Chloride Resistance of Concrete · Report Cement Concrete & Aggregates Australia Cement Concrete & Aggregates Australia

Re

po

rt Cement Concrete & Aggregates Australia

Chloride Resistance of ConcreteJune 2009

2 Chloride Resistance of Concrete

Contents

1 INTRODUCTION 3

2 CHLORIDE-INDUCED STEEL CORROSION 3

3 FACTORS AFFECTING CHLORIDE RESISTANCE 4

3.1 General 4

3.2 Factors Relating to Concrete 5

3.3 Factors Relating to the Structure 7

4 CHLORIDE TRANSPORT 7

4.1 Transport Mechanisms 7

4.1.1 Diffusion 8

4.1.2 Capillary suction and absorption 9

4.1.3 Permeability 10

4.1.4 Migration 11

4.1.5 Absorption and desorption 11

4.1.6 Mixed modes 11

4.2 Electrochemical Properties 12

4.3 Marine Exposures 12

4.4 Relative Severity of Exposure Conditions 13

5 CHLORIDE RESISTANCE TESTS 14

5.1 Indirect Measures 15

5.1.1 Cement type and water-cement ratio 15

5.1.2 Compressive strength 17

5.2 Direct Measures 18

5.2.1 Chloride diffusion coefficients 18

5.2.2 Absorption, sorptivity, ISAT and AVPV 20

5.2.3 Coefficient of permeability 26

5.2.4 Rapid chloride permeability test 27

5.3 Electrochemical Properties 29

6 COMPARATIVE PERFORMANCE DATA 29

7 CHLORIDE RESISTANCE ENHANCING MEASURES 31

7.1 Sealers 31

7.2 Corrosion Inhibitor 31

7.3 Other Admixtures 32

8 CONCLUSIONS 33

9 REFERENCES 34


1 INTRODUCTION

Corrosion of steel reinforcement in concrete is the most common problem affecting the

durability of reinforced concrete structures. Chloride-induced corrosion is one of the main

mechanisms of deterioration affecting the long-term performance of such structures1.

Concrete provides physical and chemical protection to the reinforcing steel from penetrating

chlorides which may cause steel depassivation leading to increased risk of steel corrosion.

The chloride resistance depends on the permeability of the concrete and the thickness of

cover to the reinforcement. The integrity of the concrete cover under service load, in terms of

cracking and crack width, also influences the resistance to penetrating chlorides. Corrosion

of steel reinforcement is an electrochemical process. Hence electrochemical properties of

concrete, such as resistivity, are important inherent properties affecting the corrosion rate of

reinforcing steel.

Metha2 reconfirmed from a review of case studies that it is the permeability of concrete, rather

than its chemistry, which is the key to overall durability. The causes of high permeability are not

limited to poor concrete proportion but poor concreting practice, such as incomplete mixing,

inadequate consolidation and curing after placement, insufficient cover to reinforcing steel,

and badly constructed joints. In service, concrete may exhibit various forms of cracking for

reasons such as settlement, premature loading, overloads, and repeated impact. To obtain

long-term durability of concrete marine structures, the control of concrete cracking in service

through proper mix proportioning and concreting practice is of as much importance as the

control of concrete permeability.

This report discusses the various factors affecting chloride resistance of concrete, mechanisms

of chloride transport, related test methods and performance specifications. It also assesses

additional measures to enhance the chloride resistance of concrete.

2 CHLORIDE-INDUCED STEEL CORROSION

Steel reinforcement embedded in concrete is inherently protected against corrosion by

passivation of the steel surface due to the high alkalinity of the concrete. When a sufficient

amount of chlorides reaches the steel reinforcement it permeates the passivating layer and

increases the risk of corrosion. The resistivity of concrete can also be reduced, affecting the

corrosion rate of the steel.

Anodicprocess

Cathodi cprocess

Fe 2e-

4e-

2e- 2(OH-)

4(OH-)

Fe++

2Fe++

FeO(H O)2 x

½O2

O2

H O2

H O2

+ +

+


For use in reinforced or prestressed concrete structures the chloride concentrations in

cements, mixing water, aggregates, and admixtures are strictly controlled, and the maximum

permissible concentrations are given in building standards. AS 1379 3 restricts the acid-soluble

chloride of fresh concrete to 0.8 kg/m3 of concrete.

In most cases, however, excessive amounts of chloride in concrete originate from external

sources. The penetration of chlorides into the concrete occurs by various transport

mechanisms depending on the exposure conditions. There are significant amounts of chlorides

in seawater but chlorides are more limited in groundwater and soil. In many countries de-icing

salts, used to combat the build-up of snow and ice on transport infrastructures, are the greatest

source of chlorides.

In seawater, chlorides usually pose a greater threat to steel in concrete than sulfates do to

concrete as calcium sulphoaluminate or ettringite (the expansive reaction product of sulfate

and tricalcium aluminate in the cement) is more soluble in the presence of chloride and hence

does not cause the disruptive expansion.

Portland cement reacts with sodium chloride to form chloroaluminates or Friedel’s salt, thus

immobilizing the chloride and reducing the free chloride ions available to depassivate the steel.

The results of a 34-year long-term exposure of plain and reinforced concrete beams in a tidal

seawater exposure in Los Angeles Harbour in California in 1959 and 19614, where freezing

and thawing does not occur, found that all plain concrete mixtures (stored at approximately

mean tide) display excellent resistance to seawater attack, regardless of cement composition,

water-cement ratio, cement content, the use of SCM, and method of curing. However, severe

cracking due to corrosion of embedded reinforcing steel developed in some beams stored

above high tide, while only minor or no cracking developed in companion beams stored in

seawater near mean tide level. The most severe corrosion-induced cracking occurred in

concrete with the highest w/c of 0.49 and least cover of 25 mm. The relatively greater degree

of steel corrosion in beams stored above high tide is attributed to the greater availability of

oxygen at the reinforcing steel surface. Corrosion-related distress was found to be sensitive to

concrete cover and water-cement ratio. Prestressed steel appeared to be no more vulnerable

to galvanic corrosion than ordinary deformed reinforcing bars. The nickel and painted-on

epoxy coatings appeared to provide little, if any, additional protection against corrosion.

3 FACTORS AFFECTING CHLORIDE RESISTANCE

3.1 General

In Australia, a large majority of structures are built either near the coast (where they are

exposed to airborne chlorides) or in direct contact with seawater. The durability of reinforced

and prestressed concrete structures is thus highly dependent on the resistance of concrete to

chloride penetration. The physical resistance of concrete to chloride penetration is influenced

by factors relating to the concrete itself, such as the porosity of concrete and interconnectivity

of the pore system; and to factors relating to the concrete structure such as the stress

conditions and the integrity of the cover.

The total chlorides content, being the combined free and bound chlorides, does not give

a realistic indication of the risk of corrosion to the reinforcement. It does, however, give an

assessment of the long-term risk to structures exposed to chlorides which are also prone to

carbonation under certain exposure conditions. Carbonation results in lower pH, enabling the

chlorocomplexes to release free chlorides.


3.2 Factors Relating to Concrete

External chlorides penetrate into the interconnecting pores in concrete as bulk liquid by

convection, and chloride ions diffuse further into the saturated pore system. Diffusion is

controlled by concentration gradients of the free chlorides; thus the capacity of the concrete

to physically adsorb and to chemically react with chloride ions affects the free chloride ions

concentration in the concrete.

The chloride resistance of concrete is thus highly dependent on the porosity of concrete in

terms of pore size, pore distribution and interconnectivity of the pore system. The porosity of

concrete is determined by:

n the type of cement and other mix constituents;

n concrete mix proportions;

n compaction and curing.

The type of cement influences both the porosity of the concrete and its reaction with chlorides.

The porosity of concrete is highly dependent on the water-cement and aggregate-cement

ratios whereas the type and amount of cement affect the pore size distribution and chemical

binding capacity of the concrete. The influence of cement type and water-cement ratio on the

chloride resistance of concrete, measured in terms of effective diffusion coefficient, is shown in

Figure 3.1.

Figure 3.1 Effect of water-cement ratio and cement type on the chloride resistance

of concrete (CSIRO6)

Water/cement

Dif

fusi

on

co

effi

cien

t, D

e.36

5 (1

0–12

m2 /

s)


Figure 3.2 Variation of porosity with depth of concrete slabs cured using

different methods (Gowripalan et al 5)

The porosity or permeability of insitu concrete is highly dependent on the degrees of

compaction and curing during placing and the early life of concrete respectively. Curing

greatly affects the porosity of the concrete cover which protects the steel from chloride-

induced corrosion. The effectiveness of various curing regimes on the porosity of concrete is

illustrated in Figure 3.2.

It has been found that curing can improve the chloride resistance of concretes, measured

in terms of water sorptivity, to different degrees depending on the type of cement. The

effectiveness of early curing on sorptivity is shown in Table 3.1.

TAbLE 3.1 Influence of curing on sorptivity of various Grade 50 concretes (Khatri et al 6)

RTA sorptivity (mm in 24 hours)

Type of cement1-day sealed 27-day air-cured

7-day sealed 21-day air-cured

7-day wet 21-day air-cured

GP 14 17 5

GB1 33 27 10

GB2 40 28 1

GB3 >50 35 0

In the past, the durability property of concrete was specified by maximum water-cement ratio

and minimum cement content. With the availability of a range of chemical admixtures and

supplementary cementitious materials (SCM), it has become increasingly difficult to specify

durability prescriptively. Performance-based specifications are becoming more common and

are quantified by the use or adaptation of test methods that measure the principal chloride

transport mechanism for specific exposure conditions.

Depth from surface (mm)

Po

rosi

ty (

%)


3.3 Factors Relating to the Structure

Concrete structures in service are subjected to varying stress conditions resulting in both

macro and microcracking. Flexural cracks are expected and are controlled by limiting the

stress in the steel and/or spacing of reinforcement. Thermal cracking is controlled by limiting

the differential temperature and restraint conditions during casting and in service7.

Research under the Concrete in the Oceans programme8 found that while macrocrack

width may influence corrosion in the short term, the influence decreases with time and that

in the long term the influence of crack width on corrosion is likely to be insignificant. There

are two possible reasons for the reduced influence of crack width with time: unfavourable

electrochemical process and healing of cracks. Research by Japan Port and Airport Research

Institute9 revealed that narrow cracks (<0.5 mm) were healed during 15 years of exposure of

specimens in a tidal pool created by using seawater directly from the sea irrespective of the

cement types. It was noted that the deposits, found at the root of the crack, created a highly

alkaline media and thereby stopped the corrosion reaction over the steel bars at the root of

the crack. It has been reported10 that chloride concentration increases with crack width being

more significant for crack width exceeding 0.5 mm. Cracks 0.3 mm or less in width undergo

self-healing, especially in seawater when magnesium and carbonates combined with calcium

hydroxide to form insoluble products of magnesium hydroxide (brucite) and calcium carbonate

(aragonite) and subsequently fill the cracks.

Concrete under tension has been found to allow greater chloride penetration than the same

concrete under compression11. This was attributed to the microcrack at the aggregate-paste

interface in the tension zone which expedites chloride diffusion. On the contrary, a reduction in

the porosity of the concrete in the compression zone may impede chloride diffusion.

In practice, structural concrete with crack widths within a specified limit is considered to have

the same chloride resistance as stress-free concrete. The effect of stress condition can be

factored into the chloride resistance of structural concrete such as the diffusion coefficient

used to model service life12.

The integrity of the cover concrete depends on structural design, attention to detailing to

control crack width and the compaction the concrete around the steel reinforcement. With such

measures in place, the resistance of concrete to chloride penetration will largely depend on the

quality of the concrete and to a lesser extent on the stress condition of the structural concrete.

4 CHLORIDE TRANSPORT

The transportation of chloride ions into concrete is a complicated process which involves

diffusion, capillary suction, permeation and convective flow through the pore system and

microcracking network, accompanied by physical adsorption and chemical binding13. With

such a complex transportation process, it is necessary to understand individual transport

mechanisms and the predominant transport process in order to pinpoint the appropriate

method for quantifying the chloride resistance of concrete.

4.1 Transport Mechanisms

Individual transport mechanisms and the associated test methods are described in this

section. There is usually more than one transport mechanism involved. Most performance-

based specifications are based on the predominant transport mode applicable to the specific

exposure condition.


Table 4.1 summarises the primary chloride transport mechanisms applicable to structures in

various exposure conditions.

TAbLE 4.1 Primary chloride transport mechanism for various exposures

Exposure Example of structuresPrimary chloride transport mechanism

Submerged Substructures below low tide. Diffusion.

Basement exterior walls or transport tunnel liners below low tide. Liquid containing structures.

Permeation, diffusion and possibly Wick’s action.

Tidal Substructures and superstructures in the tidal zone.

Capillary absorption and diffusion.

Splash and spray Superstructures about high tide in the open sea.

Capillary absorption and diffusion. (Also carbonation)

Coastal Land based structures in coastal area or superstructures above high tide in river estuary or body of water in coastal area.

Capillary absorption. (Also carbonation)

4.1.1 Diffusion

Diffusion is a transfer of mass of free molecules or ions in the pore solution resulting in a net

flow from regions of higher concentration to regions of lower concentration of the diffusing

substance. This mode of transport operates in fully saturated media such as fully submerged

concrete structures. For porous material like concrete, the diffusion coefficient, D, is the

material characteristic property describing the transfer of a given substance driven by

concentration gradient.

In steady-state chloride diffusion, the effective driving force is the gradient of the free chloride

ions in the pore solution; the diffusion coefficient is referred as Dfree or Df. Df can be determined

from the difference in concentration of chloride ions in the two cells separated by the concrete.

This type of test is used only in research as it is extremely time consuming and thin sections of

mortar or concrete are tested.

In the transient or non-steady-state diffusion process, the mass balance equation describes the

change of concentration in a unit volume with time. This can be accounted for by Fick’s second

law of diffusion:

∂C = D ∂2C (1)

∂t

∂x 2

An analytical solution to Equation 1 is given as:

Cx = Cs [1 – erf ( x ) ] (2)

2 √Dt

where Cx is the chloride concentration at distance x,

x is the distance from the exposed surface,

Cs is the surface chloride concentration,

D is the chloride diffusion coefficient,

t is the exposure time, and

erf is the error function.


The diffusion coefficient, D, is determined from the best-fit curve represented by Equation 2 for

the measured chloride profile.

For concrete, the diffusing chloride may be partially immobilized due to chemical interaction or

physical adsorption. Then the balance equation must be extended by a sink, s:

∂C = ∂ (D ∂C ) – s ∂t

∂x

∂x

Immobilization of diffusing material is of considerable importance for the experimental

determination of diffusion coefficients. As long as the binding capacity of the test specimen

is not yet exhausted, the net flow of diffusing material appears to be low. Then, the diffusion

coefficient is underestimated. In cases where the sink, s, is not taken into account an

‘apparent’ diffusion coefficient, Dapparent or Da, may be deducted from the acid-soluble chloride

profile from the experiments, which then depends on time, t.

ASTM C 155614 is a standard method for determining the apparent chloride diffusion

coefficient of cementitious mixtures by bulk diffusion. Precision estimates are given for both

apparent diffusion coefficient and surface chloride concentration. It refers to a precision data

source from an interlaboratory study of NORDTEST NT Build 44315.

In a non-steady-state diffusion process, the gradient of the free chloride ions in the pore

solution is the effective driving force; the diffusion coefficient derived from measured free

chloride concentration profiles is the effective diffusion coefficient, Deffective or De. The effective

diffusion coefficients can also be derived from measured concentration profiles in existing

structures and serve as a basis for estimates on the future progress of the chloride penetration.

4.1.2 Capillary suction and absorption

Capillary absorption is the transport of liquids due to surface tension acting in capillaries. It

is influenced by the viscosity, density and surface tension of the liquid and the pore structure

(radius, tortuosity and continuity of capillaries) and surface energy of the concrete. Chloride

can therefore be transported in the liquid solution.

When concrete in not in permanent contact with a liquid such as in the tidal zone, a non-steady

state transport of the liquid prevails. In this case, the amount of liquid absorbed at the surface

of the concrete as well as the amount of liquid transported at any distance from the surface is a

function of time.

For short-term contact of the concrete with a liquid, the velocity of the take-up is referred to as

initial absorption rate, a, given by the mass of the liquid absorbed per unit area and a function

of the contact time, t:

a = ∆m

A x f (t n)

where a = absorption rate (g/m2 sn)

∆m = take-up of liquid (g)

A = area in contact with water (m2)

f (t n) = time function

A time function = √t is usually valid but functions other than √t may also be valid.


In practice, non-steady state capillary absorption is the mode of transport measured as

sorptivity or initial surface absorption test (ISAT). In Australia, a variation of sorptivity has been

used. CSIRO measures the take-up of liquid per unit area as a depth. This technique has been

adopted by the Roads & Traffic Authority of New South Wales16.

Capillary suction may also develop into a steady-state transport phenomenon if suitable

boundary conditions are kept constant over time. A concrete member in contact with water

on one side will take up the liquid by capillary suction. If the evaporation of the water at the

opposite side is in equilibrium with the take-up of water, capillary suction transports the water

over a certain section of the concrete member in a steady-state process.

Capillary absorption is an important mechanism with respect to the ingress of chlorides into

concrete. Non-saturated concrete in contact with a salt solution will take up the salt solution by

capillary forces; chlorides thus penetrate into the concrete much faster than by diffusion alone.

Simultaneously, chlorides are transported by diffusion to increase the depth of penetration.

4.1.3 Permeability

Permeability is a measure of the flow of liquids or gases through a porous material caused by a

pressure head. The permeability of concrete depends on the pore structures and the viscosity

of the liquids or gases. This mode of transport is applicable for concrete structures in contact

with liquid under a pressure head, such as in liquid-retaining structures. Dissolved chlorides

and gases are therefore transported by convection with the permeating water into concrete.

Water permeability can be determined in two ways: steady-state and non-steady-state water

permeability. A coefficient of permeability is a material characteristic which is obtained in

saturated flow which occurs under a steady-state flow when a constant flow rate is established.

It is expressed as the volume of water per unit area of surface per unit time, flowing through

a concrete under a constant pressure head and at a constant temperature. Thus water

permeability has a unit of m3/(m2·s) or m/s. The permeability of concrete varies between

10–16 and 10–10 m/s. In a saturated sample the water flow in the pores due to hydraulic

pressure can be described by D'Arcy’s law:

Q = Kw ∆PA / l

where Q = water flow (m3/s)

Kw = coefficient of water permeability (m/s)

∆P = pressure difference (m) across the sample

A = surface area of sample (m2)

l = thickness of sample (m)

The time to obtain steady-state flow depends on the pressure applied as well as the

composition, size and degree of saturation of the sample. A period of 2 days to 2 weeks

may be required for concrete with a water-cement ratio of 0.75 to 0.35 for vacuum-saturated

samples under a 3.5-MPa pressure head. A coefficient of variation for repeatability of around

10% has been found for such test.

The CSIRO constant flow water permeability test17 is based on the steady-state saturated

water flow through a 50-mm-thick concrete sample under a pressure gradient of 0.7 MPa.

For concrete with very low water permeability less than 1.3 x 10–13 m/s, a non-steady-state

water permeability is preferred by measuring the depth of water penetration under a constant

pressure. The coefficients of water permeability determined by both methods have been found


to correlate well and a guide to the selection of the more suitable method was developed

based on the 28-day compressive strength and the age of concrete at testing17. The guide is

that the constant flow method is preferred if:

2.3 x (T )2 + 1.1 x (Fc28)2 ≤10,400

Where T is the age of the concrete when tested (days), and

Fc28 is the 28-day compressive strength in MPa

The DIN 1048 is one version of the penetration method which reports the water penetration

depth as a simplified indicator of water permeability.

4.1.4 Migration

Migration is the transport of ions in electrolytes due to the action of an electrical field as

the driving force. In an electrical field, positive ions will move preferentially to the negative

electrode and negative ions to the positive one. Migration may generate a difference in

concentration in a homogeneous solution or may provoke a species flux in the direction of

concentration gradients. This mode of transport may occur accidentally when there is a stray

current leakage, or intentionally in concrete rehabilitation techniques.

One popular technique which measures chloride ion migration or the electrical conductance of

concrete is the AASHTO T277 or ASTM C120218, developed by D Whiting. In these methods,

a potential difference of 60 V(DC) is maintained across the ends of 51-mm-thick slices of

a concrete cylinder, one of which is immersed in a sodium chloride solution, the other in a

sodium hydroxide solution. The amount of electrical current passing through the concrete

during a six-hour period is measured and the total charge passed, in coulombs, is used as an

indicator of the resistance of the concrete to chloride ion penetration. ASTM C1202 provides

precision in terms of single-operator and multi-laboratory precision.

The technique has also been used in combination with others to accelerate the transport of

chloride ions such as in the CTH rapid method19.

4.1.5 Adsorption and desorption

Adsorption is a fixation of molecules on solid surfaces due to mass forces in mono- or

multi-molecular layers. Desorption is liberation of adsorbed molecules from solid surfaces.

Adsorption of chlorides is controlled by the micropore structure and the characteristics of

the hydrated products, and in particular the specific surface area and surface charge of the

pore walls.

4.1.6 Mixed modes

The transport mechanism depends on the boundary conditions as well as on the moisture

state and its distribution in the concrete element. Pure permeation of a chloride solution as well

as pure diffusion of chloride ions will prevail only for a moisture-saturated concrete in which

no capillary forces can be active. If dry or non-saturated concrete is exposed to a chloride

solution, however, capillary absorption is the dominant mechanism. Nevertheless, small

hydraulic pressure heads can support the ingress by permeation, and the diffusion of ions

simultaneously carries the ions also into narrow pore spaces where no capillary flow any longer

occurs. Except for concrete elements that are continuously submerged in seawater, these mixed

modes of chloride transport obviously prevail in most cases for concrete structures in service.


The relative importance of various modes of chloride transportation has been evaluated by

examination of the sensitivity of key parameters on the chloride concentration profile20. A

transport model based on diffusion, convection and chloride binding was used. It was found

that the chloride penetration is sensitive only to the permeability coefficient for concrete at

higher permeability of the order of 1.0 x 10–13 m/s (applicable to concrete with w/c of around

0.35) and at great cover depth, whereas for low permeability of the order of 2.0 x 10–16 m/s,

the chloride ingress is virtually dependent only on the diffusion coefficient. The range of the

diffusion coefficient used was 0.33-2.53 x 10–12 m2/s. Other parameters considered include the

diffusivity decay constant, temperature of exposure, critical chloride threshold, surface chloride

concentration and depth of cover. It was found that the chloride ingress is extremely sensitive

to the diffusion decay constant, temperature and depth of cover, and is linearly dependent on

critical threshold.

4.2 Electrochemical Properties

Some of the electrochemical properties used in studying corrosion of steel reinforcement in

concrete include resistivity and conductivity of concrete, and half-cell potential of reinforced

concrete. The resistivity of concrete measures the resistance of concrete against the flow of

electrical current and is highly dependent on the moisture condition and the ionic nature of the

pore solution. Resistivity therefore reflects the chloride resistance of concrete. The resistivity

of concrete increases with time due to continued hydration of cement, and will also decrease

with ingressing chloride ions21. With macro corrosion, the resistivity of concrete will impact

corrosion current and hence the rate of corrosion. Resistivity therefore affects the corrosion

resistance rather than the chloride resistance of concrete.

4.3 Marine Exposures

Marine structures are exposed to chlorides from seawater in four exposure conditions:

n submerged zone

n tidal zone

n splash and spray, and

n coastal zone.

Submerged structures are subject to sustained direct contact with seawater. Chlorides

penetrate into concrete mainly by ion diffusion, and to some extent permeation of the salt

solutions. The concrete surface zones may form protective coatings with a low permeability

due to ion exchange reactions with other compounds of seawater, resulting in films of

Mg(OH)2 and CaCO3. Therefore, the penetration rate of chlorides into these structures is often

considerably lower than estimated from laboratory experiments, where no protective films can

be formed due to the test method chosen.

Structures in tidal or splash and spray zone are subject to cyclic exposure to seawater. Ingress

of chlorides into the concrete is supported by capillary absorption of the seawater upon

direct contact. Capillary absorption gains importance as the degree of drying between the

individual wetting periods increases. The splash and spray zone is sometime referred to as the

atmospheric zone22.

Coastal structures may be subject to considerable chloride concentration in the atmosphere,

which may be deposited or washed out with rain at the surface of structures. Ingress of

chlorides into the concrete is supported by capillary absorption of the seawater upon direct


contact, and chloride removal during wash out is possible through reverse diffusion. With long

drying periods, carbonation of the concrete surface may lead to the release of the bound

chlorides in the carbonated zone.

There are exposure conditions where concrete is in contact with seawater under significant

hydrostatic pressure. In cases where the opposite face of a concrete element is subject to

drying condition such as immersed transport tunnel or basements, Wick action needs to be

considered23.

Table 4.1 summarises the primary chloride transport mechanisms applicable to structures in

various exposure conditions.

