

be dissipated quickly due to the negative length changecaused by shrinkage As a result of insufficient expansion,shrinkage cracks may occur The following paragraphs ex-plain three different possible placing sequences that can beused successfully with shrinkage-compensating concrete

In each case, the compressible filler strip joint should bemade from a material that has a maximum compressibility

of 25 psi (170 kPa) when the material is reduced to 50 cent of its original thickness

per-Fig 3.7.1(a) shows a plan of a rectangular tank with a ing sequence leaving the corners open until all side wallshave been cast

cast-Fig 3.4.4(f)—Joint details for movement parallel and zontally perpendicular to the joint.

hori-Fig 3.4.4(e)—Integral footing for partition walls.

The recommended sequence for placing concrete wall

sec-tions is also shown The sequence may be varied as long as

no wall section is restrained between two previously cast

sec-tions and the corners are cast last Each cast section shown is

based on a maximum length of 150 ft (46 m) The minimum

time period, generally 3 to 7 days, between casting adjoining

sections should be sufficient to ensure that adequate

expan-sion or volume change can take place before the next section

is cast The size of the corner section shall be just large

enough to develop hooks and laps as required by design This

figure shows the preferred sequence of wall placement

Fig 3.7.1(b) shows a plan with continuous sequential

cast-ing of a tank wall with all placements stoppcast-ing short of the

corners A compressible filler strip joint is placed betweenthe last section and the existing first corner section The planshows the recommended sequence for placing concrete wallsections with continuous casting The length of each castsection shown is the same as stated for Fig 3.7.1(a) except at

Fig 3.6.5—Details for post-tensioned slab on grade.

Fig 3.4.8(a)—Center adjacent slab placement pattern.

Fig 3.4.8(b)—Center rotation slab placement pattern.

Fig 3.4.8(c)—Lag slab placement pattern.

Fig 3.6.1—Stress in slab with full external restraint sell, 1973).