A field study conducted by Veritec’s Seawater Laboratory in Norway24 on the effect of static

and dynamic loads on corrosion behaviour of reinforced concrete structures in various

exposures has found that for completely submerged concrete, diffusion of oxygen to the

embedded steel is nearly totally restricted due to blocking of pores and the cracks by the

formation of calcareous products on the concrete surface. No corrosion was found and

the danger of localised attack on reinforcing steel, exposed in the bottom of a submerged

crack with a surface width up to 1 mm was negligible. The corrosion conditions for concrete

structures spanning through several environmental zones (multi-zone) differ from those found

on a completely submerged structure due to a more ample oxygen supply. No significant

effects of loading condition on the corrosion behaviour were found for both the submerged and

multi-zone structures.

4.4 Relative Severity of Exposure Conditions

The severity of various exposures is reflected in specifications. Mehta2 highlighted that, for

example, both FIP and ACI 357R have similar recommendations of 50-mm cover thickness

over principal reinforcing steel and 75-mm over prestressing tendons in submerged zone.

In the splash zone and the atmospheric zone, which are subject to seawater spray, the

recommended cover thickness is 65 mm for reinforcing steel and 90 mm for prestressing steel.

Stirrups may have 13 mm less cover than the preceding recommended values.

The Australian Standard on Concrete Structures, AS 3600, has three exposure classifications:

B2 – coastal (up to 1 km from coastline and usually referred to as atmospheric exposure),

B2 – permanently submerged, and C – tidal or splash zones. A study25 comparing chloride

ingress into structural and fender prestressed concrete piles of Iluka Wharf, at the mouth of the

Clarence River in northern New South Wales, revealed the tidal or splash zone to be the most

severe followed by the fully submerged and atmospheric zone. The severity was determined

by comparing the level of water-soluble chloride found at an depth of 65 mm from the surface

after up to 25 years of exposure. The durability performance of Port Kembla Olympic Pool, with

structural components situated in various marine environments, was investigated after over

60 years of exposure26. The findings in terms of chloride profiles also supported the relative

classification of exposure conditions in AS 3600 of B2 for atmospheric and submerged zones

and C for tidal zone.

In contrast to the above, a Norwegian study of a 60-year-old reinforced concrete pier in

Oslo harbour27 found higher chloride concentrations in concrete pillars in the continuously

submerged zone than in those in the tidal zone. However, it was noted that the locations just

above and within the tidal zone the chloride penetration was measured on concrete cores

drilled out through jackets of repaired pillars. Hence the concretes are different and exposed

for less than 60 years.


5 CHLORIDE RESISTANCE TESTS

The chloride resistance of concrete is quantified in terms of its properties and the minimum

thickness of the concrete cover protecting the reinforcement from the external environment.

Concrete has traditionally been proportioned for structural strength capacity and for durability.

For durability, concrete mixes are usually prescribed by the maximum water-cement ratio and/

or the minimum cement content. ACI 257R-84 (Reapproved 1997) Guide for the Design and

Construction of Fixed Offshore Concrete Structures, for example, specified maximum w/c and

corresponding nominal concrete cover for various marine exposures.

The advent of chemical admixtures and the use of supplementary cementitious materials

(SCM) has provided great flexibility in the production of workable concrete with specific

hardened concrete properties to suit each application. These technological developments

mean that concrete can have very different properties at the same water-cement ratio

depending on the type and amount of cement and chemical admixture. As an example, the

relationships between 28-day compressive strength of concrete and water-cement ratio of two

types of cement (GP and GB) are shown in Figure 5.1.

Figure 5.1 Compressive strength and water-cement ratio relationship (CSIRO6)

A range of concrete properties have been used as a measure of its resistance to chloride

penetration. Each will be described and discussed in terms of its relevance to chloride

transport mechanism, its adoption in specifications and testing precision, and its relationship

to measured long-term chloride profile quantified as effective or apparent chloride diffusion

coefficients. The long-term chloride profile is considered the best measure of the mixed modes

of chloride transportation and in particular the chloride level at cover depth.

Long-term research data Two comprehensive sets of research data along with other specific

experimental outcomes will be used to relate each concrete property to the long-term chloride

diffusion coefficients. CSIRO6 examined a range of properties of concretes proportioned from

one Type GP and three Type GB cements. Type GB cements comprised cement with either

a 30% fly ash, a 65% ground granulated blast-furnace slag or a 10% silica fume content. All

Water/cement

Mea

n 2

8-d

ay c

om

pre

ssiv

e st

ren

gth

(M

Pa)


the concretes were membrane-cured for 7 days and tested at 28 days, except compressive

strength specimens which were moist cured for 28 days. A set of specimens were immersed

in 3% sodium chloride solution and the effective chloride diffusions, De.28 and De.365, were

obtained after a 28-day and one-year immersion period respectively. Sherman et al28 also

investigated a range of concrete properties and corresponding apparent chloride diffusion

coefficients, Da.365, after 365-day ponding conducted using AASHTO T259. Fifteen mixtures of

concrete were proportioned from Type I cement with and without silica fume (at 5% and 7.5%)

at w/c of 0.30, 0.40 and 0.46. The concretes were cured in a number of ways including water

cured, wet burlap covered and heat cured.

5.1 Indirect Measures

5.1.1 Cement type and water-cement ratio

The porosity of concrete is highly sensitive to water-cement ratio. The connectivity of the pore

system depends on the amount of original mixing–water filled space and the degree to which

it has been filled with hydration products. Capillary pores are those voids remaining that were

originally filled with mixing water: ie pores with diameters in the range of 3.2 to 3,000 nm29.

These capillary pores will cease to be connected at different times in the age of the concrete

as a function of w/c and curing condition30. If stored moist, these times are approximately:

w/c 0.4 0.5 0.6 0.7 >0.7

time 3 d 14 d 6 m 1 y never

Water-cement ratio has therefore been used to specify for durability. However, various types

of cements have been found to affect chloride resistance of concrete differently. The cement

chemical composition affects both the porosity and chloride binding capacity of cement. The

effective chloride diffusion coefficient, De, of concrete incorporating a Type GP and three Type

GB cements, derived from concrete after one year immersion in 3% sodium chloride solution,

are shown in Figure 5.2. They show great divergence ranging from 5 to 30 x 10–12 m2/s at high

w/c of 0.6 to a narrow range of 1 to 4 x 10–12 m2/s at low w/c of 0.4. Similar findings are shown

for concrete made from an ASTM Type I cement and a blend of ASTM Type I and silica fume.

For high chloride resistance concrete, both the type of cement and maximum water-cement

ratio must be specified.

The Roads and Traffic Authority of New South Wales (RTA) has used such prescriptive

requirements in the B 80 specifications31. These are shown in Table 5.1.

TAbLE 5.1 Wet curing and other requirements in RTA specifications

Exposureclassification

Curing period requirement (days) Other requirements

SL cement

Blended cement containing BFS* and/or FA**

Blended cement containing AS***

Minimum cement content (kg/m3)

Maximum w/c (by mass)

Min strength for durability, fc.min(d) (MPa)

B2 7 14 7 370 0.46 40

C N/A 14 7 420 0.40 50

*blast furnace slag, **fly ash, ***amorphous silica


Figure 5.2 Effect of water-cement ratio and cement type on the chloride resistance

of concrete (CSIRO 6) (Sherman et al. 28)

Water/cement

Dif

fusi

on

co

effi

cien

t, D

e.36

5 (1

0–12

m2 /

s)

Water/cement

Dif

fusi

on

co

effi

cien

t, D

365

(10–1

2 m

2 /s)


5.1.2 Compressive strength

Both strength and transport characteristics are linked to the pore structure of the concrete.

Concrete with low porosity usually has high strength and high resistance to the penetration of

aggressive ions. It has been shown32, for example, that oxygen permeability (K in 10–16 m2)

and the compressive strength of concrete (fc.28 in MPa) made of different types of cement can

be related. For a range of concretes manufactured from three Type GP cements, the water

permeability (K28 tested at 28-day in 10–12 m/s) and the mean 28-day compressive strength

(fc.28 in MPa) has shown excellent correlation, with correlation coefficient R exceeding 95% as

shown in Figure 5.3. Increases in compressive strength reflect reduced porosity.

Figure 5.3 Influence of concrete strength on water permeability

Figure 5.4 Influence of strength on medium-term chloride diffusion (CSIRO 6) (Sherman et al. 28)

Mean 28-day compressive strength (MPa)

Wat

er p

erm

eab

ility

, K28

(10

–12

m/s

)

Mean 28-day compressive strength, fc.28 (MPa)

Dif

fusi

on

co

effi

cien

t, D

365

(10–1

2 m

2 /s)


Australian Bridge Design Standard AS 5100.5 specifies compressive strength grade and

corresponding nominal cover for concrete exposed to different environments including B2 and

C (‘permanently submerged’ in sea water and ‘in tidal or splash zones’ respectively).

For a range of concrete manufactured from Type GP and three Type GB cements, and ASTM

Type I cement and Type I and silica fume, the diffusion coefficient of concrete after 365-day

exposure in 3% sodium chloride solution, D365, have been found to correlate well with their

corresponding mean 28-day compressive strength, as shown in Figure 5.4. However, the type

of cement has been found to greatly influence these longer-term chloride diffusion coefficients,

especially for medium to low strength grade C20 to C32 concrete. For higher strength C40 to

C50 concrete commonly specified for concrete structures in aggressive environments, such

differences diminish as shown in Figure 5.4.

5.2 Direct Measures

5.2.1 Chloride diffusion coefficients

It is generally recognised that diffusion is the principal chloride transport mechanism from

the external environment into concrete. This is especially true in dense concrete where pore

water or vapour is well maintained in the concrete except for a few millimetres at the surface.

Chloride diffusion coefficient. D, is a measure of the resistance of concrete to chloride diffusion.

The ‘diffusion coefficient’ and ‘concrete cover’ combine to provide a measure of the resistance

of concrete to chloride ingress. They also enable the prediction of service life and hence are

the most direct means of specifying chloride penetration resistance.

The diffusion coefficient can be determined from a non-steady-state or a steady-state of

chloride test using Fick’s second and Fick’s first law respectively. Due to time constraints and

specimen configuration, it is more common for the chloride diffusion coefficient of concrete to

be determined from a non-steady-state test.

In non-steady-state chloride diffusion, the diffusion coefficient derived from the total or

acid-soluble chloride profile, which accounts for both the free and bound chloride, is

the apparent diffusion coefficient, Da. The diffusion coefficient derived from the free or

water-soluble chloride profile is the effective diffusion coefficient, De. Both Da and De are highly

dependent on the exposure period with an order of reduction found from a 28-day exposure to

tens of years of exposure as shown in Figure 5.5. The curing period and age of the concrete

at the commencement of exposure to chloride also influences the resultant diffusion coefficient

but to a lesser extent than exposure period as shown in Table 5.2. In this case, the diffusion

coefficients are of the same order. In durability specifications in term of chloride diffusion

coefficient, both the age of concrete and exposure period must be clearly specified.

TAbLE 5.2 Influence of curing and age of concrete on De (Cao et al 33 1993)

Curing period prior to exposure (days)

Diffusion coefficient after 28-day exposure, De.28 (10–12 m2/s)

Concrete 1 Concrete 2 Concrete 3

28 3.17 3.50 3.17

56 2.25 1.21 1.22


Figure 5.5 Dependence of non-steady-state chloride diffusion coefficient

on exposure period (Vallini and Aldred 34)

Figure 5.6 Short-term exposure diffusion coefficient De.28 as indicator for longer-term diffusion

coefficient, De.365 (CSIRO 6)

The chloride diffusion coefficient, determined from a steady-state chloride diffusion, is

the effective chloride diffusion coefficient and is considered to be the asymptote of the

corresponding non-steady-state effective chloride diffusion coefficient, De, derived from an

indefinite exposure period.

In Australia, the first major project where concrete was specified for durability by diffusion

coefficient was the West Tuna and Bream B concrete gravity-base offshore structures in

1993. The apparent chloride diffusion coefficient of the 56-day cured concrete after 28-day

immersion in substituted ocean water was 0.8 x 10–12 m2/s.

The apparent diffusion coefficient results of two properly conducted tests, according to

ASTM C1566, should not differ by more than 39.8% of the mean value.

Exposure period (years)

Dif

fusi

on

co

effi

cien

t, D

a (1

0–12

m2 /

s)

Diffusion coefficient, De.28 (10–12 m2/s)

Dif

fusi

on

co

effi

cien

t, D

e.36

5 (1

0–12

m2 /

s)


Diffusion coefficients from a short-term exposure period (say 28 days, De.28) are the more

practical specification used for major infrastructure projects when there is sufficient leadtime

for the appropriate concrete mix design to be developed and tested prior to commencement

of construction. Such diffusion coefficients are reasonably good indicator of the longer-term

chloride resistance, De.365, as shown in Figure 5.6 for a set of one Type GP and three Type GB

cement concretes. The correlation coefficient, R, of 74% indicates that short-term diffusion

coefficients are not necessarily a very good indicator of longer-term chloride resistance. The

results also show a reduction of diffusion coefficients with exposure period, with the value of

De.28 being about a third of De.365.

5.2.2 Absorption, sorptivity, ISAT and AVPV

A number of absorption-related properties are used to indicate the porosity and durability

potential of concrete. Absorption, initial surface absorption (ISAT), sorptivity and apparent

volume of permeable void (AVPV) are each fundamentally a measure of capillary absorption or

absorption rate of concrete.

Absorption is a measure of multi-dimensional capillary absorption of water into concrete. The

amount of water absorption of a concrete depends on the porosity and its interconnectivity, the

moisture condition or degrees of internal drying and the temperature of the water. AS 1012.21.

BS 1881: Part 122 and ASTM C642 are the respective Australian, British and American

Standard test methods for water absorption determination.

Absorption has been used as a measure for durability in concrete products such as pipes and

blocks. It has not been commonly used to measure the chloride resistance of concrete, possibly

because it is not a good indicator of chloride diffusion coefficient as shown in Figure 5.7.

Figure 5.7 Poor correlation between water absorption and medium-term chloride diffusion

coefficient (Sherman et al 28)

Sorptivity is a measure of one-dimensional capillary absorption rate as a function of time

and provides a good indication of the pore structure and its connectivity of the near-surface

concrete. There are two methods of measuring sorptivity. The first (RTA T362) measures the

depth of penetrating water front into the concrete and the second (ASTM C1585-04) measures

Absorption (%)

Dif

fusi

on

co

effi

cien

t, D

a.36

5 (1

0–12

m2 /

s)


the depth of water penetration indirectly through the weight gain in the concrete. The sorptivity

is calculated from the slope of the curve of the depth of water penetration with time in

mm/hour½ or the weight gain against time and converted into mm/min½ or mm/s½ for the

respective methods.

The Roads and Traffic Authority of New South Wales (RTA) has used sorptivity16 as an

alternative performance-based specification (Provision A) in RTA B 80 specifications 31. The

RTA sorptivity for exposure classes B2 and C Table 5.3 is measured over a soaking period of

24 hours, hence RTA sorptivity of 1 mm is equivalent to 0.026 mm/min½, and RTA sorptivity of

3.8 mm is equivalent to 0.1 mm/min½.

TAbLE 5.3 Durability requirements for concrete (Provision A) in RTA B8031

Exposure classification

Minimum cement content (kg/m3

)

Maximum water/cement ratio(by mass)

Maximum sorptivity penetration depth (mm)

Portland cement Blended cement

B2 370 0.46 17 20

C 420 0.40 N/A 8

Figure 5.8 Influence of RTA sorptivity on medium-term effective chloride diffusion

Sorptivity has been found to be a good performance indicator of chloride resistance of concrete.

In Figure 5.8, the relationship between sorptivity and medium-term effective chloride diffusion

coefficients of a range of concretes (one Type GP and three Type GB) is shown to have

reasonably good correlation (R = 76%). Lower sorptivity indicates better chloride resistance.

For Class C exposure, a maximum sorptivity limit of 8 mm or 0.21 mm/min½ is specified for

blended cement. It can be observed from Figure 5.8 that such low sorptivity would indicate a

concrete with a very low diffusion coefficient. Such low sorptivity has been found to be difficult

to measure and no precision statement is available for the RTA sorptivity.

RTA sorptivity (mm)

Dif

fusi

on

co

effi

cien

t, D

e.36

5 (1

0–12

m2 /

s)


Figure 5.9 Sorptivity of 19-year-old concretes and apparent chloride diffusion

coefficient, Da, after 19-year exposure in splash zone in Port Fremantle 34

The sorptivity of matured concrete (tested at 19 years old) has also been found to give a

good indication of the long-term chloride resistance of a range of concretes exposed in the

splash zone in Port Fremantle in Western Australia 34. The concrete mixes were proportioned

from a range of cements including Type GP cement, Type GB cement with 30–65% GGBS

and Type GB cement with 10% silica fume, all at nominal water-cement ratio of 0.4. A good

correlation coefficient R of 85% is found as shown in Figure 5.9. The sorptivity is measured

using a method similar to ASTM C1585.

ASTM C1585 states that the repeatability coefficient of variation is found to be 6.0% in

preliminary measurements for the absorption for a single laboratory and single operator.

It should be noted that the non-steady-state apparent diffusion coefficient, Da.19yr, is about

an order of magnitude below the non-steady-state effective diffusion coefficient, De.365, and

the sorptivity of the matured concrete is correspondently an order of magnitude below the

sorptivity of the 28-day-old concrete. These are good examples of the reduction of diffusion

coefficient with exposure time, and the limitation of specifying diffusion coefficient from

short-term exposure tests.

The initial surface absorption test (ISAT) is a British Standard method used to measure the

water absorption of a concrete surface35. The ISAT is reported to be sensitive to small changes

in concrete mix constituents, strength grade and curing. The test is particularly useful for

field measurement of insitu concrete surface absorption as the equipment is portable and is

inexpensive. However, the surface absorption, like all permeation properties, is very sensitive

to the moisture condition of the concrete and hence conditioning of the surface prior to testing

is important. In principle, ISAC measures the sorptivity of concrete. However, the geometry

of the opening of the cap and its contact with concrete surface mean that the complex two-

dimensional absorption measured needs to be analytically converted into the standard one-

dimensional sorptivity. In addition, it cannot be used to estimate the porosity of the concrete36.

Sorptivity of mature concrete (mm/min1/2)

Dif

fusi

on

co

effi

cien

t, D

a.19

yr (

10–1

2 m

2 /s)


A tentative classification for concrete durability based on ISAT has been proposed as shown

in Table 5.4. Experimental evaluation of ISAT-10, the ISAT measurement taken at 10 minutes,

showed that localised material differences within the same concrete can cause variations of

ISAT-10 up to ± 24% whereas various curing regimes can produce ISAT-10 values of ± 63%

about a mean value of the same concrete37.

TAbLE 5.4 Tentative classification for concrete durability based on ISAT37

Durability classification ranking ISAT-10 (ml/m2/sec X 10–2)

1 <50

2 51–70

3 71–90

4 91–110

5 >110

The relationship between ISAT-10 and coefficient of chloride diffusion has been evaluated

and is reproduced in Figure 5.10 37. ISAT-10 is an excellent indicator of chloride resistance of

concrete quantified in terms of the effective diffusion coefficient. The relationship between ISAT

and diffusion found in Figure 5.10 is very similar to that between sorptivity and longer term

effective diffusion coefficients shown in Figure 5.9.

Figure 5.10 ISAT-10 and chloride diffusion coefficient of OPC concrete (Dhir et al.37)

Volume of permeable void is a method of determining the water absorption after immersion

in water at room temperature, after immersion and boiling, and the 'volume of permeable voids

(VPV)' or volume of water absorption after a period in boiling water of a hardened concrete

sample. The high temperature affects both the viscosity and the mobility of the water molecules

which may enable the greater displacement of pore system within the hardened concrete. This

is shown in the relationship between absorption and the absorption after boiling in Figure 5.11.

Boiling resulted in a 6% increase in absorption with the exception of a few outlying points.

ISAT-10 (ml/m2/sec x 10–12)

Dif

fusi

on

co

effi

cien

t, D

(10

–12

m2 /

s)


The ASTM C642 method measures the volume of permeable voids (VPV) as a percentage of

the volume of the solid. The Australian Standard AS 1012.21 test method has been adapted

from the ASTM method. It measures the apparent volume of permeable voids (AVPV) as a

percentage of the volume of the bulk materials, ie solid and voids. The AVPV has been used by

VicRoads to classify concrete durability as shown in Table 5.5.

Figure 5.11 Effect of boiling on absorption (Sherman et al. 28)

TAbLE 5.5 VicRoads classification for concrete durability based on the AVPV

Durability classification indicator

Vibrated cylinders (AVPV %)

Rodded cylinders (AVPV %)

Cores (AVPV %)

1 Excellent <11 <12 <14

2 Good 11–13 12–14 14–16

3 Normal 13–14 14–15 16–17

4 Marginal 14–16 15–17 17–19

5 Bad >16 >17 >19

The chloride resistance of concrete has been found to improve with the reduction in VPV

as shown in Figure 5.12. The resistance based on medium-term effective chloride diffusion

coefficients, De.365, has been found to correlate well with the VPV with a correlation coefficient

of 86% for VPV exceeding about 13%. At lower VPV in the range of 6–14%, Sherman et al. 29

found no correlation (R=12%) between VPV and medium-term apparent chloride diffusion

coefficients, Da.365. as shown in Figure 5.12. From the two sets of slightly different data, there

appears to be a change of chloride resistance around the critical VPV level of 12–13%.

Earlier research by Whiting38 examined the relationship between VPV and the chloride content

from 2 to 40 mm below the concrete surface as percentage by weight of concrete, after a

Absorption (%)

Ab

sorp

tio

n a

fter

bo

ilin

g (

%)


short-term 90-day AASHTO T259-80 ponding test in 3% sodium chloride solution. The results

indicate improvement in chloride resistance with reducing VPV shown in Figure 5.13. It also

supports the critical range of VPV of around 12%.

There is no precision statement for VPV in ASTM and AS standards. It is therefore extremely

difficult in practice to set narrow compliance limits.

Figure 5.12 Relationships between ASTM VPV28 and De.365 (CSIRO 6) and

ASTM VPV42 and Da.365 (Sherman et al. 28)

Figure 5.13 Influence of ASTM VPV on chloride resistance (Whiting 38)

ASTM C642 volume of permeable voids (%)

Dif

fusi

on

co

effi

cien

t, D

e.36

5 (1

0–12

m2 /

s)

VPV (%)

Ch

lori

de

afte

r 90

-day

po

nd

ing

(%

)


5.2.3 Coefficient of permeability

The term ‘permeability’ has been used to describe the ease with which a fluid moves through

concrete and in general to describe its durability. Strictly this mode of transport is applicable

only to concrete structures in direct contact with a liquid under a pressure head such as

occurs in liquid-retaining structures. Dissolved chlorides and gases are therefore transported

by convection with the permeating water into concrete.

The coefficient of water permeability, K28, of concrete tested at 28 days has been found

to correlate well with the compressive strength of concrete as shown in Figure 5.5. There

is, however, limited information on the relationship between water or air permeability and

chloride resistance. Earlier research38 has examined the relationship between hydraulic

and air permeability (m2) and chloride from 2 to 40 mm below the concrete surface after

90-day AASHTO T259-80 ponding test in 3% sodium chloride solution. Figure 5.14 shows

improvements in the chloride resistance of concrete with reduction in both hydraulic and air

permeability. However, the air permeability is found to be a more sensitive indicator of chloride

resistance than the hydraulic permeability.

Figure 5.14 Hydraulic permeability as an indicator of chloride resistance (Whiting 38)

Ch

lori

de

afte

r 90

-day

po

nd

ing

(%

)

Hydraulic permeability (m2)

Ch

lori

de

afte

r 90

-day

po

nd

ing

(%

)

Air permeability (m2)


The DIN 1048 Part 5 is one version of the penetration method which reports the water

penetration depth as a simplified indicator of water permeability. There is no precision

statement available for water permeability coefficient nor for DIN 1048 test.

Oxygen or air permeability A prototype surface air flow (SAF) device for the estimation

of concrete permeability has been developed 39. The method measures the rate of air flow

through a vacuum plate placed on a concrete surface under a vacuum of approximately

25 inches of mercury (16.6 kPa absolute pressure). The effective depth of measurement

is approximately 12 mm below the surface. The method has been found to correlate well

with chloride diffusion constants derived from 90-day ponding tests, as well as with true air

permeabilities measured using a pulse decay technique on concretes with a variety of water-

cement ratios as well as admixtures such as latex and silica fume.

5.2.4 Rapid chloride permeability test

The rapid chloride permeability test (RCPT) was first developed as a rapid means of assessing

permeability of concrete to chloride ions. It was adopted as a standard test method for rapid

determination of the chloride permeability of concrete by the American Association of State

Highway and Transport Officials as AASHTO T277 in 1989 and subsequently as ASTM C120218.

The technique basically measures chloride ion migration or the electrical conductance of

concrete. In this method, a potential difference of 60 v DC is maintained across the ends of

51-mm-thick slices of a concrete cylinder, one of which is immersed in a sodium chloride

solution, the other in a sodium hydroxide solution. The amount of electrical current passing

through the concrete during a six-hour period is measured and the total charge passed, in

coulombs, is used as an indicator of the resistance of the concrete to chloride ion penetration.

A table classifying concrete resistance to chloride ion penetrability has been proposed40 as

reproduced in Table 5.6.

Whiting40 examined the relationship between the rapid chloride permeability tested at 90 days,

RCPT90, and chloride penetration determined from the depth of 2 to 40 mm concrete surface

after 90-day ponding test in 3% sodium chloride solution, Cl(2–40mm).90. A very good correlation

was found between the charge passed and the amount of penetrating chloride as shown in

Figure 5.15.

TAbLE 5.6 Chloride ion penetrability based on charge passed (Whiting40)

Charge passed Chloride ion penetrability

>4,000 High

2,000–4,000 Moderate

1,000–2,000 Low

100–1,000 Very Low

<100 Negligible

In Australia, the RCPT was first used in the early 1990s to specify highly durable concrete for

the sea wall at Sydney Airport Parallel Runway project at the 1000-coulomb limit. According to

ASTM C1202, the results of two properly conducted tests in different laboratories on the same

material should not differ by more than 51%. The average of three test results in two different

laboratories should not differ by more than 42%.


Figure 5.15 ASTM C1202 as indicator of chloride resistance (Whiting 40)

Figure 5.16 RCPT28 versus De.365 and RCPT42 versus Da.365

The RCPT charge passed has been found to be a good indicator of the chloride resistance

of concrete17, 28. The charge passed measured at 28 days or 42 days, RCPT28 or RCPT42,

correlates well with the medium-term apparent or effective chloride diffusion coefficients,

Da.365 or De.365, as shown in Figure 5.16. The increase in the age of testing from 28 to 42 days

is not expected to affect the charge passed by concrete with ‘low’ and ‘very low’ chloride

penetrability and hence RCPT42 Data (≤1000 coulombs) are combined with RCPT28 to show

the combined relationship for the full range of charge passed shown in Figure 5.16.

Concrete with RCPT charge passed below 2000 coulombs shows consistently good chloride

resistance with Da.365 below 3 x 10–12 m2/s whereas concretes in the 2000- to 3000-coulomb

range exhibit highly variable chloride resistance.

ASTM C1202 RCPT90 (coulomb)

Ch

lori

de

pen

etra

tio

n, C

l (2–4

0 m

m).

90(w

t%)

Dif

fusi

on

co

effi

cien

t, D

365

(10–1

2 m

2 /s)

ASTM C1012 charge passed (coulomb)


Sherman et al. 28 found that the correlations between RCPT and long-term chloride diffusion,

the surface chloride concentration, and the time-to-corrosion to be highly variable and that it

requires individual correlation between the test and every concrete mixture. The widely used

1000-coulomb cut-off limit was claimed to be arbitrary and misleading for many concretes, due

to the widely different chloride permeability observed for concretes both meeting and failing

such a coulomb limit-based specification. The use of heat curing was found in the same study,

for example, to increase the coulomb values of concrete without increasing its actual chloride

permeability.

5.3 Electrochemical Properties

Electrical resistivity of concrete is one of the most significant parameters controlling the rate

of active corrosion of the embedded steel reinforcement. Insitu measurement of concrete

resistivity can be made by using four equi-spaced electrodes with surface contacts. An

alternating current, l, is passed via the concrete through the outer pair of contacts, and the

resulting voltage, V, between the inner contacts is measured. For a semi-infinite homogeneous

material, the resistivity is given by:

r = 2 p a(V/ I)

where a is the contact spacing. A commercial instrument is available. However, results

obtained from insitu measurement have been found to be highly variable. Recent

investigation41 has revealed field measured electrical resistivity to vary with electrode spacing,

concrete cover and the presence of embedded steel. Measurement should be taken with

electrode spacing of less than 30 mm and as far away as possible from embedded steel.

The resistivity can be measured using the Wenner Bridge four-electrode method42. However,

it is essential for the electrodes to be cast into the concrete to ensure good electrical contact.

Such a method is therefore more commonly used in corrosion research and provides more

consistent and repeatable measurements.

In a 5-year research on chloride-induced corrosion of steel in a range of concretes exposed

to simulate tidal exposure, it was found43 that concrete with high resistivity resulted in a

substantially longer time before a probable active corrosion state. High resistivity concrete is

defined as concrete with 28-day resistivity of greater than 4000 ohm cm or a 56-day resistivity

of greater than 5000 ohm cm. The cover to reinforcement required is:

For low resistivity concrete C = ( 2822 ) x √t

S1.28

For high resistivity concrete C = ( 226,537 ) x √t

S 2.61

where C = cover to reinforcement (mm)

S = mean 28-day compressive strength (MPa)

t = time taken to reach a probable active corrosion (years).

6 COMPARATIVE PERFORMANCE DATA

The suitability of each property as a performance-based indicator for chloride resistance of

concrete is considered from the precision of the test method, its relationship with medium-term

chloride diffusion resistance, De.365, evaluated as the correlation coefficient, R (determined

from least square method), for the corresponding range of measureable values. These are

shown in Table 6.1.


TAbLE 6.1 Precision and correlation of each property with chloride resistance

De.28(10–12 m2/s)

f28 (MPa)

Sorptivity28 (mm/s½)

ISAT-10(10–2 ml/m2/s) VPV28 (%)

RCPT28(coulomb)

K (m/s)

Precision from relevant standards or other sources

Sources of precision ASTM C1566

ASTM C39

ASTM C1585

Dhir et al. ASTM C642 ASTM C1202

DIN 1048

Reportable to 0.001 0.1 0.1x10–4 NA(1) NA NA NA

Repeatability CV 14% 3% 6% ±24% NA 12.3% NA

Reproducibility 20% 5%(2) NA NA NA 18% NA

Correlation to De.365 of a GP and three Gb cement concretes (CSIRO)

Range of value correlation coef, R

10–60 74%

20–60 58%(3)

15–100mm 76%

– –

6–13, 13–18 12%, 86%(4)

2–5 x103

76%– –

Critical range to one order change in De.365 1 to10 m2/s

28 37 41 mm – 1.5 1270 –

Increment used in classifications

–

10th MPa

RTA B80 12 mm

20

VicRoads 1–2%

C1202 1000

–

Correlation to other chloride resistance indicators, %

Sherman Da.365 – – – – 12(4) 89 –

Dhir De.steady – – – 99(5) – – –

Whiting Cl2–40 mm.90 – – – – 99 97 78

(1) Not available

(2) For 150-mm-diameter cylinders

(3) Better correlation R of 94% for Type GP cement and 80% for Type GB cements

(4) Correlation R of 86% and 12% are applicable to VPV range of 13–18% and 5–14% respectively

(5) Correlation to steady-state diffusion coefficient, De.steady

With the exception of compressive strength, sorptivity, VPV and RCPT correlate reasonably well

with the medium-term chloride diffusion coefficient, De.365, with better than 75% correlation and

around 74% between short-term De.28 and De.365. The correlation improves in all properties for

high chloride resistance concretes. Compressive strength shows a poor correlation coefficient

of 58% to the medium-term chloride diffusion coefficient unless the type of cement is also

considered (94% for Type GP cement and 80% for Type GB cements). Hence strength and

cement type are the most suitable parameters for a semi-prescriptive specification. For the

ISAT and coefficient of water permeability, K, their relationships with other measures of chloride

resistance are also determined and shown in Table 6.1. Very good correlation was found for

ISAT and chloride in the 2- to 40-mm cover zone. The VPV, however, shows no correlation with

the medium-term chloride diffusion coefficient in the VPV range of 6–13 %, the range specified

by VicRoads for ‘good’ and ‘excellent’ concrete (vibrated).


In setting the performance limits for each chloride resistance property, the sensitivity of

each property should be evaluated in terms of the ‘critical range’ and the corresponding

repeatability of the test. The critical range chosen for comparison in Table 6.1 is the range

of each property required to detect a change of one order of magnitude in De.365 from

1.0 to10.0 x10–12 m2/s.

7 CHLORIDE RESISTANCE ENHANCING MEASURES

This section examines the alternative methods used to improve the chloride resistance of

concrete, viz surface treatments and the use of integral admixtures and corrosion inhibitors in

concrete.

7.1 Sealers

Damp-proofing is a process of treating a concrete surface to reduce absorption. Damp-

proofing by the use of sealers is considered more economical than waterproofing. There are

two categories of sealing materials: coatings and penetrating sealers. Surface coatings can be

clear or coloured. Penetrating sealers can be either inert pore-plugging materials or chemically

reactive products. Sealers are useful on surfaces exposed to cyclic wetting and drying but are

not recommended in continually submerged situations.

A survey revealed that highway agencies in the United States and Canada were very interested

in sealers but use remained limited. Linseed oil has been used although a number of agencies

no longer use it because of its poor long-term performance. Highway agencies are more often

using materials such as silanes and siloxanes. Preliminary experimental work carried out under

the SHRP program44 indicated that penetrating sealers could have a significant effect on the

surface electrical properties of concrete, with electrical resistance in the near-surface layers

of concrete staying higher after wetting than in untreated concretes. Additionally, regaining

of insulation characteristics after removal of surface water was more rapid for sealed than for

unsealed specimens.

Two test procedures have been developed to evaluate the effectiveness of sealers. The first

‘surface resistance test’ with a criterion of 200 k-ohms after 4 minutes of testing was selected

for differentiating between effective and ineffective sealers. The second ‘absorption method’, a

modified European procedure (RILEM II.4) indicated that column drops of less than 10 mm in

4 minutes are generally associated with effective sealers, while drops over 20 mm in 4 minutes

are associated with ineffective sealers. Field trials in the states of Vermont, California and

Minnesota showed good agreement between the test methods and the expected performance

of the sealers in structures.

A recommended practice for the use of sealers is outlined in ACI 345.1R Routine Maintenance

of Concrete Bridge Members.

7.2 Corrosion Inhibitor

Corrosion inhibitors are admixtures incorporated into fresh concrete. ACI 222.3R Design and

Construction Practices to Mitigate Corrosion of Reinforcement in Concrete Structures defines

them as chemical substances that decrease the corrosion rate when present at a suitable

concentration, without significantly change the concentration of any other corrosion agent.

These admixtures act on the steel surface, either electrochemically (anodic, cathodic, mixed-

inhibitor) or chemically (chemical barrier) to inhibit chloride-induced corrosion above the

chloride-corrosion threshold level.


Commercial systems include an inorganic admixture containing calcium nitrite and several

organic admixtures containing: alkanolamines; amines and esters; alkanolamines and amines

and their salts with organic/inorganic acids; alkanolamines, ethanolamine and phosphate.

Combined organic/inorganic concrete admixture containing amine derivatives and sodium

nitrite are also available. There are also surface applied coatings which incorporate a corrosion

inhibitor.

Laboratory evaluation and field performance Independent investigations45 on the

effectiveness and harmlessness of calcium nitrite as a corrosion inhibitor in cracked and

uncracked concrete have found that calcium nitrite does not affect the chloride diffusion

coefficient of concrete. The threshold value for the corrosion inducing chloride content to a

depth of 20 mm is at least doubled in comparison to the untreated concrete. The initial results

suggest that calcium nitrite could be effective under simulated severe exposure conditions

even for ‘poor’ concrete (w/c = 0.5). Results after 400 days of exposure showed that calcium

nitrite was harmless for the ‘poor’ concrete in the region of cracks. On the contrary, for the

good quality concrete (w/c = 0.38), calcium nitrite seems to be effective even in the region of

cracks with a width of 0.30 mm.

The effectiveness of a number of corrosion inhibiting systems were evaluated in field exposure

of concrete barrier walls of the Vachon bridge located North of Montreal by the National

Research Council (NRC) of Canada. The bridge is subject to the simultaneous effects of

de-icing salt contamination, freeze-thaw and wet-dry cycles. The commercial systems

investigated include an inorganic admixture containing calcium nitrite and several organic

admixtures. The concrete is reported to have a water-cement ratio of 0.36, a CSA Type 10

cement content of 450 kg/m3, an air content of 6.5%, a slump of 80 mm, and an average

28-day compressive strength of 45 MPa. After 10 years exposure between 1996 and 2006 with

measurements of half-cell potential, corrosion rate (Gecor 6 measuring polarisation resistance)

and concrete resistivity, as well as coring and determination of chloride contents at various

depths, it was concluded46 that the inorganic admixture system gave consistently good

performance, with a reduced risk of corrosion, followed by other organic concrete admixtures

in comparison to the control system.

7.3 Other Admixtures

Apart from water reducers and high-range water reducers (superplasticisers) used to produce

better chloride resistance concrete through the reduction of water-cement ratio, a range of

admixtures are used to provide a physical barrier to reduce the rate of penetration of corrosive

agents into the concrete. Bitumen, silicates and water-based organic admixtures consisting

of fatty acids, such as oleic acid; stearic acid; salts of calcium oleate; and esters, such as

butyloleate, are typically used in these types of admixtures. There are limited long-term studies

evaluating the effectiveness of such admixtures.

The potential benefit of the use of an hydrophobic admixture on the chloride resistance of

concrete has been studied in a long-term field exposure experiment at Port Fremantle in

Western Australia. The mix proportions and long-term diffusion coefficients of mixes with and

without the admixture have been reported34.

It appears that the addition of this particular admixture resulted in an adjustment of water

content resulting in GP cement concrete and possibly 65% GGBS concrete with higher water-

cement ratios than corresponding concrete without the admixture. However, for concrete with

the same cement content, the admixture has resulted in substantial improvement in chloride


resistance observed in terms of the diffusion coefficients after 6-year exposure in a splash

zone. The improved chloride resistance remained moderate after 19-year exposure, with Da

around three times that of concrete with 10% silica fume.

8 CONCLUSIONS

Concrete provides physical and chemical protection to the reinforcing steel against penetrating

chlorides which may cause steel depassivation, leading to increased risk of steel corrosion. The

chloride resistance of concrete depends on the permeability properties of the concrete and the

cover thickness to the reinforcement. The integrity of the concrete cover under service load, in

terms of cracking and crack width, also influences the resistance to penetrating chlorides.

The resistance of the concrete depends largely on the porosity and interconnectivity of the

pore system in the concrete, and to a lesser extent on the chemical binding capacity of

the cement. The chloride resistance of concrete thus depends on the mix constituents, mix

proportions, the degree of compaction and curing given to the fresh and hardened concrete.

A range of concrete properties has been used to measure the chloride resistance of the

concrete. As diffusion and capillary absorption are the primary chloride transport mechanisms

into most concrete structures exposed to chlorides, the relevant concrete property is thus

quantified in terms of its relationship to a 365-day chloride profile or a 90-day penetrating

chloride in the 2- to 40-mm cover region. The chloride diffusion coefficient, derived from the

chloride profile after long-term exposure, is considered one of the best indicators of chloride

resistance.

A number of parameters and corresponding test methods have been examined and, where

appropriate, were related to various chloride transport mechanisms as follows:

Diffusion Chloride diffusion coefficient, De.28 (10–12 m2/s)

Capillary absorption Absorption (%)

Sorptivity (mm/min1/2),

Initial Surface Absorption Test, ISAT (10–2 ml/m2/s)

Volume of Permeable Voids, VPV (%)

Compressive strength, f28 (MPa), an indirect measure of porosity

Permeability Water Permeability Coefficients, K (m/s)

Migration Rapid Chloride Permeability Test, RCPT (Coulomb)

All transport Water-cement ratio, w/c.

The use of prescriptive specifications of cement type and w/c has been found to be extremely

effective in specifying chloride resistance. The effect of the type of cement on chloride

resistance of concrete is quite evident for concrete at higher w/c, but such distinctions diminish

for concrete at low w/c below about 0.45 (0.35 for concrete with silica fume). There is, however,

no acceptable method of testing w/c of hardened concrete; control of w/c can be only

conducted at the batch plant.

A semi-prescriptive specification of cement type and compressive strength has also been

found to be extremely effective in specifying chloride resistance. Like w/c, the distinction in

chloride resistance between different cements diminishes as compressive strength increases

beyond 40 MPa. This represents the most practical approach to specification and quality

control using compressive strength grade and characteristic-strength-compliance criterion.


For performance-based specifications, capillary absorption properties (sorptivity, ISAT

and VPV) and migration property (RCPT) have all been found to be as good as indicators

of longer-term chloride resistance as the short-term chloride diffusion coefficient. Critical

comparative performance data are given in Section 6. It was found that there are limits to the

effective range of sorptivity (15–100 mm) and VPV (13–17%) that are sensitive to chloride

resistance. There are no published precision data for these tests and narrow performance

classes can be set only with known precision data. The water permeability has also been

found to be a good indicator of only the short-term chloride resistance, hence its use should be

limited to liquid-retaining structures.

The resistivity of concrete has been found to be a good indicator of chloride resistance in

laboratory-based measurement. The variability of insitu resistivity measurement and the lack of

correlation data with longer-term diffusion coefficient have so far prevented its use in chloride

resistance specifications.

9 REFERENCES

1 Durable Concrete Structures Cement & Concrete Association of Australia, 1989.

2 Mehta P K ‘Durability of Concrete Exposed to Marine Environment – A Fresh Look’, Proc,

Second Int. Conf. on Concrete in Marine Environment, SP-109, St. Andrews by-the-Sea,

Canada, 1988.

3 Specification and supply of concrete, AS 1379 Standards Australia.

4 Stark D Long-Time Performance of Concrete in a Seawater Exposure PCA Research and

Development Information RP337, R&D Serial No. 2004, 1995.

5 Gowripalan N, Wainwright P J, Cabrera J G and Cusens A R Effect of Curing on Concrete

Durability with particular reference to Curing Membranes, Final Report to the Science and

Engineering Research Council, UK, Sept 1989.

6 Khatri R P, Hirschausen D and Sirivivatnanon V Broadening the use of fly ash concretes

within current specifications CSIRO Report BRE045, 1998.

7 Early-age Thermal Crack Control in Concrete (CIRIA C660), Construction Industry

Research and Information Association.

8 Beeby A W Concrete in the Oceans – Cracking and Corrosion, Technical Report No. 1,

CIRIA/UEG, Cement and Concrete Association, Dept of Energy, UK, 1978.

9 Mohammed T U, Hamada, H and Otsuki N ‘Several factors regarding sustainability of

marine concrete structures’, Proceedings Sixth CANMET/ACI Int. Conf. on Durability of

Concrete, SP-212, pp 23–44.

10 Lim C C ‘Influence of cracks on the service life prediction of concrete structures in

aggressive environments’, UNSW PhD thesis, 2000.

11 Gowripalan N, Sirivivatnanon V and Lim C C ‘Chloride Diffusivity of Concrete Cracked in

Flexure’, Cement and Concrete Research, Vol. 30, No. 5, pp. 725–730, May 2000.

12 Cao H T and Sirivivatnanon V ‘Service Life Modelling of Crack-freed and Cracked

Reinforced Concrete Members subjected to Working Load’, Proceedings CIB Building

Congress 2001, Wellington, New Zealand, 2-6 April, 2001.

13 Performance Criteria for Concretre Durabilty edited by J Kropp and H K Hilsdorf, E & F N

Spon, London, 1995.

14 Standard Method for Determining the Apparent Chloride Diffusion Coefficient of

Cementitious Mixtures by Bulk Diffusion (ASTM C 1556 – 04).


15 Luping T and Sorensen H ‘Evaluation of the Rapid Test Methods for Measuring the Choride

Diffusion Coefficients of Concrete’, NORDTEST Project No. 1388-98, Swedish National

Testing and Research Institute, SP Report 1998:42.

16 Interim test for verification of curing regime – Sorptivity, RTA Test Method T362, February

2001.

17 Khatri R and Sirivivatnanon V ‘Methods for the Determination of Water Permeability of

Concrete’, ACI Materials Journal, Vol. 94, No. 3, May-June, 1997.

18 Standard Test Method for Electrical Indication of Concrete’s Ability to Resist Chloride Ion

Penetration ASTM C 1202 – 05.

19 Tang L and Nilsson L-O ‘Rapid Determination of the Chloride Diffusivity in Concrete by

Applying an Electrical Field’ ACI Materials Journal, Vol. 89, No. 1, pp. 49–53, 1992.

20 Boddy A, Bentz E, Thomas M D A and Hooton R D, ‘An overview and sensitivity study of a

multimechanistic chloride transport model’, Cement and Concrete Research 29,

pp 827–837, 1999.

21 Baweja D, Roper H and Sirivivatnanon V ‘Quantitative Descriptions of Steel Corrosion in

Concrete using Resistivity and Anodic Polarisation Data’, Proceedings 4th CANMET/ACI

Int. Conf. on Durability of Concrete, edited by V M Malhotra, SP 170-3, pp 41–63.

22 Mehta P K ‘Durability of Concrete Exposed to Marine Environment – A Fresh Look’,

Proceedings Second Int. Conf. on Concrete in Marine Environment, SP-109, 1988.

23 Buenfeld N R, Shurafa-Daoudi M T and McLoughlin I M ‘Chloride transport due to wick

action in concrete’ Nilsson L O, Ollivier J P (Eds.), Chloride Penetration into Concrete,

Proceedings Int. RILEM Workshop, 1997, pp. 315–324.

24 Espelid B and Nilsen N ‘A field study of the corrosion behaviour on dynamically loaded

marine concrete structures’, Proceedings Second Int. Conf. SP109-4, 1988, pp 85–104.

25 Driscoll S, Sirivivatnanon, V and Khatri R P ‘Performance of a 25-year-old Coastal Concrete

Wharf Structure’, Proceedings of the AUSTROADS 1997 Bridge Conference, Sydney,

Australia, December 1997.

26 Khatri R P, Guirguis S and Sirivivatnanon V ‘60 Years Service Life of Port Kembla Concrete

Swimming Pool’, Proceedings of Concrete 2001, Concrete Institute of Australia Biennial

Conference 2001, pp. 55–62.

27 Gjorv O E and Kashino N ‘Durability of a 60-year old reinforced concrete pier in Oslo

Harbour’ Materials Performance, Vol. 25, No. 2, pp. 18–26.

28 Sherman M R, McDonald D B and Pfeifer D W ‘Durability Aspects of Precast Prestressed

Concrete. Part 2: Chloride Permeability Study,’ PCI Journal, Vol. 41, No. 4, July-Aug, 1996,

pp. 76–95.

29 Philleo R E ‘Freezing and Thawing Resistance of High-Strength Concrete,’ NCHRP

Synthesis of Highway Practice 129, Transportation Research Board, 1986.

30 Power T C, Copeland L E and Mann H M ‘Capillary Continuity or Discontinuity in Cement

Paste,’ Journal of the PCA Research and Development Lab, Vol. 1, No. 2, 1959, pp. 38–48

(Reprinted as PCA R&D Bulletin 110 1988).

31 ‘QA Specification B 80 – Concrete Work for Bridges’, Edition 5, Revision 3, RTA, June 2008.

32 Torrent R J and Jornet A ‘The quality of the covercrete of low-, medium- and high-strength

concretes’, Proceedings Second CANMET/ACI Int. Conf. Durability of Concrete, Canada,

1991.


33 Cao H T et al., ‘Chloride Diffusion Resistance of Fly Ash Concrete’, CSIRO Report BIN 069,

ADAA and Pacific Power, October, 1993.

34 Vallini D and Aldred M ‘Durability Assessment of Concrete Specimens in the Tidal and

Splash Zones in Fremantle Port’, Coasts & Ports Australasian Conference, 2003.

35 BS 1881, Part 122 Test for determining the initial surface absorption British Standards

Institution 1983.

36 Wilson M A, Taylor S C and Hoff W D ‘The initial surface absorption test (ISAT): an analytical

approach’, Magazine of Concrete Research Vol. 50, No. 2, June 1998, pp. 179–85.

37 Dhir R K, Jones M R, Byars E A and Shaaban I G ‘Predicting concrete durability from its

absorption’, Third International Conference Durability of Concrete, SP-145, 1994,

pp. 1177–1194

38 Whiting D ‘Permeability of Concrete,’ SP-108, American Concrete Institute, 1988, pp. 195–222.

39 Condition Evaluation of Concrete Bridges Relative to Reinforcement Corrosion, Vol. 7

Methods for Field Measurement of Concrete Permeability, (SHRP-S/FR-92-109), National

Research Council, Washington, DC 1992.

40 Whiting D Rapid Determination of the Chloride Permeability of Concrete Final Report No.

FHWA/RD-81/119. Federal Highway Administration, August 1981, NTIS No. PB 82140724.

41 Sengul O. and Gjorv O E ‘Effect of Embedded Steel on Electrical Resistivity Measurements

on Concrete Structures’ ACI Materials Journal Vol. 106, No. 1, Jan-Feb 2009.

42 ‘Standard Test Method for Field Measurement of Soil Resistivity Using Wenner Four-

Electrode Method’’ ASTM G 57-78, 1984.

43 Baweja D, Roper H and Sirivivatnanon V ‘Specifications of Concrete for Marine

Environments: A Fresh Approach’, ACI Materials Journal, Vol. 96, No. 4, July 1999,

pp 462–470.

44 Condition Evaluation of Concrete Bridges Relative to Reinforcement Corrosion, Vol. 5

Methods for Evaluation the Effectiveness of Penetrating Sealers, (SHRP-S/FR-92-107)

National Research Council, Washington, DC 1992.

45 Schiebl P and Dauberschmidt C ‘Evaluation of Calcium Nitrite as a Corrosion Inhibitor’,

Supplementary Papers, Fifth CANMET/ACI Int. Conf. on Durability of Concrete, 2000,

pp 795–811.

46 Cusson D and Qian S ‘Corrosion inhibiting systems for concrete bridges – 10 years of field

performance evaluation’, Fifth Int. Conf. on Concrete under Severe Conditions Environment

and Loading, Tours, France, June 2007, pp. 1–10.


CCAA OFFICES

SYDNEY OFFICE:Level 6, 504 Pacific Highway St Leonards NSW Australia 2065POSTAL ADDRESS: Locked Bag 2010 St Leonards NSW 1590TELEPHONE: (61 2) 9437 9711 FACSIMILE: (61 2) 9437 9470

BRISBANE OFFICE: Level 14, IBM Building 348 Edward Street Brisbane QLD 4000TELEPHONE: (61 7) 3831 3288FACSIMILE: (61 7) 3839 6005

MELBOURNE OFFICE:2nd Floor, 1 Hobson Street South Yarra VIC 3141 TELEPHONE: (61 3) 9825 0200FACSIMILE: (61 3) 9825 0222

PERTH OFFICE:45 Ventnor Avenue West Perth WA 6005 TELEPHONE: (61 8) 9389 4452FACSIMILE: (61 8) 9389 4451

ADELAIDE OFFICE:PO Box 229 Fullarton SA 5063TELEPHONE: (61 8) 8274 3758

PrEMIxED CONCrEtE AND ExtrACtIVE INDUStrIES OFFICE PO Box 243Henley Beach SA 5022TELEPHONE: (61 8) 8243 2505FACSIMILE: (61 8) 8125 5822

TASMANIAN OFFICE: PO Box 246Sheffield tAS 7306TELEPHONE: (61 3) 6491 1509FACSIMILE: (61 3) 6491 2529

WEBSITE: www.ccaa.com.au

EMAIL: [email protected]

Disclaimer: Cement Concrete & Aggregates Australia is a not for profit organisation sponsored by the cement, concrete and aggregate industries in Australia to provide information on the many uses of cement, concrete and aggregates. this publication is produced by CCAA for that purpose. Since the information provided is intended for general guidance only and in no way replaces the services of professional consultants on particular projects, no legal liability can be accepted by CCAA for its use.

CCAA respects your privacy. Your details have been collected to provide you with information on our activities, publications and services. From time to time your details may be made available to third party organisations who comply with the Privacy Act such as affiliated associations, sponsors of events and other reputable organisations whose services we think you may find of interest. If you do not wish to receive information from CCAA or wish to be taken off the database please write to the Privacy Officer, CCAA, Locked Bag 2010, St Leonards, NSW, 1590

INTRODUCTIONThis Technical Note discusses the mechanisms of external sulfate attack which include chemical reactions and the physio-chemical effects on concrete. It reports the outcome of a research project conducted by Cement Concrete & Aggregates Australia (CCAA). In this project the performance of Australian concrete mixes, proportioned using sulfate-resisting cements (AS 3972 Type SR1) and non sulfate-resisting cements were evaluated in both neutral and acidic sulfate conditions. The results were examined in relation to the long-term concrete exposure data from the US Portland Cement Association (PCA) and the 40-year non-accelerated exposure programme at the US Bureau of Reclamation (USBR). Current specifications for sulfate resisting concrete in relevant Australian Standards and some Road Authorities' specifications are reviewed in the context of CCAA research findings which are applicable to concrete structures in sulfate or acid sulfate soil conditions.

1 SUlfaTe expOSURe CONDITIONS IN aUSTRalIaSulfates may occur naturally in soil and groundwater, in industrial effluents and wastes from chemical and mining industries, as well as in sea water. Acid sulfate soils are associated with naturally occurring sediments and soils containing iron sulfides usually found in mangroves, salt marsh vegetation or tidal areas and low lying parts of coastal floodplains, rivers and creeks.

Very few cases of aggressive sulfate soil and groundwater conditions have been reported in Australia. In certain sections of the Parramatta Rail Link in the Lane Cove Valley in Sydney, aggressive sulfate and carbon dioxide in groundwater were found; concretes with and without protective membrane were therefore used to satisfy the 100 years design life. The concrete at the base of water cooling towers have been found to be exposed to high sulfate levels in the closed circuit cooling systems. In the case of cooling towers at Bayswater Power Station in NSW, the concrete was found2 to show no sign of attack when inspected after 10 years of service. In the M5 East Motorway project at Cooks River Crossing near Kingsford Smith airport in Sydney, sulfate-resisting

Sulfate-resisting Concrete

Sulfate-resisting concrete

MAY 2011

68T

ec

hn

ica

ln

ote

page 2 Sulfate-resisting Concrete

concrete was used for the piles and diaphragm wall constructed in areas where the groundwater was found to have very high sulfate contents, possibly caused by effluents from the industrial areas around the Cooks River.

According to the NSW Acid Sulfate Soils Management Advisory Committee3, acid sulfate soils are soils containing highly acidic soil horizons or layers resulting from the aeration of soil materials that are rich in iron sulfides. The oxidation produces hydrogen ions in excess of the sediment's capacity to neutralise the acidity, resulting in soils of pH of 4 or less. The field pH of these soils in their undisturbed state is pH 4 or more and may be neutral or slightly alkaline. Organic acids are common in coastal ecosystems and can produce acid water and sediment. The pH of these sediments is usually around 4.5–5.5. As they do not have the ability to generate additional acid when exposed to air, they do not exhibit the same kinds of environmental risks that are associated with acid sulfate sediments.

In New South Wales, acid sulfate soil conditions have been reported by the Roads and Traffic Authority4 in the Pacific Highway upgrading programme, eg at the Chinderah Bypass which involved the dredging and disposal of potential acid sulfate soil from a site near a major bridge over the Tweed River at Barneys Point. Other locations include the floodplains of many rivers including Clarence River, Clyde River, Hawkesbury River, Hunter River, Macleay River, Manning River, Myall River, Nambucca River, Richmond River and Shoalhaven River. In Queensland, acid sulfate soils have also been found in the coastal regions including sulfide-bearing source rock and sodic soils which cover 45% of Queensland5. This has led Queensland Main Roads to draw designers' attention to detailed analysis of the chemistry of the soil and groundwater, and the design of concrete to withstand these potentially harsh conditions.

2 MeChaNISMS Of SUlfaTe aTTaCkThe deterioration of concrete exposed to sulfate is the result of the penetration of aggressive agents into the concrete and their chemical reaction with the cement matrix. The three main reactions involved are:■ Ettringite formation – conversion of hydrated

calcium aluminate to calcium sulphoaluminate,■ Gypsum formation – conversion of the calcium

hydroxide to calcium sulfate, and■ Decalcification – decomposition of the

hydrated calcium silicates.These chemical reactions can lead to

expansion and cracking of concrete, and/or the loss of strength and elastic properties of concrete. The form and extent of damage to concrete will depend on the sulfate concentration, the type of cations (eg sodium or magnesium) in the sulfate solution, the pH of the solution and the microstructure of the hardened cement

matrix. Some cements are more susceptible to magnesium sulfate than sodium sulfate, the key mechanism is the replacement of calcium in calcium silicate hydrates that form much of the cement matrix. This leads to a loss of the binding properties. Formation of brucite (Mg(OH)2) and magnesium silicate hydrates is an indication of such attack.

The presence of chloride in soil and groundwater may be beneficial since there is considerable evidence, from seawater studies6,7, that the presence of chloride generally reduces expansion due to sulfate attack. The risk of corrosion of embedded metals in buried concrete in non-aggressive soil is generally lower than in externally exposed concrete. However, high chloride concentrations in the ground may increase the risk of corrosion since chloride ions may permeate the concrete, leading to a depassivation of the metal surface.

Above the soil or water table level in the soil profile where the concrete surface is exposed to a wetting and drying condition, the concrete will also be subjected to a physio-chemical sulfate attack. Folliard and Sandberg8 reported that the physio-chemical process is more prevalent in the field, in which concrete is physically, rather than chemically attacked by sodium sulfate. The only reactions involved are within the sodium sulfate-water system; the phase changes from a solution to a solid, or from an anhydrous solid, thenardite (Na2SO4), to its hydrated form, mirabilite (Na2SO4.10H2O). The amount of deterioration is a function of the potential crystallisation pressures or the volume increase associated with a given mechanism. Any of the mechanisms can potentially produce pressures that are an order of magnitude greater than the tensile strengths of the concrete. Further, the same pressures can be reached by any one of several crystallisation mechanisms by simply varying the temperature and concentration of the sulfate solution in the system. The volume increase could cause severe deterioration of the concrete but may be partially accommodated in air-entrained concrete.

3 phySICal aND CheMICal ReSISTaNCe Of CONCReTeBoth the physical resistance of concrete to the penetration and capillary-induced migration of aggressive agents and the chemical resistance of the concrete to the deleterious reactions described above are important attributes of sulfate resisting concrete. Thus factors influencing the permeability and surface porosity of the concrete and the chemical resistance of cement are prime performance parameters of concrete exposed to sulfate attack.

The physical resistance of concrete is traditionally achieved by specifying mix design parameters such as maximum water–cement ratio and minimum cement content, while the

Sulfate-resisting Concrete page 3

chemical resistance is by the use of sulfate-resisting cement. This is the approach adopted in codes and guideline such as ACI 3189 and BRE Special Digest 110 and directly or indirectly in relevant Australian Standards. Recent research has focused on performance-based specification for sulfate resisting concrete. A specification based on water permeability was proposed by Sirivivatnanon and Khatri (1999)11. In this research, a rapid electrochemical test procedure similar to ASTM C 1202 Indication of Concrete's Ability to Resist Chloride Ion Penetration which was proposed by Tumidajski and Turc (1995)12 has been used to rapidly assess the ability of concrete to resist sulfate penetration. Long-term concrete performance tests are evaluated by CCAA to substantiate the validity of these approaches.

The role of concrete quality on the resistance to both the chemical and physio-chemical attack by sulfates has been studied by researchers at the PCA. It involved long-term exposure of concrete prisms in the laboratory and in the field. Findings have been reported by Verbeck (1967)13 and Starks (2002)14. Interestingly, it was found that a continuous immersion in sulfate solution was a relatively mild condition compared with cyclic wetting and drying. The physical resistance of the concrete to the physio-chemical sulfate attack was achieved by limiting the maximum water–cement ratio and minimum cement content of the concrete, and the application of a sealer to the surface of concrete.

4 aUSTRalIaN ReSeaRCh ON SUlfaTe‑ReSISTINg CONCReTeIn 2002, CCAA initiated a research project to develop a performance–based specification for sulfate-resisting concrete. The research was undertaken and completed in 2010. In this research project, nineteen concrete mixes were proportioned using six Type SR sulfate-resisting cements and two non sulfate-resisting cements, at water–cement ratios (w/c) of 0.4, 0.5 and 0.65. The concrete was proportioned with a fixed dosage of water–reducing admixture and a variable dosage

of superplasticiser to produce concrete with a slump of 120 ± 20 mm. The minimum cement contents were 415, 335 and 290 kg/m3 for the mixes with w/c of 0.4, 0.5 and 0.65 respectively. The concrete specimens were moist cured for three days and kept in the laboratory until testing at 28 days. (Hence there was a limited depth of carbonation at the surface of the concrete at the commencement of sulfate exposure.) Compressive strength, water permeability and rapid sulfate permeability of the concretes were determined at 28 days.

At 28 days, the concrete specimens were exposed by full immersion in 5% (50,000 ppm) sodium sulfate solutions maintained at pH of 7± 0.5 and 3.5 ± 0.5. The performance of the concrete was measured in terms of expansion of 75 x 75 x 285 mm duplicate prisms and strength retention of 100 mm x 200 mm duplicate cylinders throughout the exposure period of three years.

The 28 day compressive strength of the concrete varied widely from 45.5–75.5 MPa, 32.5–64.0 MPa and 29.5–37.0 MPa for w/c of 0.4, 0.5, and 0.65 respectively reflecting the influence of different cements. Results are shown in figure 1.

4.1 performance of buried concreteFrom previous CSIRO and PCA studies11,14 of long-term expansion of concrete immersed in sodium sulfate solution, an expansion performance limit of 220 microstrain per year within the first three years of exposure has been found to indicate long-term dimensional stability of the concrete. Small concrete specimens which maintain its 28-day strength within the first three years are indicator of good long-term strength retention. After three years of exposure, all Type SR cement concretes with water-cement ratios of 0.4 and 0.5 performed well both in terms of expansion and strength retention. As shown in figures 2 and 3 and Tables 1 and 2, all the concretes were stable in both neutral and acidic sulfate solutions with increases in expansion rate within the performance limit of 220 microstrain per year, and with strength retentions above 100% of the 28-day compressive strength. The results suggest that all concretes of 0.4 and 0.5 water–cement ratio, irrespective of the strength, will provide good resistance to sulfate attack in the long-term and could be classified as sulfate-resisting concretes.

In both pH 3.5 and 7 sulfate solutions, S2C and S3C (w/c 0.65) showed expansion rates significantly exceeding 220 microstrain per year during the first two years of exposure. Some S3C prisms were found to be badly cracked and expansion cannot be measured after two-year exposure as shown in plates 1 and 2.

The expansion performance limit was derived from a long-term study by the PCA of concretes exposed to accelerated field and laboratory-simulated sulphate environments reported by

C1 C2 C3 C4 C5 S1 S2 S3

CONCRETE DESIGNATION

28-D

AY C

OM

PRES

SIVE

STR

ENG

TH (M

Pa)

0

20

40

60

80w/c = 0.4w/c = 0.5w/c = 0.65

figure 1 28-day compressive strength of the concretes at w/c of 0.4, 0.5 and 0.65


Stark14. In Sacramento, California concrete prisms from 50 concrete mixtures were partially buried in sodium sulfate-rich soils, maintained at about 6.5% or 65,000 ppm sodium sulfate concentration, and exposed to cyclic immersion and atmospheric drying condition since 1989. The performance of the prisms was rated visually from 1.0 to 5.0 with a rating of 1.0 indicating excellent performance with virtually no evidence of deterioration, while a rating of 5.0 represented major loss of paste matrix and widespread exposure and loss of coarse aggregate particles. It was found that the main deterioration mechanism of concrete in this wetting and drying condition was due to the physio-chemical process of sulfate attack.

A second set of companion concrete prisms were immersed in a 6.5% or 65,000 ppm sodium sulfate solution in PCA's Construction Technology Laboratories (CTL) in Skokie, Illinois and their expansion monitored for over 12 years. All the concrete prisms were reported to perform very well after a 12-year exposure period. More importantly, all concrete with low expansion rate (within 220 microstain) per year in the first three years of exposure did not exhibit rapid increase in the rate of expansion in subsequent years nor did their maximum expansion reach 3000 microstrain – an elastic strain limit for most concrete. This PCA study concluded that sulfate resistance of concrete was mainly governed by water-cement ratios at w/c of 0.4 and below, whereas cement composition would influence the performance of concrete with intermediate w/c of 0.4 to 0.55.

plate 1 Failure of S3C (w/c 0.65) prisms after 570 days in 5% Na2SO4 at pH 7

plate 2 Failure of S3C (w/c 0.65) prisms after 570 days in 5% Na2SO4 at pH 3.5

plate 3 Cylinders after 3 years exposure prior to compression test. Grey in neutral sulfate solution (left). Rustic red in acidic sulfate solution (right)

0 200 400 600 800 1000 1200

DAYS IMMERSED

EXPA

NSI

ON

(mic

rost

rain

s)

0

500

1000

1500

2000pH 7, w/c = 0.65

S1CS2CS3CLimit

0 200 400 600 800 1000 1200

DAYS IMMERSED

EXPA

NSI

ON

(mic

rost

rain

s)

0

500

1000

1500

2000pH 7, w/c = 0.4

S1AS2AS3ALimit

0 200 400 600 800 1000 1200

DAYS IMMERSED

EXPA

NSI

ON

(mic

rost

rain

s)

0

500

1000

1500

2000pH 7, w/c = 0.5

S1BS2BS3BLimit

figure 2 Expansion of concrete prisms in 5% Na2SO4 solution at pH 7

0 200 400 600 800 1000 1200

DAYS IMMERSED

EXPA

NSI

ON

(mic

rost

rain

s)

0

500

1000

1500

2000pH 3.5, w/c = 0.65

S1CS2CS3CLimit

0 200 400 600 800 1000 1200

DAYS IMMERSED

EXPA

NSI

ON

(mic

rost

rain

s)

0

500

1000

1500

2000pH 3.5, w/c = 0.4

S1AS2AS3ALimit

0 200 400 600 800 1000 1200

DAYS IMMERSED

EXPA

NSI

ON

(mic

rost

rain

s)

0

500

1000

1500

2000pH 3.5, w/c = 0.5

S1BS2BS3BLimit

figure 3 Expansion of concrete prisms in 5% Na2SO4 solution at pH 3.5


The US Bureau of Reclamation (USBR) non-accelerated sulfate testing programme, on concrete cylinders partially submerged in 2.1% or 21,000 ppm sodium sulfate solution at ambient temperature, showed concrete with w/c ratio of 0.45 and lower to be intact even after 40-year exposure period15. The Bureau defined failure when expansion reached 0.5% or 5000 microstrain. The results also showed the importance of permeability and cement composition for concrete with w/c exceeding 0.45. USBR results support the validity of current service life performance specification.

It can be observed from Tables 1 and 2 that the compressive strength of the concrete increased well above the 28 day strength in the first 1–2 years of immersion, followed by a gradual reduction in strengths. After three-year exposure in both neutral and acidic sodium sulfate solutions, the strengths remained at or above the 28-day strength level for Type SR cement concretes with

water-cement ratios of 0.4 and 0.5. This clearly showed the integrity of the concrete and its mechanical resistance to sulfate attack.

4.2 performance of partly buried concreteWhile most buried concrete elements such as piles and footings are likely to be kept moist throughout their service life, parts of some of them (eg the top of footings and pile caps) may be exposed to periodic wetting and drying conditions. The PCA study confirmed that the exposure to alternate immersion and atmospheric drying in the sodium sulfate-rich soil was a more severe exposure condition than continuous immersion in the same solution. Attention must therefore be given to the sulfate resistance of concrete under such exposure conditions. Stark14 found a consistently improved trend in the rating of the surface deterioration of concrete with increased cement content irrespective of the type of cement. In the PCA's 17 concrete mixtures with a cement

Table 1 Retention of cylindrical compressive strength as % of 28-day strength in pH 7

C1 C2 C3 C4 C5 exposureperiod (days) 0.4 0.5 0.4 0.5 0.4 0.5 0.4 0.5 0.4 0.5

0 100 100 100 100 100 100 100 100 100 100 514 126 126 129 164 129 161 125 131 116 116 776 134 135 125 159 131 158 124 132 115 115 939 128 130 121 152 145 158 116 126 112 105 1240 118 123 122 161 136 159 117 130 111 101

S1 S2 S3

exposure 0.4 0.5 0.65 0.4 0.5 0.65 0.4 0.5 0.65period (days) S1A S1B S1C S2A S2B S2C S3A S3B S3C

0 100 100 100 100 100 100 100 100 100 365 132 139 154 103 103 112 93 110 87 570 130 133 149 99 99 111 95 90 10 1075 125 127 152 97 93 65 110 45 0

Table 2 Retention of cylindrical compressive strength as % of 28-day strength in pH 3.5

C1 C2 C3 C4 C5 exposureperiod (days) 0.4 0.5 0.4 0.5 0.4 0.5 0.4 0.5 0.4 0.5

0 100 100 100 100 100 100 100 100 100 100 514 125 130 136 151 146 158 117 137 125 123 776 120 116 141 153 136 151 109 128 118 109 939 120 122 138 151 135 146 113 129 115 106 1240 123 127 136 163 131 149 111 136 115 108

S1 S2 S3

exposure 0.4 0.5 0.65 0.4 0.5 0.65 0.4 0.5 0.65period (days) S1A S1B S1C S2A S2B S2C S3A S3B S3C

0 100 100 100 100 100 100 100 100 100 365 119 123 156 87 97 112 90 102 90 570 120 128 136 77 95 97 102 86 6 1075 101 116 135 90 91 87 93 24 11

The conditions of the concrete cylinders after 3-years exposure were quite varied with most retaining their integrity but some were badly cracked especially around the top edges. See Plate 3 showing contrast in colour of cylinders after 3-year exposure.


content of 390 kg/m3, most concretes had a rating between 1.4 and 3.8 after 12–years exposure in the sulfate-rich soil ground in Sacramento. This is considered to be a good performance of the concrete under such an aggressive sulfate environment. Stark found that the observed severe deterioration in the outdoor exposure was due largely to cyclic crystallisation of NaSO4 salts after sufficient evaporation of moisture from the outdoor soils exposure as postulated by Folliard and Sandberg8. This is probably the reason for the effectiveness of a sealer, such as silicon and linseed oil, in limiting the capillary-induced migration of sulfate, and thus improving the performance of concrete including concrete with higher w/c of 0.49–0.52.

With all Type SR cement concrete mixes performing exceedingly well under full immersion in sodium sulfate solutions at both neutral and acidic conditions, and a minimum cement content of 415 kg/m3 in the 0.4 w/c series, it is likely that the low water–cement ratio concretes will also perform very well in the severe wetting and drying condition. With appropriate surface protection, the 0.5 w/c series of concrete with a minimum cement content of 335 kg/m3 would also be expected to perform well in the more aggressive wetting and drying condition.

5 SpeCIfyINg SUlfaTe‑ReSISTINg CONCReTeSulfate-resisting concrete has traditionally been specified prescriptively by the type of cement and mix proportion limits in terms of maximum water-cement ratio and minimum cement content. In highly acidic and permeable soils where pH is below 3.5, additional protective measures are required to isolate the concrete from direct contact with the aggressive ground condition. ACI 318 9 and BRE SD110 are examples of these specifications. BRE SD1 is particularly progressive in recommending specifications for sulfate-resisting concrete for intended working life of 50 years for building works and 100 years for civil engineering structures.

5.1 australian StandardsIn the revision of the Australian Standard for concrete structures AS 360016, specifications for concrete in sulfate soils with a magnesium content of less that 1000 mg/L have been introduced. For each exposure classification, concrete is specified in terms of concrete grade and minimum concrete cover, see Table 3. The current Australian Standards for piling, AS 215917 and for concrete structures for retaining liquids, AS 373518, recommend the specification of certain concrete grades and corresponding covers for a design life 40–60 years in a range of exposure classifications. The exposure classification is defined by the magnitude of sulfate in the soil or in groundwater, pH level and the soil conditions in term of its permeability. In severe and very severe conditions, where sulfate levels exceed 2000 ppm in groundwater or 1% in soil, AS 3735 Supplement19 recommends a minimum cement content of 320 kg/m3 and a maximum water–cement ratio of 0.5 and the use of Type SR cement.

The findings from CCAA research, described in the previous section, supported the specification of sulfate resisting concrete by strength grade and cover (AS 2159 and AS 3600), and in particular, confirmed the expected performance of the sulfate resisting concrete in the moderate (B2) and severe to very severe (C1 and C2) exposure classifications shown in Table 3.

It should be noted that the Australian Standard for bridge design AS 5100 20 provides no specific guidance on specifying concrete for 100 years design life in sulfate conditions.

5.2 Other specificationsRoad authorities, such as the RTA in New South Wales and the Queensland Department of Main Roads, are specifying sulfate-resisting concrete based on exposure classifications in Austroads Bridge Design Code (superseded by AS 5100), but with additional limits on maximum w/c and minimum cement content. Queensland Department of Main Roads refers to MRS11.70 with the additional requirements shown in Table 4.

Table 3 Strength and cover requirements for sulfate soils(Summarised from Tables 4.8.1 and 4.10.3.2 in AS 3600—2009)

SO4 Characteristic MinimumIn groundwater In soil exposure strength cover(mg/L) (%) classification (MPa) (mm)

<1000 < 0.5 A2 25 50 1000–3000 0.5–1 B1 32 501

3000–10,000 1–2 B2 40 501,2

>10,000 > 2 C1 and C2 ≥ 50 651,2,3

Notes:1 It is recommended that cement be Type SR.2 Additional protective coating is recommended.3 The cover may be reduced to 50 mm if protective coating or

barriers are used.

Table 4 Additional requirements (From Table 8 of QDMR MSR11.70)

Minimum Maximum Strengthexposure cementitious water–cementitious gradeclassification content (kg/m3) ratio (MPa)

B1 320 0.56 32B2 390 0.46 40C 450 0.40 50


As can be noted, the findings from CCAA research project also support the above specifications.

5.3 performance‑based specificationsSulfate resisting concrete has traditionally been specified prescriptively by the maximum water–cement ratio and a specific type of SR cement. This is to ensure good physical resistance of the concrete to limit the penetrating sulfate ions, and good chemical resistance of the cement matrix to the deleterious sulfate reactions. A performance specification based on water permeability of the concrete has been proposed by Sirivivatnanon and Khatri 11. As part of CCAA research, a further attempt has been made to develop a performance-based specification for sulfate resisting concrete based on the physical resistance of the concrete (eg water permeability, rapid sulfate permeability) and the chemical resistance of the cement (sulfate expansion). A six-hour accelerated test method for a rapid sulfate permeability determination was developed and is shown in figure 4.

The dimension stability and strength retention properties of the nineteen concrete mixes were evaluated in accordance with the criteria established in Section 4.1. Concrete passing both expansion and strength retention criteria is considered sulfate-resisting concrete. The mix prescription in water-cement ratio, and

performance properties: water permeability coefficient and rapid sulfate permeability; are summarised in Table 5 and shown in figures 5 and 6.

Based on these properties, a semi-prescriptive and performance-based specifications for sulfate-resisting concrete are proposed as follows.1 Type SR cement and water-cement ratio ≤ 0.5,

and2 Type SR cement and a water permeability

coefficient ≤ 2x10-12 m/s or rapid sulfate permeability ≤ 2000 coulombs.The result also showed that concrete with

Type SR cement at w/c greater than 0.5 or non sulfate-resisting cement at very low w/c no greater than 0.4 can perform well especially in neutral sulfate condition.

0.3N NaOH

Thermocoupleand gas aperture

Metal mesh Machinedperspexchamber

+ 60 V 0 VInlet

Testspecimen

8.8% Na2SO4

figure 4 An accelerated test set-up for the rapid sulfate permeability determination

figure 5 Water permeability of the concretes at w/c of 0.4, 0.5 and 0.65

figure 6 Rapid sulfate permeability of the concretes at w/c of 0.4, 0.5 and 0.65

C1 C2 C3 C4 C5 S1 S2 S3

CONCRETE DESIGNATIONW

ATER

PER

MEA

BILI

TY (1

0–12 m

/s)

0

3

2

1

4

5w/c = 0.4w/c = 0.5w/c = 0.65

C1 C2 C3 C4 C5 S1 S2 S3

CONCRETE DESIGNATION

RAP

ID S

ULF

ATE

PER

MEA

BILI

TY (c

oulo

mb)

0

3000

2000

1000

4000w/c = 0.4w/c = 0.5w/c = 0.65

Table 5 Summary of long-term performance and possible specifications

C1–C5 S1 S2 S3 Concreteproperties 0.4 0.5 0.4 0.5 0.65 0.4 0.5 0.65 0.4 0.5 0.65

Water permeability(x10–12 m/s) 0.07–0.28 0.34–1.70 0.16 1.5 70.3 0.14 0.35 13.4 0.13 0.44 16

Rapid sulfate permeability(coulombs) 940–1260 1180–1450 1475 1965 2260 2580 3225 4010 1780 2265 3060

Water-to-cement 0.40–0.41 0.50 0.39 0.50 0.63 0.39 0.5 0.66 0.40 0.50 0.66

28-day compressive strength(MPa) 47.5–75.5 32.5–59.0 52.5 49.5 29.5 68.0 64.0 37.0 68.0 58.0 34.5


MAY 2011

68

6 CONClUSIONSSulfate resistance of the concrete is a function of its physical and chemical resistance to penetrating sulfate ions. Good physical resistance of the concrete is directly related to the water-cement ratio and the cement content. Good chemical resistance is related to the resistance of the cement matrix to the deleterious sulfate reactions.

Sulfate-resisting concrete can be achieved using a sufficient quantity of a sulfate-resisting cement (Type SR complying with AS 3972) and a low water–cement ratio to obtain a concrete with low water permeability. For fully buried concrete structures in saturated soils, a sulfate-resisting concrete can be achieved from Type SR cement at a cement content of 335 kg/m3 and a water–cement ratio of 0.5. For partially buried structures exposed to a wetting and drying condition, the same sulfate-resisting concrete can be used but with additional protective measure such as the application of an appropriate sealer to the surface of the exposed concrete. Alternatively, a sulfate-resisting concrete can be achieved from Type SR cement at a cement content of 415 kg/m3 and a water–cement ratio of 0.4. The AS 3600 specifications for concrete structures in acid sulfate soils, based on minimum compressive strength and Type SR cement, is shown to produce adequate sulfate-resisting concrete for the exposure condition indicated. Alternatively, performance-based specifications based on Type SR cement and a concrete with a limit on either water permeability or rapid sulfate permeability can be used.

7 RefeReNCeS1 AS 3972 General purpose and blended cements.

Standards Australia, Sydney 2010.2 Sulphate Attack on Concrete. Fly Ash Technical

Notes No. 2, Ash Development Association of Australia, 1995

3 Acid Sulfate Soils Assessment Guidelines, NSW Acid Sulfate Soils Management Advisory Committee, 1998.

4 Selim H, Forster G and Chirgwin G 'Concrete Structures in Acid Sulfate Soils', Proceedings of Austroads Bridging the Millennia Conference, edited by G.J. Chirgwin, Vol. 2, Sydney, 1997,

pp 393–409.5 Carse A 'The design of concrete road structures

for aggressive environments', Proceedings of Concrete Institute of Australia 22nd Biennial Conference, Melbourne, 2005.

6 Lea F M Chemistry of Cement and Concrete. 4th Edition, Elsevier, London, 2004.

7 Biczok I Concrete Corrosion: Concrete Protection. 8th edition, Akademiai Kiado, Budapest, 1980.

8 Folliard K J and Sandberg P 'Mechanisms of Concrete Deterioration by Sodium Sulfate Crystallization', American Concrete Institute,

SP145, Durability of Concrete, Third International Conference, Nice, France, pp 933–945.

9 Building Code Requirements for Reinforced Concrete, ACI 318–05, American Concrete Institute.

10 Building Research Establishment. Concrete in aggressive ground, BRE Special Digest 1:2005.

11 Sirivivatnanon V and Khatri R P 'Performance Based Specification for Sulphate Resisting Concrete,' Proceedings of International Conference on a Vision for the Next Millennium, edited by R N Swamy, Sheffield, UK, June 1999, pp 1097–1106.

12 Tumidajski P J and Turc I, A Rapid Test for Sulfate Ingress into Concrete, Cement and Concrete Research, Vol. 25, No. 5, 1995, pp 924–928.

13 Verbeck G J 'Field and Laboratory Studies of the Sulphate Resistance of Concrete', Proceeding of a symposium in honour of Thorbergur Thorvaldson, Candana, 1967, 113–124.

14 Stark D C 'Performance of Concrete in Sulfate Environments, RD129, Portland Cement Association, (USA), 2002

15 Monteiro P J M and Kurtis K E, 'Time to failure for concrete exposed to severe sulfate attack', Cement and Concrete Research, Vol. 33, No. 7, 2003 pp 987–993.

16 AS 3600 Concrete Structures, Standards Australia, Sydney, 2009.

17 AS 2159 Piling – Design and installation, Standards Australia, Sydney, 1995.

18 AS 3735 Concrete structures for retaining liquids, Standards Australia, Sydney 2001.

19 AS 3735 Concrete structures for retaining liquids Supplement – Commentary, Standards Australia, Sydney, 2001.

20 AS 5100 Australian Standard for Bridge Design, Standards Australia, Sydney, 2004.

CCAA OFFICES

SYDNEY: (61 2) 9437 9711 BRISBANE: (61 7) 3227 5200MELBOURNE: (61 3) 9825 0200PERTH: (61 8) 9389 4452ADELAIDE: (61 8) 8274 3758TASMANIA: (61 3) 6491 1509

WEBSITE: www.ccaa.com.au

EMAIL: [email protected]

Disclaimer: Cement Concrete & Aggregates Australia is a not for profit organisation sponsored by the cement, concrete and aggregate industries in Australia to provide information on the many uses of cement, concrete and aggregates. This publication is produced by CCAA for that purpose. Since the information provided is intended for general guidance only and in no way replaces the services of professional consultants on particular projects, no legal liability can be accepted by CCAA for its use.

CCAA respects your privacy. Your details have been collected to provide you with information on our activities, publications and services. From time to time your details may be made available to third party organisations who comply with the Privacy Act such as affiliated associations, sponsors of events and other reputable organisations whose services we think you may find of interest. If you do not wish to receive information from CCAA or wish to be taken off the database please write to the Privacy Officer, CCAA, Locked Bag 2010, St Leonards, NSW, 1590

Guidelines for

the Use of High Volume Fly Ash Concretes

DBCE Technical Report TR95/2

Prepared by

V. Sirivivatnanon, H. T. Cao, R. P. Khatri and L. Bucea

CSIRO Division of Building, Construction and Engineering

14 March, 2012

Table of Content

1. 1

1. INTRODUCTION 1

1.1 Optimum Mix Design 1

1.2 Specifications 1

1.3 Nomenclatures 1

2. DEVELOPMENTS 1

2.1 Overseas 1

2.2 Australia 1

3. MIX DESIGN AND FRESH CONCRETE PROPERTIES 1

3.1 Mix Design 1

3.1.1 Compressive Strength and Water-to-Binder Ratio 2

3.1.2 Water Demand 2 3.1.3 Aggregate Grading 3

3.2 Consistency, Setting Times and Early Strength 3

3.3 Effect of Sources of Materials 4

4. HARDENED CONCRETE PROPERTIES 4

4.1 Mechanical Properties 4

4.2 Volume Stability 4

4.3 Durability Properties 6

4.3.1 Corrosion of Steel Reinforcement 6 4.3.2 Sulphate Resistance 9

4.3.3 Optimum Dosage for Durability 11

5. APPLICATIONS 11

5.1 Suitability based on Inherent Properties 11

5.2 Built Structures 11

5.3 Current Developments 12

6. SUMMARY 12

7. ACKNOWLEDGMENT 12

8. REFERENCES 13

1

Guidelines for the Use of High Volume Fly Ash Concretes

1. INTRODUCTION

Australian experience with the use of fly ash and slag in concrete has been reviewed comprehensively by Ryan

1. The use of fly ash in structural concrete dates back to 1958 when a monolith crest section of a concrete gravity dam structure was built with concrete using fly ash from Pyrmont Power Station in Sydney. The acceptance of the use of fly ash in concrete has now reached a point where it has been claimed that 90% of concrete produced in Sydney contains fly ash. The traditional dosage of fly ash in typical concrete mixes varies from 60-100 kg/m3. This represents about 20 percent or less by mass of the total binder used. This typical dosage probably reflects the optimum fly ash content in terms of cost related to compressive strength ($/MPa). However, when durability is of prime concern, the optimum dosage would need to be re-examined.

High Volume Fly Ash (HVFA) concrete usually refers to structural concrete with fly ash content substantially higher than that used in conventional fly ash concretes. In Australia, concrete with fly ash making up to 40 percent by weight (wt.%) or 50 percent by volume of binder can be classified as HVFA concrete.

1.1 Optimum Mix Design

Most codes of practice, both past and present, call for compressive strength in both specification and compliance requirements. The adopted philosophy has been in using fly ash as a portland cement replacement with its contribution to strength measured in terms of an equivalent cement ratio. A simple procedure has also been developed by Butler

2 to optimise the amount of replacement material in terms of the binder cost.

In these guidelines, it is emphasised that a balance between both compressive strength and durability

properties should be adopted. The appropriate means of selecting a concrete to benefit a specific application is based on life-cycle costing. In most cases, the cost/MPa of HVFA will be higher than portland cement concrete. Cost benefit will usually be derived from a longer service life.

1.2 Specifications

The current Australian Standard for Concrete Structures, AS 3600, recognises the use of Supplementary Cementitious Materials (SCM) such as fly ash and ground granulated blast furnace slag (slag). AS 3600 uses the term cement to describe both portland cement and fly ash or slag blended cements. It is interesting to note that the minimum percentage by mass of portland cement in the binders is given in a formula which translates into the use of no more than 40 wt.% fly ash or 65 wt.% slag of the total binder.

1.3 Nomenclatures

Cement is a term used to describe portland cement only. This differs from its normal use in AS 3600 to describe all types of binders.

Binder is used to describe the cementitious material which could be portland cement or combinations of portland cement and other supplementary cementitious materials.

Supplementary Cementitious Material (SCM) describes materials such ash fly ash or slag as defined in AS 3582.

2. DEVELOPMENTS

The development of structural grade HVFA concrete follows the successful development and use of roller compacted concrete (RCC) in dam and hard core construction. RCC is basically a low grade concrete,, with a large volume of fly ash used for improving the workability and to lower the heat of hydration of the concrete. It has a consistency which enables its placing by asphaltic concrete placing technique.

2.1 Overseas

Structural grade high fly ash content concrete has been trialed at Didcot Power Station in 19813. The Canadian Centre for Mineral and Energy Technology (CANMET) has carried out a major research project developing high volume fly ash concrete since 1985. CANMET has adopted the approach of producing HVFA concrete with 56 wt.% fly ash (ASTM C618 Type F) and a relatively high dosage of superplasticizer to achieve the required consistency. Work is continued at CANMET on the durability and fracture toughness of HVFA concretes.

2.2 Australia

In Australia, The development of structural grade concrete with a high volume of fly ash (HVFA) began at the CSIRO in 1991. Concretes with fly ash contents in the binder fraction between 30 and 70 percent by weight were investigated. The results of this work have been reported by Sirivivatnanon et al.4,5. Work is continuing on the long term durability properties in high chloride and sulphate environments. A major field trial was successfully completed in 1992 and reported by Nelson

et al.6. HVFA concrete is gaining wider acceptance in

the industry. The largest volume produced so far was a 40,000 m3 of 40 wt.% fly ash, 40 MPa concrete used for the construction of the basement of the Melbourne Casino project in 1995.

3. MIX DESIGN AND FRESH CONCRETE

PROPERTIES

3.1 Mix Design

2

HVFA concretes can be designed using conventional mix design philosophy such as the ACI method7 or those originated from Road Note No. 48.

In designing mixes with the use of a chemical admixture, two possible approaches can be adopted. The first is by the use of a fixed dosage of admixture such as a water reducer or a superplasticizer. The second is by a variable dosage of superplasticizer beyond the optimum dosage to contain the amount of free water and binder. In the fixed dosage approach, the optimum dosage of chemical admixture was derived from the cost of the admixture to achieve the minimum water demand for a particular binder system. The optimum dosage of the chemical admixture becomes the fixed dosage used in subsequent concrete mixes. The variable dosage is only relevant when a superplasticizer is used. It was found that water demand reduces with increasing superplasticizer dosage.

Mix design data given in these guidelines are those based on the fixed dosage of chemical admixture. They enable the design of concrete mixes with a reasonable amount of free water and hence a workability which does not differ significantly from conventional concrete. The increased

dosage technique is useful in designing concrete mixes with low heat of hydration by limiting the amount of binder used.

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.90

10

20

30

40

50

60

70

Water-to-Binder

28-day Compressive Strength (MPa)

Portland cementconcrete

40 wt.% FA

50 wt.% FA

Grade 25Mean 32 MPa

Figure 1 28-day compressive strength to water-to-binder

ratio relationship

3.1.1 Compressive Strength and Water-to-Binder

Ratio

The compressive strength of HVFA concrete can be related to water-to-binder or water-to-cement ratio in a similar manner to portland cement. In Figure 1, the relationships between 28-day compressive strength and the water-to-binder ratio of HVFA concretes manufactured from Vales Point fly ash, a GP portland cement, and a 20 mm maximum size crushed basalt aggregate are given. While these relationships are distinguishable for different percentages of fly ash in the binder, they appear to be independent of the type of

chemical admixture used. At a fixed water-to-binder ratio, lowering the percentage of fly ash results in an increase of corresponding strength of the concrete.

3.1.2 Water Demand

The water demand is found to depend on the type of chemical admixture used and the total binder content. The optimum dosage with respect to water demand of the two water reducers examined, GWR and Pozzolith 370, are found to be 0.4 litre per 100 kilograms of binder which is in the upper range of the recommended dosage for portland cement concrete. In the case of superplasticizers, Daracem or Rheobuild 1000, the water demand was reduced with increased dosage as shown in Figure 2. However, the optimum dosage with respect to workability (not just slump) was found to be around one litre per 100 kilograms of binder. A combined dosage of 0.4 litre of Pozzolith 370 and 0.7 litre of Rheobuild 1000 is as effective as the use of 1.0 litre of Rheobuild 1000 in the control of water demand. With the current admixture price structure, the combined admixtures represents the cheapest way to effect high range water reduction.

Typical free water demands are given in Table 1. It has been found that the water demand depended more on the binder content than on the percentage of the fly ash in the binder.

Table 1 Approximate free-water content (kg/m3) required to produce HVFA concrete with 20 mm maximum size crushed aggregate and natural sand.

Chemical Admixture

Water Reducer Superplasticizer

Slump, mm

Binder, kg/m3 Binder, kg/m3

450 550 650 450 550 650

50-100 160 180 205 140 145 160

100-150 180 200 225 160 165 180

500 1,000 1,500 2,000 2,500 3,000120

130

140

150

160

170

180

Superplasticizer Dosage (ml/100 kg binder)

Free Water Demand (litre/m3)

3

Figure 2 Effect of increasing superplasticizer dosage on the water demand of HVFA concrete at a fixed slump.

3.1.3 Aggregate Grading

It was found that conventional methods of selecting the composition of aggregates to give a good total grading worked well with HVFA concretes. Combined grading such as those recommended in Road Note No. 49 and given in Figure 3 could be used as a guide. The coarser zones B and C can be used for mixes with high binder contents, but a combined grading such as 555G which falls completely within one zone (zone B) appears to be more important in producing good workable HVFA concretes. The effect of grading on workability has been described in details elsewhere by Sirivivatnanon et al.10.

Percentage Passing

0.075 0.15 0.3 0.6 1.18 2.36 4.75 9.5 19

0

10

20

30

40

50

60

70

80

90

100

Zone C

Zone B

Zone A

0.075 0.15 0.3 0.6 1.18 2.36 4.75 9.5 19

0

10

20

30

40

50

60

70

80

90

100

555G

555F

555E

555E

555F

555G

Sieve Size, mm

Road Note No. 4

Figure 3 Combined aggregate grading with respect to recommended grading zones

3.2 Consistency, Setting Times and Early Strength

HVFA concretes generally contain higher binder contents than equivalent portland cement concrete. This usually results in fresh HVFA concretes which are more cohesive or sometimes very sticky. These concretes tend to give a lower flow value than portland cement concretes of similar slump as shown in Figure 4.

40 60 80 100 120 1400

100

200

300

400

500

Slump (mm)

Flow - BS1881 part105 (mm)

PorlandCement HVFA

Figure 4 Relationship between flow and slump of portland cement and HVFA concretes.

There is usually some delay in the setting times of HVFA concrete compared to portland cement concrete. The extent of the delay depends on the particular cement and fly ash combination. In Figure 5, the setting times of Waurn Pond (WP) cement and its combinations with 40 wt.% fly ashes from Eraring (E) and Vales Point (VP) are given. The delays in initial and final set are of the order of one and one and a half hours respectively. While these lengths of delay are quite acceptable in most applications, the use of a certain type of chemical admixture with a particular cement / fly ash combination could cause an unacceptable length of delay. Precaution should therefore be taken in checking the compatibility between all concreting materials prior to the production of HVFA concretes.

WP WP/E WP/VP0

100

200

300

400

500

Setting Times (Minutes)

Initial set

Final set

Figure 5 Initial and final setting times of concrete made from Waurn Pond cement and its combinations with Eraring or Vales Point fly ash.

The early strength development of HVFA concrete is found to be slightly lower than portland cement concrete as shown in Figure 6. The 3-day to 28-day compressive strength ratio was 0.61 and 0.53 for portland cement and HVFA concretes respectively.

4

0 5 10 15 20 25 300

10

20

30

40

50

Age of Curing, days

Compressive St rength ( MPa)

WP WP/ E WP/ VP

Figure 6 Early compressive strength development.

3.3 Effect of Sources of Materials

Like any other conventional concretes, the properties of HVFA concrete do vary with the source of portland cement and fly ash used. Examples of the effect of the source of materials on the setting times and compressive strength have been published5. The influence of the source of materials on durability properties is available 11,12 and will be published in future.

4. HARDENED CONCRETE PROPERTIES

Three specific types of hardened concrete properties are of interest to engineers. They are mechanical, volume stability and durability properties. In evaluating these properties of HVFA concrete, comparisons are usually made to portland cement concrete of equivalent 28-day compressive strength.

For specific structural grades, the nominal water-to-binders of portland cement and HVFA concrete are given in Table 2. It should be noted that HVFA concretes have W/B ratio ranging from 0.12 to 0.16 below portland cement of equivalent grade. These differences will prove to be of significance in durability performance as discussed in a subsequent section.

Table 2 Water-to-binder ratio of portland cement and HVFA concrete at corresponding grade.

Structural Grade

Portland cement concrete

40% high volume fly ash concrete

25 0.64 0.48

32 0.56 0.41

40 0.49 0.35

50 0.40 0.28

Typical mix designs of two series of concretes are given

in Table 3. The first is a low 50±15 mm slump (L series) HVFA and portland cement concretes designed for pavements (and other slab applications). The others are

pump mixes with 100±25 mm slump (H series) HVFA and portland cement concretes developed for other structural works.

4.1 Mechanical Properties

TThhee mmeecchhaanniiccaall pprrooppeerrttiieess ooff tthhee mmiixxeess ggiivveenn iinn TTaabbllee 33

aarree ssuummmmaarriisseedd iinn TTaabblleess 44 && 55.. TThhee rreessuullttss iinnddiiccaatteedd

tthhaatt tthhee fflleexxuurraall ssttrreennggtthh ((FFff)) aanndd eellaassttiicc mmoodduulluuss ((EE)) ooff

HHVVFFAA ccoonnccrreetteess aarree ssiimmiillaarr ttoo ppoorrttllaanndd cceemmeenntt

ccoonnccrreetteess ooff eeqquuiivvaalleenntt 2288--ddaayy ccoommpprreessssiivvee ssttrreennggtthh ((FFcc))..

HHVVFFAA ccoonnccrreetteess ccaann tthhuuss bbee uusseedd ffoorr ccoonnccrreettee ssttrruuccttuurreess

iinn tthhee ssaammee mmaannnneerr aass ppoorrttllaanndd cceemmeenntt ccoonnccrreetteess..

TThheeiirr eellaassttiicc pprrooppeerrttiieess ccaann aallssoo bbee pprreeddiicctteedd ffrroomm tthhee

ccoommpprreessssiivvee ssttrreennggtthh aanndd ddeennssiittyy ((ρρ)) iinn tthhee ssaammee mmaannnneerr

aass ppoorrttllaanndd cceemmeenntt ccoonnccrreetteess ggiivveenn iinn AASS33660000 1133..

FFoorr HHVVFFAA ccoonnccrreetteess

FFff == 00..6699 ((FFcc))11//22

EE == 00..005599 ρρ11..55((FFcc))11//22

FFoorr ppoorrttllaanndd cceemmeenntt ccoonnccrreetteess

FFff == 00..7700 ((FFcc))11//22

EE == 00..005566 ρρ11..55((FFcc))11//22

..

4.2 Volume Stability

DDrryyiinngg SShhrriinnkkaaggee

HHVVFFAA ccoonnccrreetteess ccaann hhaavvee aa ssiimmiillaarr oorr uupp ttoo 2200%% lloowweerr

ddrryyiinngg sshhrriinnkkaaggee tthhaann ppoorrttllaanndd cceemmeenntt ccoonnccrreetteess.. TThhee

rreedduuccttiioonnss iinn sshhrriinnkkaaggee aarree mmoorree ssiiggnniiffiiccaanntt iinn ccoonnccrreetteess

ooff lloowweerr ggrraaddeess.. TTyyppiiccaall ddrryyiinngg sshhrriinnkkaaggeess aatt 5566 ddaayyss ffoorr

bbootthh tthhee llooww aanndd hhiigghh sslluummpp ccoonnccrreetteess aarree ggiivveenn iinn

TTaabblleess 44 aanndd 55.. TThhee sshhrriinnkkaaggee vvaalluueess aarree wweellll bbeellooww

770000xx1100--66 rreeccoommmmeennddeedd iinn AASS 33660000 TThhee iinnccrreeaassee iinn

ddrryyiinngg sshhrriinnkkaaggee wwiitthh ttiimmee uupp ttoo 9911 ddaayyss ffoorr ppoorrttllaanndd

cceemmeenntt aanndd HHVVFFAA ccoonnccrreettee iiss sshhoowwnn iinn FFiigguurree 77 ffoorr tthhee

ppuummpp mmiixx HH--sseerriieess.. TThhee rreedduuccttiioonn iinn sshhrriinnkkaaggee ooff

HHVVFFAA ccoonnccrreettee ccaann bbee oobbsseerrvveedd aass eeaarrllyy aass 2288 ddaayyss ffoorr

tthhee lloowweerr ggrraaddee 2200 ccoonnccrreettee.. TThhiiss ttrreenndd rreemmaaiinnss uupp ttoo 9911

ddaayyss aanndd bbeeyyoonndd..

CCrreeeepp CChhaarraacctteerriissttiiccss

TThhee ccrreeeepp cchhaarraacctteerriissttiiccss ooff ccoonnccrreettee aarree ssiiggnniiffiiccaannttllyy

iimmpprroovveedd wwiitthh tthhee uussee ooff hhiigghh vvoolluummee ooff ffllyy aasshh aass

sshhoowwnn iinn FFiigguurree 88.. AA ccrreeeepp rraattee,, FF((KK)),, iiss ddeetteerrmmiinneedd

ffrroomm tthhee ssllooppee ooff tthhee lliinnee rreellaattiinngg ccrreeeepp ssttrraaiinn ppeerr uunniitt

ssttrreessss ttoo tthhee nnaattuurraall llooggaarriitthhmm ooff ttiimmee llooggee((tt++11)) wwhheerree tt iiss

tthhee ttiimmee ooff llooaaddiinngg iinn ddaayyss.. TThhee ccrreeeepp rraatteess wweerree

rreedduucceedd bbyy 4400 aanndd 5555 ppeerrcceenntt iinn ggrraaddee 3322 aanndd 2200 HHVVFFAA

ccoonnccrreetteess rreessppeeccttiivveellyy ccoommppaarreedd ttoo ppoorrttllaanndd cceemmeenntt

ccoonnccrreetteess ooff eeqquuiivvaalleenntt 2288--ddaayy ccoommpprreessssiivvee ssttrreennggtthh aass

sshhoowwnn iinn TTaabbllee 55..

TThheerree ccaann tthheerreeffoorree bbee cclleeaarr aaddvvaannttaaggeess iinn tthhee uussee ooff

HHVVFFAA ccoonnccrreetteess ffoorr ssttrruuccttuurraall mmeemmbbeerrss wwhhiicchh aarree

sseennssiittiivvee ttoo hhiigghh ccrreeeepp ssttrraaiinn ssuucchh aass lloonngg ssppaann bbrriiddggee

ggiirrddeerrss aanndd ccoolluummnnss iinn hhiigghh rriissee bbuuiillddiinnggss..

5

Table 3 Mix design of concretes given in kilogram per cubic metre. Aggregates at s.s.d. All mixes had either a water reducing agent (WRA) or a superplasticizer (SP) at a dosage of 0.4 and 1.0 litre per 100 kg of binder respectively

Mix

Desig-

nation

Cement Fly

Ash

Free

Water

20 mm

Agg.

10 mm

Agg.

Sand

Admixture

W/B

W/C

200AL 245 0 170 605 605 820 WRA 0.7 0.7

205AL 185 185 160 605 605 680 WRA 0.43 0.87

320AL 315 0 175 600 595 765 WRA 0.57 0.57

324AL 235 160 160 615 610 645 WRA 0.41 0.68

450BL 355 0 155 620 605 770 SP 0.43 0.43

454BL 265 175 135 730 480 660 SP 0.31 0.51

200AH 270 0 175 590 580 810 WRA 0.65 0.65

205AH 200 200 165 580 575 680 WRA 0.41 0.82

320AH 335 0 195 570 570 770 WRA 0.58 0.58

324AH 255 170 180 565 560 655 WRA 0.43 0.72

450BH 340 0 160 620 615 760 SP 0.47 0.47

454BH 340 225 170 535 535 585 SP 0.3 0.5

Table 4 Mechanical and drying shrinkage properties of the low slump HVFA and portland cement concrete of equivalent 28-day compressive strength

Mix

Desig-

nation

Fly

Ash

Grade

Slump

Flow

Compressive

Strength,

(MPa)

28-day

Elastic

Modulus

28-day

Flexural

Strength

Drying

Shrinkage

at 56 days

% MPa mm mm 7-day 28-day GPa MPa x10-6

200AL 0 20 50 370 22.0 27.0 35.5 3.7 575

205AL 50 20 55 315 18.5 26.0 35.5 3.5 455

320AL 0 32 60 350 30.5 41.0 44.5 4.3 605

324AL 40 32 50 365 28.0 37.0 44.5 4.6 525

450BL 0 45 60 370 49.5 55.5 49.0 5.3 615

454BL 40 45 85 290 39.5 52.5 49.0 4.7 525

Table 5 Mechanical, drying shrinkage and properties of the high slump HVFA and portland cement concrete of equivalent 28-day compressive strength

Mix

Desig-

nation

Fly

Ash

Grade

Slump

Flow

Compressive Strength,

(MPa)

28-day

Elastic

Modulus

Creep

Rate

Drying

Shrinkage

at 56 days

% MPa mm mm 7-day 28-day GPa 10-6*1 x10-6

200AH 0 20 115 465 23.0 28.0 36 38.3 600

205AH 50 20 125 425 18.5 26.0 36 16.7 465

320AH 0 32 110 435 34.5 40.0 47 23.9 650

324AH 40 32 100 420 26.5 36.5 42 14.0 605

450BH 0 45 80 345 44.5 54.0 50 - 540

454BH 40 45 130 390 40.5 54.0 53 - 595

Note : 1. Creep rate is given in 10-6 MPa/ln(t+1).

6

DDrryyiinngg SShhrriinnkkaaggee ((mmiiccrroossttrraaiinn))

0 20 40 60 80 100100

200

300

400

500

600

700

800

Number of Days

32 MPa Porland cement concrete

32 MPa HVFA concrete

20 MPa

20 MPaHVFA concrete

Porland cement concrete

Figure 7 Drying Shrinkage of portland cement and high volume fly ash concrete of various

grades with high slump of 100±25 mm.

CCrreeeepp SSttrraaiinn ((mmiiccrroossttrraaiinn ppeerr MMPPaa))

1 2 5 10 20 50 100 200 5000

50

100

150

200

Number of Days + 1

200AH

205AH

320AH

324AH

20 MPa Portlandcement concrete

32 MPa Portlandcement concrete

20 MPaHVFA concrete

32 MPaHVFA concrete

F(K)=38.3

F(K)=23.9

F(K)=16.7

F(K)=14.0

Figure 8 Strain due to creep of portland cement and high volume fly ash concrete of various

grades.

4.3 Durability Properties

The durability of HVFA concretes with respect to the

protection of steel reinforcement against corrosion and the resistance to deterioration in sulphate environments is considered. Three fly ashes, FA1, FA2 and FA3

from three different States in Australia are examined. Most durability studies were carried out using mortars.

It is well known that when the pH of the concrete

surrounding steel reinforcement is sufficiently lowered by carbonation or when there is a sufficient level of chloride ions at the steel surface, steel corrosion

occurs. This could eventually result in cracking of concrete and loss of structural integrity. The service

life of reinforced concrete structure is closely related to properties of the concrete such as its resistance to

carbonation, carbonation-induced steel corrosion, resistance to chloride penetration and chloride-induced

steel corrosion. These properties of HVFA concretes are discussed in this section.

The benefit of the use of fly ash concretes in moderate sulphate environment has been recognised in current

British and Australian Standards 14,15. In this work, the sulphate resistance of fly ash blended cement concretes in a 5% sodium sulphate solution as well as solutions at

pH’s of 7 and 3 are presented.

4.3.1 Corrosion of Steel Reinforcement

Corrosion of steel reinforcement is one of the most

common durability problems in reinforced concrete structures. This problem is caused by either carbonation or chloride penetration or both. Generally

the deterioration of concrete due to corrosion of steel reinforcement is characterised into three stages, ie.

initiation, propagation and accelerated corrosion. In the initiation stage, steel is protected by its passivation in high pH condition and the absence of chloride ions.

The corrosion rate (steel loss) during this stage is very small and is considered negligible for engineering purposes. When the alkalinity of the concrete

surrounding the steel is lowered sufficiently by carbonation and/or when there is sufficient chloride

ions at the steel/concrete interface, steel passivation is destroyed. Corrosion rate of steel becomes significant. This is the propagation stage. The accelerated

corrosion stage occurs when there is severe cracking and damage to the concrete cover caused by the

cumulative effect of the steel corrosion.

a) Carbonation-induced corrosion

Carbonation of concrete is the reaction between

atmospheric carbon dioxide and the cement hydrates (the presence of these products maintains the high

alkalinity in the pore solution). Both calcium hydroxide and calcium silicate hydrates react with carbon dioxide to form carbonate compounds. The result is the

lowering of the pH in the pore solution. When carbonation is advanced the pH of the concrete can be

about 9 or lower. In such a condition, steel passivity is lost. The rate of carbonation or the advance of the carbonation front depends on many factors. Some of

the important factors are time of exposure, the nature of cementitious matrix, its “permeability” to carbon

dioxide and the condition surrounding the concrete (moisture, temperature). The carbonation rate can vary significantly with the climate. For example, the

carbonation front (as detected by phenolphthalein solution) can be three or four times higher in a concrete “sheltered” from rain as compared to a similar concrete

in an “exposed” environment. The important issue with regard to carbonation is that the carbonation problem

should be considered in conjunction with the prevalent

7

climatic and exposure conditions. Furthermore, carbonation is not a problem in itself but carbonation-

induced corrosion of steel reinforcement is. Hence it is necessary to considered the corrosion rate in conjunction with the carbonation rate. For example, a

sheltered concrete may be subjected to a high carbonation rate. However, the corrosion rate will be

low due to the high electrolytic resistance of dry concrete. These considerations must be taken in the analysis of service life.

With a limited length of long-term exposure of two

years in a standard laboratory condition of 23°C 50% RH, a condition which results in a significantly higher

carbonation rate than that expected in an exposed outdoor condition, the depth of carbonation of a range of portland cement and HVFA concretes are compared

on the basis of equivalent 28-day compressive strength (Fc). Figures 9 and 10 show the relationship between carbonation depth and the reciprocal of the square root

of compressive strength 1/sqrt Fc of concretes designed without and with chemical admixture respectively.

For concretes without chemical admixture, the

carbonation depth of HVFA concretes can be higher than that of corresponding portland cement concrete as shown in Figure 9. The differences in the carbonation

depth of the two concretes increase with the reduction in the strength level.

1/sqrt Fc

Carbonation Depths after 2 years (mm)

0.13 0.14 0.15 0.16 0.17 0.18 0.19 0.24

6

8

10

12

14

16Mixes without chemical admixture

R = 95.3%

R = 98.8%

Grade 25

0%FA 30-60%FA

Grade 32Grade 40

Figure 9 Carbonation depth of concretes without

chemical admixture after 2 years exposure in

23°C 50% RH.

Mixes with chemical admixture

R = 95.2%

R = 95.3%

Grade 32 Grade 20

0.12 0.14 0.16 0.18 0.2 0.220

5

10

15

20

1/sqrt Fc

Carbonation Depths after 2 years (mm)

0%FA

40-60%FA

Figure 10 Carbonation depth of concretes with

chemical admixture after 2 years exposure in

23°C 50% RH.

For concretes designed with a standard chemical admixture dosage, that is 400 ml of water reducer or

1000 ml of superplasticizer per 100 kilogram of binder, the differences in the carbonation depth of HVFA and portland cement concretes, as shown in Figure 10, are

not significant especially for concretes in the grade range of 25-32 MPa. According to AS 3600, grades 25

and 32 are recommended for exposure classification A2 and B1. These classifications cover the conditions where carbonation could pose a threat to the durability

of concrete structures.

It is noted that HVFA concretes designed with chemical admixtures are the type of structural concrete which

should be chosen for most building and civil engineering works. Based on the short-term data, these HVFA concretes would be expected to perform as well

as portland cement concretes.

b) Chloride-induced corrosion

Chloride ions can penetrate into concrete through the

effect of concentration gradient and/or through the effect of capillary action. The mechanism of chloride penetration is often described as diffusion. This may

be an over simplification for the process. The transportation of ionic species into a concrete medium

is complicated. This is because of the possible reactions between the chloride ions and the hydrates which constantly alter the pore system. Regardless of

the mechanism of transportation of chloride ions, it is known that when the amount of chloride ions at the

steel/concrete interface is higher than a critical concentration, steel corrosion will occur. This critical chloride concentration is called the chloride threshold

level. It is also known that chloride threshold level depends on binder content and the chemistry of the pore solution.

8

It has been suggested that the hydroxyl concentration is the controlling factor with regard to the chloride

threshold level. A relationship, such as Cl-/OH- = 0.6, has been suggested for the estimation of chloride threshold level. Recent work by Cao et al.16 indicated

that this is not the case since blended cements can have similar chloride threshold level to portland cements

despite having lower OH- concentrations in their pore solution.

When passivity of steel cannot be maintained, corrosion of steel occurs. The service life of a

reinforced concrete structure is directly related to the development of the corrosion and its rate. However, it

must be stressed that the mode of corrosion should also be considered. For example, metal loss due to pitting corrosion may be much smaller than that of general

corrosion. However, the effect of pitting corrosion (concentrated metal loss in a small area) can be very

dangerous to the integrity of the structure in terms of loss in load carrying capacity.

The detection of corrosion and the measurement of corrosion rate of steel can be used to compare the

behaviour of different binders in the initiation and propagation stages of the deterioration. One of the

possible methods of comparatively assessing the initiation stage is by monitoring the corrosion rate of steel embedded in concrete or mortar. The change in

corrosion rate from ‘negligible’ to ‘significant’ can be used to determine the effect of binder type on initiation period. These data are presented in Figure 11. The

corrosion rate was determined by using polarisation resistance technique. It must be stressed that the

initiation stage is controlled by :

• the rate of chloride penetration,

• the chloride threshold level and

• the concrete cover thickness.

The effect of 40 wt.% fly ash binder systems on the

development of corrosion of steel is shown in Figure 11. From this figure, it can be seen that the use

HVFA binders leads to a similar or longer initiation period as compared to portland cement mortar of the same W/B and with limited initial curing period of 7

days. The cover of the mortars over steel sample in this case is about 7 mm. It can be seen that the initiation

period where the corrosion rate of steel is negligible for this configuration is about 3 months for portland cement and about 3 to 4 months for both HVFA

binders. It is expected that for a realistic concrete cover in a marine environment, say about 50 mm, the effect of HVFA binder in terms of the increase of the

initiation period will be further magnified. This conclusion is based on better chloride penetration

resistance performance data at larger cover depths17,18and the lower W/B ratio used to produce

HVFA concrete of the same strength grade as portland cement concrete (described in 3.1.1).

0 2 4 6 8 100

0.05

0.1

0.15

0.2

0.25

Time of immersion in 3% NaCl (months)

Corrosion rate (micro A/cm2)

Type A/1

40% FA1

40% FA2

W/B = 0.67 d moist cured

Figure 11 Effect of binder on the corrosion rate of steel embedded in 7 mm thick mortars.

The propagation period, characterised by significant corrosion rate, is also a very important to the maintenance-free service life. The main reason is that,

for a binder system in which a low corrosion rate can be maintained, the service life will be prolonged by an

extended propagation period.

The effect of fly ash on the corrosion rate of steel

embedded in mortars W/B ration of 0.4 and cured for 7 days, is shown in Figure 12. It is clear that the use of high volume fly ash binder systems can result in

reduced corrosion rate of steel in comparison to the portland cements. The extent of the reduction depends

on the source of the fly ash.

Type A/1 40%FA1 40%FA2 40%FA3 Type A/20

0.05

0.1

0.15

0.2

0.25

0.3

0.35

Corrosion rate (micro A/cm²)

W/B = 0.47 d moist cured1 year immersion

Figure 12 Effect of binder on the corrosion rate of

steel embedded in 7 mm thick mortars after one year of immersion in 3 % NaCl.

It should be noted that the corrosion rate mentioned

above is that of microcell corrosion rate where the anode and cathode of the corrosion cell are in close proximity and sometimes are not physically

distinguishable. For most reinforced concrete applications, there are situations where the anode and

cathode of the corrosion cell can be physically

9

separated. In such a situation, termed macrocell corrosion, the characteristics of the concrete, such as

the resistance to ionic transportation and its resistivity, will have important influence on the corrosion rate of steel. By using fly ash blended cement, both of these

characteristics of the concrete will be improved and hence the corrosion rate will be reduced. This has been

confirmed experimentally as shown in Figure 13 in which the macrocell corrosion rates were determined using a model of equal areas of anode (chloride

contaminated area) and cathode (chloride free area) and the mortar medium was 15 mm. It can be seen that the

beneficial effect of HVFA binder systems in reducing the macrocell corrosion rate compared to portland cement mortar of the same W/B is very significant.

The effect of increasing the fly ash proportion on the reduction of macrocell corrosion rate is also clearly

evident.

Type A/1 20% FA1 40% FA1 60% FA10

0.5

1

1.5

2

2.5

Macrocell corrosion rate (micro A/cm2)

W/B = 0.87 d moist cured6 months immersion

Figure 13 Effect of varying dosage of fly ash on the

Macrocell corrosion rate of steel after six months of immersion in 3 % NaCl.

The overall conclusion is that the use of HVFA

concrete can result in extended maintenance-free service life of reinforced concrete structure in marine environments. This is based on its potential in

increasing the initiation period and reducing the corrosion rate in the propagation period of the

deterioration process due to chloride-induced corrosion of steel reinforcement.

4.3.2 Sulphate Resistance

Sulphate resistance of a cementitious material can be

broadly defined as a combination of its physical resistance to the penetration of sulphate ions from

external sources and the resistance of the chemical reactivity of its components in the matrix to the sulphate ions. Both factors are important to the overall

resistance to sulphate attack of a concrete structure. However, the chemical resistance to sulphate attack is

considered to be more critical for long term performance. The physical resistance can be improved by good concreting practice such as the use of concrete

with low W/B, adequate compaction and extended curing. The extent of chemical reactivity of hardened

cement paste with sulphate ions, on the other hand, depends very much on the characteristic of the binder system. It may appear obvious that for a long service

life, a good concrete is required. However, this will be better assured if the concrete is made from a binder that

has low “reactivity” with sulphate ions.

Sulphate resistance of cementitious materials can be assessed by a variety of methods. The performance of two fly ashes and three portland cements was examined

in terms of expansion characteristic and strength

development of mortars immersed in 5% Na2SO4

solution. In addition, the performance of mortars in 5% Na2SO4 solution at lower pH’s of 3 & 7 were determined.

All the expansion mortars were made with a fixed sand-

to-binder ratio of 2.75 and variable amount of water to

give a similar flow of 110±5 %. In most cases, the fly ash mortars had lower W/B than portland cement

mortars. When a test was performed to ASTM C1012,

the samples were cured for 1 day at 35°C and

subsequently at 23°C until they reached a compressive

strength of 20 MPa before immersion in the Na2SO4 solutions. Mortars used in the evaluation of

compressive strength retention had the same sand-to-binder ratio of 2.75 and a W/B ratio of 0.6.

Apart from two fly ashes, FA1 and FA2, the three portland cements used were Type A, C and D cement

(normal, low heat and sulphate-resisting cement respectively) according the superseded AS 1315-1982.

0 10 20 30 40 500

2,000

4,000

6,000

8,000

Immersion period in 5% Na2SO4 (wks)

Mortar Expansion (microstrain)

Type A/1

Type C

Type D

40%FA1

ASTM C1012

Figure 14 Expansion of mortar bars in 5% Na2SO4 solution to ASTM C1012.

The effect of fly ash blended cement on expansion of

mortar using the ASTM C1012 procedures is shown in Figure 14. A 5% Na2SO4 solution, without any control on the pH of the solution, was used in this case.

10

It can be seen that the use of 40 wt.% fly ash blended cement greatly reduces the expansion of mortar. In fact

the expansion of fly ash blended cement is much lower than those shown by the three portland cements Type A, C and D. When all the mortars were cured for a

short period of 3 days, the effect of both fly ashes on the reduction of expansion was also very clear as

shown in Figure 15.

0 10 20 30 40 500

1,000

2,000

3,000

4,000

5,000

6,000

7,000

Immersion Period in 5%Na2SO4 (wks)

Expansion (microstrain)

Type A/1

Type C

40%FA1

40%FA2

3 day curing

Figure 15 Expansion of mortar bars in 5% Na2SO4 solution to ASTM C1012 but with 3 days

moist curing.

The improved expansion characteristic of fly ash blended cement mortars was maintained in sulphate environments of low pH’s as shown in Figures 16 and

17. In fact, the same effect has been observed for most Australian fly ashes when used at a replacement level

of about 40 percent19.

0 10 20 30 40 500

1,000

2,000

3,000

4,000

5,000

Immersion period in pH 7 Na2SO4 (wks)


Type A/1

Type C

40%FA1

40%FA2

ASTM C1012pH 7

Figure 16 Expansion of mortar bars in a 5% Na2SO4

solution (low pH 7) to ASTM C1012.

0 10 20 30 40 500

1,000

2,000

3,000

4,000

5,000

Immersion period in pH3 Na2SO4 (wks)


Type A/1

Type C

40%FA1

40%FA2

ASTM C1012pH 3

Figure 17 Expansion of mortar bars in a 5% Na2SO4 solution (low pH 3) to ASTM C1012.

Apart from a much lower expansion characteristic in sulphate environments, the use of binders with “high” fly ash percentage leads to superior strength retention

in a sulphate solution. Figure 18 shows that for most portland cements, the loss of compressive strength was observed after about 6 months in sulphate solution.

Whereas, the 40 wt.% fly ash blended cement mortars showed strength increases even after 1 year.

0 100 200 300 40010

20

30

40

50

60

Time of Immersion (days)

Compressive Strength (MPa)

Type A/1 40%FA1 40%FA2 TypeA/2 TypeC

W/B = 0.67 d moist curedNo pH control

Figure 18 Compressive strength of mortars after different periods of immersion in sulphate solution.

In low pH sulphate solutions, the beneficial effect of high replacement fly ash binder was even more pronounced as shown in Figures 19 and 20.

The results clearly indicate that for concrete application in sulphate environment, particularly in those where the pH is low, the use of HVFA concrete has the highest

probability of extending the service life.

11

0 100 200 300 40010

20

30

40

50

60

Time of immersion (days)

Compressive Strength (days)

Type A/1 40%FA1 40%FA2 TypeA/2 TypeC

W/B = 0.67 d moist curedpH 7

Figure 19 Compressive strength of mortars after different periods of immersion in a pH 7 sulphate solution.

0 100 200 300 40010

20

30

40

50

Time of Immersion (days)

Compressive Strength (MPa)

Type A/1 40%FA1 Type C

W/B = 0.67 d moist curedpH 3

Figure 20 Compressive strength of mortars after different periods of immersion in a pH 3

sulphate solution.

4.3.3 Optimum Dosage for Durability

In marine and sulphate environments, the optimum dosage of fly ash has not been strictly determined but is

expected to be around 40 wt.%. When the proportion of fly ash by weight exceed about 50 %, the lowering in

compressive strength becomes significant. With the present cost structure of concreting materials including fly ash, a HVFA concrete with 40 wt.% fly ash is

marginally more expensive than portland cement concrete of equivalent strength grade. However, with the significantly increased expected service life, it is

argued that the optimum dosage with respect to service life of HVFA concrete would be around 40% by weight

of binder.

5. APPLICATIONS

Sufficient knowledge is now available on the design, production and properties of HVFA concrete for its

applications to be identified.

5.1 Suitability based on Inherent Properties

The inherent properties and limitations of HVFA concrete are the key to its selection in suitable

applications. They are given in relative terms to the properties of portland cement and conventional fly ash

concrete as follows:

• good cohesiveness or sticky in mixes with very high

binder content;

• some delay in setting times depending on the compatibility of cement, fly ash and chemical

admixture;

• slightly lower but sufficient early strength for most applications;

• comparable flexural strength and elastic modulus;

• better drying shrinkage and significantly lower creep;

• good protection to steel reinforcement in high

chloride environment;

• excellent durability in aggressive sulphate environments; and

• lower heat characteristics.

5.2 Built Structures

High Volume Fly Ash concretes have already found applications in major structures in many countries.

Langley (1988)20 reported its use in the Park Lane and Purdys Wharf Development in Halifax, Nova Scotia,

Canada. It is also believed that a 40-50% wt. fly ash concrete was used in the construction of the caissons of the famous Thames River Flood Barrier in London and

in bridge foundations in Florida by the Florida Department of Transportation (FDOT)21.

In pavements, Nelson et al. (1992)6 reported the

pioneer application of concrete with 40% wt. fly ash in the construction of sections of road pavement and an apron slab at Mount Piper Power Station in New South

Wales, Australia. The casting of the apron slab is shown in Figure 21. At about the same time,

Naik et al. (1992)22 reported the successful use of three fly ash concrete mixtures, 20% and 50% ASTM C618 Class C fly ash and 40% ASTM C618 Class F fly ash

to pave a 1.28 km long roadway in Wisconsin, USA.

The largest volume of HVFA concrete used in Australia was in the construction of the basement slabs

and walls of Melbourne Casino in 1995. Figure 22 shows concreting activities on the site. According to Grayson

23 , of Connell Wagner, low drying shrinkage

and durable concrete was required for the construction of the 55,000 m2 basement which was located below the water table. Saline water was found on the site

situated near the Yarra River. Slabs with an average thickness of 400 mm were designed to withstand an

uplift pressure of 45 kPa. The concrete was specified to contain at least 30 kg/m3 of silica fume or 30 wt.% of

12

fly ash or 60 wt.% of a combination of slag and fly ash. Drying shrinkage within 650 microstrains was also

specified. Liversidge 24, of Grollo Contractor, selected

a 40 MPa HVFA concrete containing 40 wt.% fly ash for the 40,000 m3 of concrete required for the

basement. The fresh concrete was reported to behave similarly to conventional concrete and a drying

shrinkage of lesser than 500 microstrains was achieved. In addition, a similar concrete was used in the construction of the pile caps and two raft slabs in the

same project.

Figure 21 First pour of 32 MPa HVFA concrete in an apron slab at Mount Piper Power

Station in New South Wales in 1991.

Figure 22 Concreting activities at Melbourne Crown Casino in 1995.

In Perth, Western Australia, a 50 wt.% fly ash concrete

was reported25 to have been used for the construction of the secant piles at Roe Street Tunnel. The ground water in the area was tested and found to be abnormally

acidic, pH = 4.0. Thus it was necessary for all piles to contain a high binder content to limit the attack of the

ground water on the concrete. A minimum binder content of 350 kg/m3 and W/B = 0.5 was specified. The requirements posed problems for the low early-age

strength needed to allow the soft piles to be bored. A number of trial mixtures were cast and the preferred option for the binder was a 50:50 portland cement fly

ash blend.

5.3 Current Developments

While the potential applications of HVFA concrete are numerous, three recent developments are worth noting.

The first is in the use of fibre-reinforced HVFA shotcrete to cap degraded rock outcrops and to cover

mine waste dumps, the second is in High Performance Concrete for massive marine structure, and the third is in the upgrading of dam structures.

Morgan et al. (1992)26 found polypropylene fibre-

reinforced HVFA shotcrete to be applied satisfactorily using conventional wet-mix shotcrete equipment and

that it required a minimum amount of cementitious material and water content of around 420 and 150 kg/m3 respectively. The polypropylene fibre content

required to provide a satisfactory flexural toughness index appeared to be between 4 and 6 kg/m3. They suggested its use in capping rock outcrops which are

susceptible to degradation and for covering mine waste dumps. Seabrook (1992)27 identified the same

technology to produce a lower cost shotcrete while maintaining reasonable quality and durability for

application on waste piles to prevent leaching of acids and heavy metals.

In marine and offshore structures such as bridges,

wharfs, sea walls and offshore concrete gravity structures, concretes with excellent sulphate and

chloride durability are required. Where large or long-span structural members are used, low heat development and low creep characteristics are of vital

importance. These are applications where HVFA concretes would be considered an ideal solution. Attention would be given to the use of compatible

concreting materials if high early strength is required in precasting or slip forming construction.

In rehabilitation of massive concrete structures such as

the raising of dam height for improved safety and flood mitigation, new high strength and high elastic modulus concrete matching existing concrete is often required.

The new concrete must have low shrinkage to minimise the effects of the new concrete on existing concrete and

low heat development characteristics to avoid potential cracking problems during construction. HVFA concretes have been found to have all the required

attributes for such use.

6. SUMMARY

In these guidelines, procedures and data on the design of structural grades for High Volume Fly Ash (HVFA) concretes are given. The range of engineering properties including : consistency and

setting times of fresh concrete, mechanical, volume stability and durability properties of hardened concrete are discussed. These are vital to satisfy the structural and serviceability requirements of

concrete structures. Experiences gained from a major field trial in New South Wales and a large scale implementation in a prestigious building project in Victoria , confirmed the practicality of the

this new concrete in the Australian construction industry. Significant improved volume stability, in terms of reduced drying shrinkage and better creep characteristics, can result in new solutions to many engineering problems. Most important of all, the

improved durability performance of HVFA concrete in marine and high sulphate environments signalled the tremendous economic gains which could be derived from the expected increased service life.

Exciting new developments in the applications of HVFA concrete in Australia and overseas have been

highlighted. It remains for the construction industry to adopt and advance this new technology to its full potential in Australia.

7. ACKNOWLEDGMENT

13

The authors acknowledge the support by Pacific Power and the Ash Development Association of Australia

(ADAA) and their permission to publish the data. In particular, we wish to express our sincere gratitude to Mr Peter Nelson, formerly of Pacific Power and

presently the Administrator of ADAA, for his strong belief and constant support in the development of

HVFA concrete in Australia.

8. REFERENCES

1. Ryan, W.G., ‘Review of the Australian Experience’, Proceedings International Workshop on the use of Fly Ash, Slag, Silica Fume and other Siliceous

Materials in Concrete’, Sydney, Australia, July 1988.

2. Butler, W. B., ‘Economic Binder Proportioning

with Cement-Replacement Materials’, Cement, Concrete and Aggregates. CCAGDP. Vol. 10, No.

1, Summer 1988, pp. 45-47. 3. Proctor, R.T. and Lacey, R.A.C., ‘The Development

of High Fly Ash Content Concrete at Digcot Power

Station’, Ashtech’84, Central Electricity Generating Board, London, September 1984, pp. 461-467.

4. Sirivivatnanon, V., Cao, H.T., Baweja, D. and Nelson, P., ‘Properties of High Volume Fly Ash Concrete’, Concrete Institute of Australia 16th

Biennial Conference, Melbourne, Australia, May 1993, pp 256-273.

5. Sirivivatnanon, V., Cao, H.T. and Nelson, P. ‘Choosing the Right Cement-Fly Ash Combinations in High Volume Fly Ash Concrete’, Proceedings,

International Conference on the Concrete Future, Malaysia, 1992, pp. 211-218.

6. Nelson, P., Sirivivatnanon, V. and Khatri, R.,

‘Development of High Volume Fly Ash Concrete for Pavements’, Part 2 - Proceedings 16th ARRB

Conference, Perth, Australia, November 1992, pp. 37-47.

7. American Concrete Institute, ‘ACI Standard

Practice for Selecting Proportions for Normal, Heavyweight and Mass Concrete’, ACI 211.1-89,

ACI Manual of Concrete Practice, Part 1. 8. Teychenne, D.C., Franklin, R.E. and Erntroy, H.C.,

‘Design of normal concrete mixes’, London,

HMSO, 1975. 9. Road Research, ‘Design of concrete mixes’,

D.S.I.R. Road Note No. 4 (London, H.M.S.O.,

1950). 10. Sirivivatnanon, V., Cao, H.T. and Nelson, P.,

‘Choosing the right cement - fly ash combinations’, Proceedings, International Conference on the Concrete Future, Malaysia, 1992, pp. 211-218.

11. Cao, H.T., Bucea, L., Wortley, B.A. and Mechanikov, E.. “Influence of fly ash on chloride

induced corrosion of steel reinforcement”, CSIRO

Report BRE 022 to the Ash Development Association of Australia, April 1994.

12. Cao, H.T., Bucea, L., Yozghatlian, S., Wortley,

B.A. and Farr, M., ‘Influence of fly ash on Sulphate Resistance of Blended Cements’, CSIRO Report

BRE 023 to the Ash Development Association of Australia, June 1994.

13. Standards Association of Australia, ‘Concrete

Structures AS 3600-1988’, Sydney, Australia 14. British Standards Institution, ‘Structure use of

Concrete’, BS 8110: Part 1: 1985. 15. Standard Association of Australia, ‘Rules for the

design and installation of piling’, AS 2159-1978,

Sydney, Australia. 16. Cao, H. T., Bucea, L., Mcphee, D.E. and Christie,

E.A., ‘Corrosion of Steel Reinforcement in Concrete - Part 1 : Corrosion of Steel in Solutions and in Cement Pastes’, CSIRO Report BRE 009,

DBCE, February 1992. 17. Thomas, M.D.A., ‘Marine performance of PFA

concrete’, Magazine of Concrete Research, 43, No.

156, Sept. 1991, pp 171-186. 18. Sirivivatnanon, V. and Khatri, R., ‘Munmorah

Outfall Canal’, CSIRO Report BIN 068, DBCE, June 1995.

19. Cao, H. T., Bucea, L., Yozghatlian, B.A., Wortley,

B. A. and Farr, M., ‘Influence of Fly Ash on the Sulphate Resistance of Blended Cements’, CSIRO

Confidential Report BRE 023, DBCE, June 1994. 20. Langley, W. S., ‘Structural Concrete Utilising High

Volumes of Low Calcium Fly Ash’, Proceedings

Internation Workshop on the use of Fly Ash, Slag, Silica Fume and other Siliceous Materials in Concrete, edited by W.G. Ryan, Sydney, 1988, pp

105-130. 21. Armaghani, J.M., , Florida Department of

Transportation, private communication. 22. Naik, T.R., Ramme, B.W. and Tews, J.H.,

‘Pavement Construction with High Volume Class C

and Class F Fly Ash Concrete’, Proceedings CANMET/ACI International Symposium on

Advances in Concrete Technology, Sep-Oct, 1992. 23. Grayson, R., Principal in charge of Melbourne

Casino Project - Connell Wagner, private

communication. 24. Liversidge, P., Technical Manager - Grollo

Premixed Pty Ltd, private communication. 25. Ryan, W.G. and Potter, R.J., ‘Application of High-

Performance Concrete in Australia’, Supplementary

Papers, ACI International Conference on High-Performance Concrete, Singapore, 1994.

26. Morgan, D.R., McAskill, N., Carette, G.G. and

Malhotra, V.M., ‘Evaluation of polypropylene fibre-reinforced high-volume fly ash shotcrete’,

14

Proceedings International Workshop on Fly Ash in

Concrete, October 1990, Calgary, Alberta, Canada. 27. Seabrook, P T., “Shotcrete as an Economical

Coating for Waste Piles”, Proceedings

CANMET/ACI International Symposium on Advances in Concrete Technology, Sep-Oct, 1992.

1

8th Austroads Bridge Conference 31 October - 5 November 2011

TESTING & SPECIFYING CHLORIDE RESISTANCE OF CONCRETE

Vute Sirivivatnanon1 and Radhe Khatri

2

1Research Manager – Cement Concrete & Aggregates Australia

2Durability Engineer – Bridge Engineering, Technology and Practice, RTA New South Wales

Abstract

A wide range of test methods has been used to measure the resistance of concrete to chloride

penetration. Each method is purported to measure the property of concrete central to the chloride

transport mechanism into concrete. They include water absorption, water permeability, chloride

diffusion, chloride migration and electrochemical resistance. In this paper, an assessment of five

commonly specified properties: namely the water-cement ratio, 28-day diffusion coefficient,

sorptivity, volume of permeable voids, and ASTM C1202 rapid chloride permeability; was made in

terms of its relation to the long-term chloride diffusion coefficient taking into consideration the

precision of the test methods. The long-term chloride diffusion coefficient is derived from the

chloride penetration profile into concrete after one-year immersion in sodium chloride solution. The

results cover a range of concretes commonly used in Australia.

2


1. Designing for Durability

Engineers design structures to satisfy both the functional and sustainable requirements. For reinforced concrete structures, the choice of concrete is particularly important for long-term performance of structures built in aggressive environments. Structures in coastal and marine environments are required to be corrosion resistant as chlorides, abundant in these environments, are particularly corrosive to structural steel and steel reinforcement. In reinforced concrete structures, the concrete cover provides the necessary protection against the penetrating chlorides. Structural design and detailing limit and control the distribution and width of cracking to limit access of chlorides to the steel reinforcement.

The choice of concrete should be based on concrete with suitable fresh and hardened concrete properties: fresh concrete properties suitable for the geometry of structural members and the construction method used; hardened concrete properties and structural design which minimize crack width and provide good protection to the steel reinforcement throughout its service life.

In designing for durability against chloride-induced corrosion, the quality of concrete will affect both the initiation and propagation periods of the service life. Cao and Sirivivatnanon, 2001 presented a comprehensive service life model based on a chloride diffusion transport mechanism. The model takes into account the diffusivity decay, temperature and stress condition on the chloride diffusion coefficient. There has been considerable research bridging the knowledge gaps in terms of temperature effect Tang (1996), and the effect of crack width on corrosion (Lim, 2000 and Jang et al 2011).

Khatri and Sirivivatnanon (2004) applied a simple statistical approach in defining a characteristic service life concept similar to the familiar characteristic strength. The FIB Model Code for service life design (2006) was based on probabilistic service life design with various approaches to define the end of service life. Sirivivatnanon et al (2007) applied the concept of limit-state service-life design to better define the end of service life. The concept is currently being researched and is probably best the approach to service-life design by design engineers who apply limit-state to their structural design.

2. Chloride transport mechanisms

The transportation of chloride ions into concrete is a complicated process which involves diffusion, capillary suction, permeation and convective flow through the pore system and microcracking network, accompanied by physical adsorption and chemical binding (Kropp and Hilsdorf, 1995). With such a complex transportation process, it is necessary to understand individual transport mechanisms and the predominant transport process in order to pinpoint the appropriate method for quantifying the chloride resistance of concrete. Individual chloride transport mechanisms and the associated test methods have been described in CCAA Report on Chloride Resistance of Concrete (2009).

There is usually more than one transport mechanism involved. Most performance-based specifications are based on the predominant transport mode applicable to the specific exposure condition. Table 1 summarises the chloride transport mechanisms applicable to structures in various exposure conditions. It is generally accepted that the predominant chloride transport mechanism is that of diffusion. The strength of various test methods used in measuring the chloride resistance of concrete can therefore be assessed by its correlation to diffusion, and in particular, to the long-term diffusion coefficient.

Diffusion is a transfer of mass of free molecules or ions in the pore solution resulting in a net flow from regions of higher concentration to regions of lower concentration of the diffusing substance. This mode of transport operates in fully saturated media such as fully submerged concrete structures. For porous material like concrete, the diffusion coefficient, D, is the material characteristic property describing the transfer of a given substance driven by concentration gradient.

3


In steady-state chloride diffusion, the effective driving force is the gradient of the free chloride ions in the pore solution; the diffusion coefficient is referred as Dfree or Df. Df can be determined from the difference in concentration of chloride ions in the two cells separated by the concrete. This type of test is used only in research as it is extremely time consuming.

Immobilization of diffusing material is of considerable importance for the experimental determination of diffusion coefficients. As long as the binding capacity of the test specimen is not yet exhausted, the net flow of diffusing material appears to be low. The ‘apparent’ diffusion coefficient, Dapparent or Da, may be deducted from the acid-soluble chloride profile from the experiments.

In a non-steady-state diffusion process, the gradient of the free chloride ions in the pore solution is the effective driving force; the diffusion coefficient derived from measured free chloride concentration profiles is the effective diffusion coefficient, Deffective or De. The effective diffusion coefficients can also be derived from measured concentration profiles in existing structures and serve as a basis for estimates on the future ingress of the chlorides.

Table 1 Chloride transport mechanisms into concrete under various exposures

Exposure Example of Structures Chloride transport mechanisms

Submerged Substructures below low tide. Diffusion.

Basement exterior walls or transport tunnel liners below low tide. Liquid containing structures.

Permeation, diffusion and possibly Wick’s action.

Tidal Substructures and superstructures in the tidal zone. Capillary absorption and diffusion.

Splash and spray Substructures and superstructures above high tide in the open sea.

Capillary absorption and diffusion.

Coastal Land based structures in coastal area or superstructures above high tide in river estuary or body of water in coastal area.

Capillary absorption.

3. Testing and specifications

A range of concrete properties have been used as a measure of its resistance to chloride penetration. Each will be described and discussed in terms of its relevance to chloride transport mechanism, its adoption in specifications and testing precision, and its relationship to measured long-term chloride profile quantified as effective or apparent chloride diffusion coefficients. The long-term chloride profile is considered the best measure of the mixed modes of chloride transportation and in particular the chloride level at cover depth.

Two comprehensive sets of research data along with other specific experimental outcomes are used to relate each concrete property to the long-term chloride diffusion coefficients. CSIRO (1998) examines a range of properties of concretes proportioned from one Type GP and three Type GB cements. The Type GB cements include cement with either a 30% fly ash, a 65% ground granulated blast-furnace slag or a 10% silica fume content. All the concretes were membrane-cured for 7 days and tested at 28 days, such as the mean compressive strength. A set of specimens were immersed in 3% sodium chloride solution and the effective chloride diffusions, De.28 and De.365, were obtained after a 28-day and one-year immersion period respectively.

Sherman et al (1996) also investigated a range of concrete properties and corresponding apparent chloride diffusion coefficients, Da.365, after 365-day ponding conducted using AASHTO T259. Fifteen mixtures of concrete were proportioned from ASTM Type I cement with and without silica fume (at 5% and 7.5%) at w/c of 0.30, 0.40 and 0.46. The concretes were cured in a number of ways including water cured, wet burlap covered and heat cured.

4


3.1 Cement Type and Water-Cement Ratio

The porosity of concrete is highly sensitive to water-cement ratio. The connectivity of the pore system depends on the amount of original mixing–water filled space and the degree to which it has been filled with hydration products. Capillary pores are those voids remaining that were originally filled with mixing water. Water-cement ratio has therefore been used to specify durability. However, various types of cements have been found to affect chloride resistance of concrete differently. The cement chemical composition affects both the porosity and chloride binding capacity of cement. The effective chloride diffusion coefficient, De, of concrete incorporating a Type GP and three Type GB cements, derived from concrete after one year immersion in 3% sodium chloride solution, are shown in Figure 1. They show great divergence ranging from 5 to 30 x 10

-12 m

2/s at high w/c of 0.6

to a narrow range of 1 to 4 x 10-12

m2/s at low w/c of 0.4. Similar findings are shown for concrete

made from an ASTM Type I cement and a blend of ASTM Type I and silica fume. For high chloride

resistant concrete, both the type of cement and maximum water-cement ratio must be specified.

The Roads and Traffic Authority of New South Wales (RTA, 2008) has used such prescriptive requirements in the B 80 specifications. These are shown in Table 2.

Table 2 Wet Curing and other requirements in RTA specifications

Curing period requirement (days) Other requirements Exposure

classification SL cement

Blended cement containing

BFS* and/or FA**

Blended cement

containing AS***

Minimum cement content (kg/m

3)

Maximum w/c

(by mass)

Min strength for durability,

fc.min(d)

(MPa)

B2 7 14 7 370 0.46 40

C N/A 14 7 420 0.40 50

*blast furnace slag, ** fly ash, *** amorphous silica.

0

10

20

30

40

50

60

0.3 0.4 0.5 0.6 0.7

Water/Cement

Dif

fus

ion

Co

eff

icie

nt,

De.3

65 (1

0-1

2 m

2/s

)

GP

GB1

GB2

GB3

y = 0.0068e15.467x

R2 = 0.9603 or R=98%

y = 0.0475e7.7809x

R2 = 0.9326 or R=97%

0

1

2

3

4

5

6

0.2 0.3 0.4 0.5

Water/Cement

Dif

fus

ion

Co

eff

icie

nt,

D365 (

10

-12m

2/s

)

Type I

Type I+SF

Figure 1 Effect of water-cement ratio and cement type on the chloride resistance of concrete (CSIRO, 1998) (Sherman et al 1996)

5


3.2 Chloride Diffusion Coefficients

The diffusion coefficient can be determined from a non-steady-state or a steady-state of chloride test using Fick’s second and Fick’s first law respectively. Due to time constraints and specimen configuration, it is more common for the chloride diffusion coefficient of concrete to be determined from a non-steady-state test.

The curing period and age of the concrete at the commencement of exposure to chloride also influences the resultant diffusion coefficient but to a lesser extent than exposure period as shown in Table 3. In this case, the diffusion coefficients are of the same order. In durability specifications, both the age of concrete and exposure period must be clearly specified.

Table 3 Influence of curing and age of concrete on Da (CSIRO,1993)

Apparent Diffusion Coefficient after 28-day exposure, Da.28 (10

-12 m

2/s)

Curing period prior to exposure (days)

Concrete 1 Concrete 2 Concrete 3

28 3.17 3.50 3.17

56 2.25 1.21 1.22

Diffusion coefficients from a short-term exposure period (say 28 days, De.28) are the more practical specification used for major infrastructure projects when there is sufficient lead time for the appropriate concrete mix design to be developed and tested prior to commencement of construction. Such diffusion coefficients are a reasonably good indicator of the longer-term chloride resistance, De.365, as shown in Figure 2 for a set of one Type GP and three Type GB cement concretes. However, the correlation coefficient, R, of 74% indicates that short-term diffusion coefficients are not necessarily the best indicator of longer-term chloride resistance. The results also show a reduction of diffusion coefficients with exposure period, with the value of De.28 being about a third of De.365.

According to ASTM C1566, the repeatability and reproducibility coefficient of variation of the apparent diffusion coefficient are 14.2% and 20.2% respectively. Hence the results of two properly conducted tests by the same operator should not differ by more than 39.8%. The results of two properly conducted tests in different laboratories should not differ by more than 56.6%.

y = 0.3248x - 0.47

R2 = 0.554 or 74%

0

10

20

30

10 20 30 40 50 60

Diffusion Coefficient, De.28 (10-12

m2/s)

Dif

fus

ion

Co

eff

icie

nt,

De.3

65 (

10

-12 m

2/s

)

GP & GB

Figure 2 Short-term exposure diffusion coefficient De.28 as indicator for longer-term diffusion coefficient, De.365 (CSIRO, 1998)

6


3.3 Sorptivity and AVPV

A number of absorption-related properties are used to indicate the porosity and durability potential of concrete. Absorption, sorptivity and apparent volume of permeable voids (AVPV) are each fundamentally a measure of capillary absorption or absorption rate of concrete.

Sorptivity is a measure of one-dimensional capillary absorption rate as a function of time and provides a good indication of the pore structure and its connectivity of the near-surface concrete. There are two methods of measuring sorptivity. The first (RTA T362) measures the depth of penetrating water front into the concrete and the second (ASTM C1585-04) measures the depth of water penetration indirectly through the weight gain in the concrete. The sorptivity is calculated from the slope of the curve of the depth of water penetration with time in mm/hour

½ or the weight gain

against time and converted into mm/min½ or mm/s

½ for the respective methods.

The Roads and Traffic Authority of New South Wales (RTA) has used sorptivity as an alternative performance-based specification (Provision A) in RTA B 80 specifications. The RTA sorptivity for exposure classes B2 and C as shown in Table 4 is measured over a soaking period of 24 hours, hence RTA sorptivity of 1 mm is equivalent to 0.026 mm/min

½, and RTA

sorptivity of 3.8 mm is equivalent to 0.1 mm/min½.

Table 4 Durability Requirements for Concrete (Provision A) in RTA B80

Exposure classification

Minimum cement content

Maximum water/cement

Maximum sorptivity penetration depth (mm)

(kg/m3

) (by mass) Portland cement Blended cement

B2 370 0.46 17 20

C 420 0.40 N/A 8

y = 0.2189x - 2.8237

R2 = 0.5775 or 76%

0

10

20

30

0 10 20 30 40 50 60 70 80 90 100 110

RTA Sorptivity (mm)

Dif

fus

ion

Co

eff

icie

nt,

De

.36

5 (

10

-12 m

2/s

)

Type GP & GB

Figure 3 Influence of RTA sorptivity on longer-term effective chloride diffusion (CSIRO, 1998)

Sorptivity has been found to be a reasonable performance indicator of chloride resistance of concrete. In Figure 3, the relationship between sorptivity and longer-term effective chloride diffusion coefficients of a range of concretes (one Type GP and three Type GB) is shown to have reasonably good correlation (R=76%). Lower sorptivity indicates better chloride resistance.

For Class C exposure, a maximum sorptivity limit of 8 mm or 0.21 mm/min½ is specified for blended

cement. It can be observed from Figure 3 that such low sorptivity would indicate a concrete with a very low diffusion coefficient. Such low sorptivity has been found to be difficult to measure and no precision statement is available for the RTA sorptivity.

7


The sorptivity of matured concrete (tested at 19 years old) has also been found to give a reasonable indication of the long-term chloride resistance of a range of concretes exposed in the splash zone in Port Fremantle in Western Australia (Vallini and Aldred, 2003). The concrete mixes were proportioned from a range of cements including Type GP cement, Type GB cement with 30–65% GGBS and Type GB cement with 10% silica fume, all at nominal water-cement ratio of 0.4. A good correlation coefficient R of 85% is found as shown in Figure 4. The sorptivity is measured using a method similar to ASTM C1585.

According to ASTM C1585, the repeatability coefficient of variation is 6.0%. Hence the results of two properly conducted tests by the same operator should not differ by more than 17.0%.

y = 525.91x3.1909

R2 = 0.7287 or R=85%

0

0.2

0.4

0.6

0.8

1

0.04 0.06 0.08 0.1 0.12 0.14

Sorptivity of Mature Concrete (mm/min1/2)

Dif

fus

ion

Co

ef.

, D

a.1

9 y

r (1

0-1

2 m

2/s

) Beams at Port Fremantle

Figure 4 Sorptivity of 19-year-old concretes and apparent chloride diffusion coefficient, Da, after 19-year exposure in splash zone in Port Fremantle (Vallini and Aldred, 2003)

It should be noted that the non-steady-state apparent diffusion coefficient, Da.19yr, is about an order of magnitude below the non-steady-state effective diffusion coefficient, De.365, and the sorptivity of the matured concrete is correspondently an order of magnitude below the sorptivity of the 28-day-old concrete. These are good examples of the reduction of diffusion coefficient with exposure time, and the limitation of specifying diffusion coefficient from short-term exposure tests.

Volume of permeable voids Absorption is a method of determining the water absorption after immersion in water at room temperature, or after immersion and boiling in water. The ‘volume of permeable voids (VPV)’ is the volume of water absorption after a period in boiling water of a hardened concrete sample. The high temperature affects both the viscosity and the mobility of the water molecules which may enable the greater displacement of pore system within the hardened concrete. This is shown in the relationship between absorption and the absorption after boiling in Figure 5. Boiling resulted in a 6% increase in absorption with the exception of a few outlying points.

Table 5 VicRoads Classification for Concrete Durability based on the VPV

Durability classification indicator

Vibrated cylinders (VPV %)

Rodded cylinders (VPV %)

Cores (VPV %)

1 Excellent <11 <12 <14

2 Good 11–13 12–14 14–16

3 Normal 13–14 14–15 16–17

4 Marginal 14–16 15–17 17–19

5 Bad >16 >17 >19

8


The ASTM C642 method measures the volume of permeable voids (VPV) as a percentage of the volume of the solid. The Australian Standard AS 1012.21 test method has been adapted from the ASTM method. It measures the apparent volume of permeable voids (AVPV) as a percentage of the volume of the bulk materials, i.e., solid and voids. The AVPV has been used by VicRoads to classify concrete durability as shown in Table 5.

y = 1.0604x

R2 = 0.9892 or R=99%

0

1

2

3

4

5

6

7

0 1 2 3 4 5 6

Absorption (%)

Ab

so

rpti

on

aft

er

bo

ilin

g (

%)

Figure 5 Effect of boiling on absorption

(Sherman et al 1996)

y = 6.0363x - 77.053

R2 = 0.7353 or R=86%

R2 = 0.015 or R=12%

0

10

20

30

5 10 15 20

ASTM C642 Volume of Permeable Voids (%)

Dif

fus

ion

Co

eff

icie

nt,

De

.36

5 (

10

-12 m

2/s

) CSIRO

Sherman et al.

Figure 6 ASTM VPV28 & De.365 (CSIRO, 1998) and ASTM VPV42 & Da.365 (Sherman et al 1996)

The chloride resistance of concrete has been found to improve with the reduction in VPV as shown in Figure 6. The resistance based on longer-term effective chloride diffusion coefficients, De.365, has been found to correlate well with the VPV with a correlation coefficient of 86% for VPV exceeding about 13%. At lower VPV in the range of 6–14%, Sherman et al (1996) found no correlation (R=12%) between VPV and medium-term apparent chloride diffusion coefficients, Da.365. as shown in Figure 6. From the two sets of slightly different data, there appears to be a significant change of chloride resistance around the critical VPV level of 12–13%.

There is no precision statement for VPV in ASTM C642 and AS 1012.21. Test results compiled by Walpole (2009) of two concrete samples tested by nine laboratories showed the repeatability and reproducibility coefficient of variation of 1.8% and 6.5% respectively. Hence the results of two properly conducted tests by the same operator should not differ by more than 5.1%. The results of two properly conducted tests in different laboratories should not differ by more than 18.4%.

3.4 Rapid Chloride Permeability Test

The rapid chloride permeability test (RCPT) was first developed as a rapid means of assessing permeability of concrete to chloride ions. It was adopted as a standard test method for rapid determination of the chloride permeability of concrete by the American Association of State Highway and Transport Officials as AASHTO T277 in 1989 and subsequently as ASTM C1202-05.

The technique basically measures chloride ion migration or the electrical conductance of concrete. In this method, a potential difference of 60 V dc is maintained across the ends of 51-mm-thick slices of a concrete cylinder, one of which is immersed in a sodium chloride solution, the other in a sodium hydroxide solution. The amount of electrical current passing through the concrete during a six-hour period is measured and the total charge passed, in coulombs, is used as an indicator of the resistance of the concrete to chloride ion penetration. A table classifying concrete resistance to chloride ion penetrability has been proposed as reproduced in Table 6.

9


Table 6 Chloride Ion Penetrability Based on Charge Passed (Whiting, 1981)

Charge Passed Chloride Ion Penetrability

>4,000 High

2,000–4,000 Moderate

1,000–2,000 Low

100–1,000 Very Low

<100 Negligible

In Australia, the RCPT was first used in the early 1990s to specify highly durable concrete for the sea wall at Sydney Airport Parallel Runway project at the 1000-coulomb limit. According to ASTM C1202, the repeatability and reproducibility coefficient of variation of are 12.3% and 18.0% respectively. Hence the results of two properly conducted tests by the same operator should not differ by more than 34.8%. The results of two properly conducted tests in different laboratories should not differ by more than 50.9%.

The RCPT charge passed has been found to be a reasonable indicator of the chloride resistance of concrete. The charge passed measured at 28 days or 42 days, RCPT28 or RCPT42, correlates well with the longer-term apparent or effective chloride diffusion coefficients, Da.365 or De.365, as shown in Figure 7. The increase in the age of testing from 28 to 42 days is not expected to affect the charge passed by concrete with ‘low’ and ‘very low’ chloride penetrability and hence RCPT42 Data (≤1000 coulombs) are combined with RCPT28 to show the combined relationship for the full range of charge passed shown in Figure 7.

Concrete with RCPT charge passed below 2000 coulombs shows consistently good chloride resistance with Da.365 below 3x10

-12 m

2/s. whereas concretes in the 2000- to 3000-coulomb range

exhibit highly variable chloride resistance.

Sherman et al (1996) found that the correlations between RCPT and longer-term chloride diffusion, the surface chloride concentration, and the time-to-corrosion to be highly variable and that it requires individual correlation between the test and every concrete mixture. The widely used 1000-coulomb cut-off limit was claimed to be arbitrary and misleading for many concretes, due to the widely different chloride permeability observed for concretes both meeting and failing such a coulomb limit-based specification. The use of heat curing was found in the same study, for example, to increase the coulomb values of concrete without increasing its actual chloride permeability.

y = 0.0071x - 13.61

R2 = 0.5836 or R=76%

Sherman R=89%

y = 0.4072e0.0008x

Combined R=77%

0

10

20

30

1000 2000 3000 4000 5000 6000

ASTM C1012 Charge Passed (coulomb)

Dif

fus

ion

Co

eff

icie

nt,

D3

65 (

10

-12 m

2/s

)

CSIRO

Sherman et al.

Combined

Figure 7 RCPT28 versus De.365 and RCPT42 versus Da.365

10


4 Comparative Performance Data

The suitability of each property as a performance-based indicator for chloride resistance of concrete is considered from the precision of the test method, its relationship with the longer-term chloride diffusion resistance, De.365, evaluated as the correlation coefficient, R for the corresponding range of measured values. These are shown in Table 7.

With the exception of water-cement ratio which is binder dependent, short-term diffusion coefficient De.28, sorptivity, VPV and RCPT correlate reasonably well with the longer-term chloride diffusion coefficient De.365. The VPV, however, shows no correlation with the longer-term chloride diffusion coefficient in the range of 6–13 % and further work may be required to support this. In setting the performance limits for each chloride resistance property, the sensitivity of each property should be considered in terms of the ‘critical range’ and the precision of the test. The critical range chosen for comparison in Table 7 is the range of each property required to detect a change of one order of magnitude in De.365 from 1.0 to10.0 x10

-12 m

2/s.

Table 7 Precision and correlation of each property with chloride resistance

De.28 Sorptivity28 VPV28 RCPT28

(10-12

m2/s) (mm/s

½) (%) (Coulomb)

Precision from relevant standards or other sources

Sources of Precision ASTM C1566 ASTM C1585 ASTM C642 ASTM C1202

Reportable to 0.001 0.1x10-4 0.1 NA

Repeatability CV 14% 6% 1.8%(2) 12.3%

Reproducibility 20% NA(1) 6.5%(2) 18%

Correlation to De.365 of a GP & 3 GB cement concretes (CSIRO, 1998)

Range of value Correlation coef, R

10–60

74%

15–100mm

76%

6–13, 13–18

12%. 86%(3)

2–5 x103

76%

Critical range: one order change in De.365 from 1 to10 m2/s

28 41 mm 1.5 1270

Increment used in classifications RTA B80 12 mm VicRoads 1–2% C1202 1000

(1) Not available (2) Analysis of proficiency test data compiled by Walpole (2009). (3) Correlation R of 86% and 12% are applicable to VPV range of 13–18% and 5–14% respectively.

5. Conclusion

Concrete provides physical and chemical protection to the reinforcing steel against penetrating chlorides which may cause steel depassivation, leading to increased risk of steel corrosion. The chloride resistance of concrete depends on the permeability properties of the concrete and the cover thickness to the reinforcement. The integrity of the concrete cover under service load, in terms of cracking and crack width, also influences the resistance to penetrating chlorides.

The use of prescriptive specifications of cement type and w/c has been found to be extremely effective in specifying chloride resistance. The effect of the type of cement on chloride resistance of concrete is quite evident for concrete at higher w/c, but such distinctions diminish for concrete at low w/c below about 0.45 (0.35 for concrete with silica fume). There is, however, no acceptable method of testing w/c of hardened concrete; control of w/c can be only conducted at the batch plant.

For performance-based specifications, capillary absorption properties (sorptivity and VPV) and migration property (RCPT) have all been found to be as good an indicator of longer-term chloride resistance as the short-term chloride diffusion coefficient. Critical comparative performance data are given in Section 4. It was found that there are limits to the effective range of sorptivity (15–100 mm), VPV (13–17%), and RCPT (1000-5000 coulombs) that are sensitive to chloride resistance.

11


6. References

ASTM C 1202 – 05 Standard Test Method for Electrical Indication of Concrete’s Ability to Resist Chloride Ion Penetration.

ASTM C 1556 – 04 Standard Method for Determining the Apparent Chloride Diffusion Coefficient of Cementitious Mixtures by Bulk Diffusion.

Cao, H.T. and Sirivivatnanon, V., ‘Service Life Modelling of Crack-freed and Cracked Reinforced Concrete Members subjected to Working Load’, Proceedings CIB Building Congress 2001, Wellington, New Zealand, 2-6 April, 2001.

Cement Concrete & Aggregate Australia ‘Chloride Resistance of Concrete’, CCAA Report, June 2009.

CSIRO ‘Broadening the use of fly ash concretes within current specifications’, CSIRO Report BRE045 by Khatri R P, Hirschausen D and Sirivivatnanon V, 1998.

CSIRO ‘Chloride Diffusion Resistance of Fly Ash Concrete’, CSIRO Report BIN 069, ADAA and Pacific Power, by Cao, H. T., et al., October, 1993.

Jang et al. ‘Effect of crack width on chloride diffusion coefficients of concrete by steady-state migration tests’, C&CR 41 (2011) 9-19.

Kropp, J. and Hilsdorf, H. K., Performance Criteria for Concretre Durabilty, E & F N Spon, London, 1995.

Lim, C.C. 'Influence of Cracks on the Service Life Prediction of Concrete Structures in Aggressive Environments', a University of New South Wales PhD Thesis, 2000.

Khatri, R.P. and Sirivivatnanon, V., ‘Characteristics Service Life for Concrete Exposed to Marine Environments’, Cement and Concrete Research, Vol. 34, No.5, May 2004, pp. 745-752

RTA Test Method T362 Interim test for verification of curing regime – Sorptivity, February 2001.

RTA ‘QA Specification B 80 – Concrete Work for Bridges’, Edition 5, Revision 3, June 2008.

Sherman M R, McDonald D B and Pfeifer D W ‘Durability Aspects of Precast Prestressed Concrete. Part 2: Chloride Permeability Study,’ PCI Journal, Vol. 41, No. 4, July-Aug, 1996, pp. 76–95.

Sirivivatnanon, V., Castel, A. and François, R., A discussion on ‘Propagation of reinforcement corrosion in concrete and its effects on structural deterioration’ by C.Q Li and J.J. Zheng, Magazine of Concrete Research 2007, 59, No. 2, March 2007, 151-154.

Sirivivatnanon, V., Gurguis, S. and Castel, A., ‘Development of Limit State Service Life Design’, Proceedings Concrete 2007, Adelaide, Australia, 17-20 October 2007a, 741-750.

Tang, L., ‘Chloride Transport in Concrete – Measurement and Prediction’, Publication P-96:6, Chalmers University of Technology, Göteborg, Sweden, 1996.

Vallini, D. and Aldred, M., ‘Durability Assessment of Concrete Specimens in the Tidal and Splash Zones in Fremantle Port’, Coasts & Ports Australasian Conference, 2003.

Walpole, M., 2008-2009 AVPV Proficiency Test Results. Private communication.

Whiting, D., Rapid Determination of the Chloride Permeability of Concrete Final Report No. FHWA/RD-81/119. Federal Highway Administration, August 1981, NTIS No. PB 82140724.

Sirivivatnanon, Kuptitanhi and Lim

First Asia/Pacific Durability of Building Systems Conference

Harmonised Standards and Evaluation, Bangkok, Thailand, 8-10 September, 1999.

1

COMPLIANCE ACCEPTANCE OF CONCRETE

Vute Sirivivatnanon, Boonrawd Kuptitanhi and Lim Char Ching

*

CSIRO Building, Construction and Engineering, Australia The Concrete Products and Aggregate Co. Ltd., Thailand

*Public Works Department, Malaysia

ABSTRACT

The quality of in-situ concrete, both in strength and durability, is measured by

testing of concrete samples taken on site. The test results represent the

'potential' quality of concrete sampled. The correct assessment of such results,

together with good concreting practices, will greatly increase the certainty of

achieving the desired quality.

The concept of "average of a group of n samples" used in compliance criteria,

such as those specified in the British and American Standard, is reviewed in

comparison with the "single test" method. It is clear that the "average of a group

of n samples" criterion is more sensitive and can be more equitable in reducing

the producer's risk of having good concrete being judged as non-complying, and

the consumer's risk of accepting sub-standard concrete. The concept can be

effective provided that right passing limit is set as demonstrated in an operating

characteristic curve. In the two examples examined, the limit set in BS 5328 is

far more effective than the one set in ACI 318.

Keywords: compliance criteria, producer’s risk, consumer’s risk, operating characteristic

curve.




2

1. INTRODUCTION

More and more emphasis is being given to the use of concrete strength to control the quality

of concrete, both in terms of strength and durability. This is necessary because it is extremely

difficult, in practice, to control durability factors such as maximum water to cement ratio and

minimum cement content. Local knowledge of the relationship between water to cement ratio

and strength, and between cement content and workability of specific grade of concrete,

enables the designer to select a strength grade so as to achieve both the strength and durability

required.

The quality of concrete in the structure is likely to be attained if the potential quality of fresh

concrete can be reasonably assured, and that this potential quality is realised in the structure

through good concrete practices. Such a potential quality of fresh concrete is normally

obtained by regular, but random, sampling of the fresh concrete in cubes or cylinders. These

samples are tested and the results, interpreted in accordance with compliance criteria, are an

indication of whether the desired potential quality of the fresh concrete is being achieved.

It can be appreciated that the correct interpretation of such results is essential towards the

achievement of strong and durable concrete structures.

2. ACCEPTANCE LEVEL

The quality of each batch of concrete differs from the next because of the variation in the

quality of all the constituents. In most cases, the quality distribution conforms to a normal

distribution curve. This forms the basis for statistical analysis that follows.

Building codes specify that the quality of concrete is considered satisfactory if the majority of

the batches exceed a level of specified strength. That is a minimum number of batches are

allowed to be ‘defective’. Many codes accept up to five percent of defective concrete. The

American code ACI 318-95 (1), however, accepts up to ten percent of defective concrete. In

more critical structures, such as parts of nuclear power station, a much more stringent limit is

usually set.

3. TESTING PLAN

An effective testing plan is one that provides for a continuing assessment of the quality of all

the concrete produced, but divided into small lots for compliance decisions. The effectiveness

of a testing plan is also judged by the optimum risks of both the producer (contractor) and the

consumer (engineer), that is the minimum producer's risk of having good concrete rejected

and the consumer's risk of accepting sub-standard concrete.

The effect of using different sampling and testing schemes on the producer's and consumer's

risk can be compared by means of Operating Characteristic curves (O-C curves). The O-C

curve of a testing plan describes the probability of acceptance of any given quality defined in

terms of the percentage proportion of defectives.

If one is prepared to accept 5% defective concrete below the specified strength, the ideal

testing plan would be one in which concrete of a quality equal to that specified i.e. 0 to 5%

defective, was always accepted so that the producer's risk was nil, and that concrete of a

quality worse than that specified, i.e. 5 to 100% defective, was always rejected so that the

consumer's risk was also nil. This ideal curve is shown as curve B (acde) in Figure 1. This is

only attainable with a theoretically infinite population of results.




3

Curve A (ab) in Figure 1 is for the specification that "no single test shall fall below the

specified strength." The probability of acceptance is the same as the probability of choosing

an acceptable result from the population (represented by the unshaded area under the normal

distribution curve) shown in Figure 2.

Curve B

(ideal test)

Curve A

(single tests)

Curve C

(average of fours)

p

d

a c

b

f

e0 10 20 30 40 50

0

10

20

30

40

50

60

70

80

90

100

TRUE PERCENTAGE DEFECTIVE TO CHARACTERISTIC STRENGTH

PR

OB

AB

ILIT

Y O

F A

CC

EP

TA

NC

E -

%

Figure 1 Operating characteristic curves for various testing plans.

-100 0 100 200 3000

1

2

3

4

5

L = limiting value for group of n samples

mean strength Xmean

k x s/2

margin = k x s

Xmeanf'c

5% below f'c - kxs/2

5% below f'c

Normal distribution of

groups of n samples from

the same population


individual samples

Figure 2 Spread of individual samples represented by a normal distribution curve and

group of n samples from the same population.

The criterion of acceptance is unsatisfactory for it involves a very high consumer's risk. There

is a 50% chance of accepting concrete containing 50% defectives, ie. accepting a concrete

whose true mean strength is equal to the specified strength.




4

4. SENSITIVITY OF A GROUP OF N SAMPLES

Instead of trying to estimate the characteristic strength, it is possible to devise a plan to give a

reasonable assurance that the quality of concrete is at least as good as that specified based on

a small number of results. The ‘average of a group of n samples’ criterion is one such plan

which gives the engineer reasonable assurance that the concrete is at least as good as that

assumed in design. The criterion is based on the assumption that the overall standard

deviation of strength is constant and equal to the value adopted in setting the current margin

for the design of the concrete mix. This is reasonable, because the standard deviation remains

sensibly constant on a job provided that the overall control remains uniform.

The criterion is also based on the fact that, when n independent samples are taken at a time

from a normal distribution with a mean xmean, and a standard deviation s, the averages

calculated from the sets of n samples are also distributed normally with the same mean xmean

but with a standard deviation s/n . Hence, when n=4 for example, the standard deviation of

the averages of the four results is s/2. A limiting value, L, for the average of four tests can be

chosen such that, when the overall distribution of the individual strength results is exactly 5%

defective to the characteristic strength, 95% of the distribution of the averages of four tests

exceeds the limiting value L. Under these conditions L is 1.64 x s/2 below the mean (xmean)

and, as the margin between the specified characteristic strength and the mean is 1.64s, the

limiting value L is thus half the current margin above the characteristic strength as shown in

Figure 2.

L = limiting value for group of n samples

-100 -50 0 50 100 150 200 2500

1

2

3

4

5Original mean strength Xmean

Xmeanf'c

51.5% below (f'c - k x s/2)

20% below f'c


groups of n samples from

the same population


individual samples

An example of a reduction in

mean strength Xmean by k x s/2

Figure 3 Graphical demonstration of the effect of a drop in quality (xmean) in the

detection of defective samples below a limiting value L.

If the true quality of production deteriorates, as reflected by a shift of the mean strength

towards the characteristic strength, the standard deviation remaining the same, so that if there

is a 20% defective rate, the distribution of the average of four tests also move towards the

characteristic strength and the limit L, so that there is a 51.5% chance that a single average of




5

four tests will fall below the limit L. This is demonstrated graphically in Figure 3. The

corresponding O-C curve C given in Figure 1.

In most testing plan, there is also the requirement that no individual test result is less than a

limit, such as 85% of the specified characteristic strength. This provides a ‘long-stop’ below

which no results should fall and the means by which an individual batch can be judged.

However, if the concrete is being produced to the characteristic strength, the probability of

obtaining such a low result is remote.

5. COMPLIANCE CRITERIA

Different codes specify different compliance criteria. Most have adopted 2 criteria; the first

being the averages of a group of n samples and the second of individual samples. The more

common criteria are shown in Table 1. The derivation of the O-C curve of the "average of

four" criteria (Figure 1), specified in the British Code of practice, CP 110 (2), is shown

graphically in Figure 1.

It can be seen from the O-C curve for the "average of four" criteria (curve C, af) in Figure 1

that there is a significant improvement from the single test criterion and that both the

producer's and consumer's risks move closer towards the ideal curve B. At higher rate of

defective concrete eg. 20%, there is a consumer's chance of 50% of detecting the poor

concrete. This may seem a somewhat high risk for the consumer but, in fact, concrete with

such a high defective rate is seldom produced; in any case, the chance of detecting

sub-standard concrete is rapidly increased by the successive application of the "average of

four" test plan (3).

Table 1: Compliance criteria for strength of concrete in different codes

Acceptance Criteria, N/mm2

Acceptable

Country Rule 1 Rule 2 percentage

Code No. of

consecutive

strength

Minimum of groups

average

Minimum of

individual sample of defective

concrete

U.K.

CP 110

BS 8110

(BS 5328)

4

4

f’c *1

+ 0.82s

f’c + 3

0.85 f’c *2

f’c – 3.0

5

U.S.A.

ACI 318-95

(ACI 214-97)

3

f’c

f’c – 3.5*2

10

Germany

DIN1045

3

9

f’c + 5

f’c + 5

f’c

0.8 f’c

5

1. f’c = Characteristic strength value in N/mm2

2. s = standard deviation

3. 3.5 N/mm2 is approximately 500 psi.




6

5.1 British Standard

British Standard BS 8110:1985(4), refers to BS 5328:1990(5), for compliance criteria. The

latter requires that, where compressive strength is specified, compliance with the

characteristic strength shall be assumed if the conditions given in both (a) and (b) are met:

(a) the average strength determined from any group of four consecutive test results

exceeds the specified characteristic strength by

3 N/mm2 for concretes of grade C20 and above

2 N/mm2 for concretes of grade C7.5 to C15, and

(b) the strength determined from any test result is not less than the specified characteristic

strength minus

3 N/mm2 for concrete of grade C20 and above

2 N/mm2 for concretes of grade C7.5 to C15.

The use of a fixed margin of 6 N/mm2, for grade C20 and above, is more practical compared

with a variable current margin (1.64 s) specified in CP 110. However, the effectiveness of

BS 5328 criteria becomes as good as those in CP 110 only when the standard deviation is

equal to 3.66 (6/1.64). If the standard deviation exceeds 3.66, as will be common in practice,

the consumer's risk increases.

5.2 American Standard

American Concrete Institute ACI 318-95, Building Code Requirements for Reinforced

Concrete Structures, clause 5.6.2.3 stipulates that strength level of an individual class of

concrete shall be considered satisfactory if both of the following requirements are met:

(a) Average strength of all sets of three consecutive strength tests equal or

exceed f’c

(b) No individual strength test (average of two cylinders) falls below f’c by

more than 500 psi.

The required average strength f’cr, (in psi) given in clause 5.3.2 of the same standard,

stipulates the use of the larger of Eq. (5-1) or (5-2) as the basis for selection of concrete

proportions.

f’cr = f’c + 1.34s (5-1)

or

f’cr = f’c + 2.33s - 500 (5-2)

It was stated in the Commentary that Eq. (5-1) provides a probability of 1-in-100 that

averages of three consecutive tests will be below the specified strength f’c. Eq. (5-2) provides

a similar probability of individual tests more than 500 psi below the specified f’c. These

equations assume that the standard deviation used is equal to the population value appropriate

for an infinite or very large number of tests.




7

Based on the concept of the sensitiveness of the ‘average of a group n samples’ discussed

earlier and demonstrated in Figure 3, it can be seen that the benefits of the use of this type of

criterion can only be realised with the use of a suitable limit L. With L equal f’c as stipulated

in (a), it is believed that the criterion is less effective. This is clearly seen in the O-C curves

shown in Figure 4. While it gives high probability of acceptance for the defective percentages

between 0-10% and hence very little producer’s risk of having good concrete rejected, the

probability of acceptance remains very high when defective percentages exceed 10%. This

means a very high risk to consumer to accept defective concrete.

The O-C curves for criteria specified in ACI 318 and BS 5328 (for a standard deviation s = 4

MPa) are also compared Figure 4. It is clear that the BS criterion resulted in significant

reduced consumer’s risk to accept defective concrete than the ACI criterion.

0 10 20 30 40 500

10

20

30

40

50

60

70

80

90

100

TRUE PERCENTAGE DEFECTIVE TO CHARACTERISTIC STRENGTH

PR

OB

AB

ILIT

Y O

F A

CC

EP

TA

NC

E -

%

BS 5328

ACI 318

single test

Figure 4 Comparison of the operating characteristic curves for BS and ACI testing plans

6. CONCLUSIONS

1. A testing plan, based on a small number of n samples applied successively, is

effective in dividing concrete production into small lots for compliance decisions.

2. Provided that a reasonable limit is chosen for such a compliance criterion, it can

be effective in balancing the risks of both the producer and consumer. A good theoretical limit

is ks/n below the mean strength corresponding to the target mean strength used in the mix

design. s is the standard deviation of the population and n is the number of samples in each

group.

3. The use of a fixed margin such as the one specified in British Standard is more

acceptable in practice. The fixed margin should be chosen in accordance with the level of

quality control achieved in the locality, as reflected by the average standard deviation.




8

7. REFERENCES

1) American Concrete Institute, ‘Building Code requirements for reinforced concrete’,

(ACI 318-95), Michigan.

2) British Standards Institution, ‘CP 110: Part 1: 1972. The Structural use of concrete

Part 1 Design, materials and workmanship’, London

3) Cement and Concrete Association, ‘Handbook on the Unified Code for Structural

Concrete’, (CP 110:1972), edited by S.C.C. Bate, et al., London 1972.

4) British Standards Institution, ‘BS 8110: Part 1 : 1985. Structural use of concrete,

Part 1 Code of practice for design and construction’, London.

5) British Standards Institution, ‘BS 5328: Methods for Specifying concrete, Part 4,

1990’, London.

6) American Concrete Institute, ‘Building Code requirements for reinforced concrete’,

(ACI 214-77, re-approved 1997), Michigan.

Documents

Chloride Resistance of Concrete · Report Cement Concrete & Aggregates Australia Cement Concrete & Aggregates Australia