new NIST PUBLICATION

June 12, 1989

NISTIR 89-4086

Service Life of Concrete

James R. Clifton and Lawrence !. Knab

U.S. DEPARTMENT OF COMMERCE

National Institute of Standards and Technology
National Engineering Laboratory
Center for Building Technology
Building Materials Division
Gaithersburg, MD 20899

June 1989

Prepared for:
U.S. Nuclear Regulatory Commission
Washington, DC 20555
NISTIR 89-4086

Service Life of Concrete

James R. Clifton and Lawrence I. Knab

U.S. DEPARTMENT OF COMMERCE

National Institute of Standards and Technology
National Engineering Laboratory
Center for Building Technology
Building Materials Division
Gaithersburg, MD 20899

June 1989

National Bureau of Standards became the

National Institute of Standards and Technology
on August 23, 1988, when the Omnibus Trade and
Competitiveness Act was signed. NIST retains
all NBS functions. Its new programs will encourage
improved use of technology by U.S. industry.

Prepared for:
U.S. Nuclear Regulatory Commission
Washington, DC 20555

U.S. DEPARTMENT OF COMMERCE

Robert Mosbacher, Secretary
NATIONAL INSTITUTE OF STANDARDS
AND TECHNOLOGY
Raymond G. Kammer, Acting Director
 VMV-S-* ..viartoW^O RsW'sjf't

^ j ^'1; rq3-.Mi'(AR3C .i^ U

r '..n _.
fo isrioikH

tT.'v''of^'!l>/AT lot
. rOi^ivid QnlfWJiiiS

c'?80?. CM ,;;pucJ3i9*^fe£!

•
^ntH,

.»_!. i _ ..1A^ t’il

:
.

‘>0 vj'‘
.

),'\'''^l?ii!4

.
'
)v’v;)V<‘ -AC, 7
-.
.:r-N >iJ(;=>qnii^D
•(„ •••
.i.u.’ c'^HI.?,- T,
'k '.(.'•'iub’^Vvn'j?nf lie,.

’'0

.lojjbrf’ts.qe'?'-')

. .
(>>w) y .3 .U
(;'E30S 00

00 '-^O TMaiSATBA«lSa z:o

v)onp«?dtoM
ai2AC2'..'iAT8 ^0 Bivrnm ja^^oitam '

Y?>OJOMHOarpi^A'
‘ic;u>?!'das<t4l§A 'iO
 EXECUTIVE SUMMARY

The U.S. Nuclear Regulatory Commission (NRC) has the

responsibility for developing a strategy for storing low-level
radioactive waste (LLW). One approach being considered involves
storing radioactive wastes in concrete vaults which are either i)
buried in the earth or ii) constructed above ground and covered
with earth. In either scenario, the vaults will be below the
frost line and above the groundwater table. A service life of
500 years is required for the storage vaults. The National
Institute of Standards and Technology is carrying out a project
for the NRC aimed at determining if concrete can be developed
which will have the desired service life. If the results of the
project indicate that a 500 year life for concrete is achievable,
performance criteria will then be developed to form a technical
bases for selecting concrete materials. This report is based on
an analysis of the major degradation processes which are likely
to effect the performance of the underground storage vaults and
on methods for predicting the service life of concrete.

The major degradation processes that concrete of underground

vaults will likely encounter are sulfate attack, corrosion of
reinforcing steel, alkali-aggregate reactions, and leaching by
ground water. Freezing and thawing damage could occur before the
vaults are covered with soil and therefore are addressed. Other
degradation processes which could occur are microbiological
attack, salt crystallization, and attack by LLW, especially
acids. All of these processes involve the penetration of the
concrete by water or aqueous solutions. Therefore, it is
concluded that concrete with low permeabilities are most likely
to achieve service lives of around 500 years.

The permeability of concrete is dependent on the porosity and

pore structure of the cement matrix, especially the extent of
pore interconnectedness. Many factors affect the pore properties
of concrete with the water-to-cement ratio having the greatest
effect. Cracking also can significantly increase the
permeability of concrete and can limit the beneficial effects of
producing low water-to-cement concrete.

An approach for predicting the service life of the concrete is

presented. It involves the integration of mathematical models
and accelerated durability testing. Several models have been
reported for predicting the service lives of concrete, including
models for concrete exposed to corrosive conditions, freezing and
thawing conditions, and to groundwater. However, these models
appear to give at the most qualitative estimates of service
lives. Further work on the development of models is recommended
to form a basis for making reasonable estimate of service lives.

iii
The rates of most of the likely degradation processes can be
predicted based on modeling and accelerated testing. An
exception is alkali-aggregates reactions. Present test methods
are not adequate to predict the behavior of potentially reactive
aggregates in field concrete. Furthermore, the present knowledge
on alkali-aggregate reactions is inadequate for making service
life predictions. It is, therefore, recommended that a test
method be developed for determining if an aggregate is
potentially reactive. An aggregate would then be accepted or
rejected based on its reactivity rather than on its rate of
reaction.

The long-term performance of concrete is often controlled by the

level of its quality. Based on available knowledge, it seems
likely that concrete can be formulated which should have service
lives of 500 years in the expected environments. However, such
lives will be obtainable only if the concrete materials meet
carefully defined specifications, and good construction practices
are followed.

iv
 ABSTRACT

The U.S. Nuclear Regulatory Commission (NRC) has the

responsibility for developing a strategy for the disposal of low-
level radioactive waste (LLW). An approach being considered for
their disposal is to place the waste forms in concrete vaults
buried in the earth. A service life of 500 years is required for
the concrete vaults as they may be left unattended for much of
their lives.

This report examines the basis for making service life

predictions based on accelerated testing and mathematical
modeling of factors controlling the durability of concrete buried
in the ground. Degradation processes are analyzed based on
considerations of their occurrence, extent of potential damage,
and mechanisms. A recommended research plan for developing
methods for predicting the service life of concrete is presented.

The major degradation processes that concrete of underground

vaults will likely encounter are sulfate attack, corrosion of
reinforcing steel, alkali-aggregate reactions, and leaching by
ground water. Freezing and thawing damage could occur before the
vaults are covered with soil and therefore are addressed. Other
degradation processes which may occur are microbiological attack,
salt crystallization, and attack by LLW, especially by acidic
materials. Two important factors controlling the resistance of
concrete to these degradation processes are its quality and
permeability. Concepts of quality and factors affecting quality
of concrete are discussed. Permeability is discussed in terms of
the water-to-cement ratio, the pore structure of concrete, and
the effects of cracks.

KEYWORDS Accelerated testing; alkali-aggregate reactions;

concrete corrosion; durability; groundwater; mathematical
modeling; permeability; service life; sulfate attack.

V
 . 3

Table of Contents
Page

EXECUTIVE SUMMARY iii

1 . INTRODUCTION 1

1.1 Background 1
1.2 Definition and Concepts of Durability and Service
Life 2
1.3 Scope and Outline of Report 4

2. QUALITY OF CONCRETE 5

2.1 Concepts of Quality 6

2.2 Quality of Concrete 7

3. PERMEABILITY OF CONCRETE 8

3.1 Effect of Porosity on Permeability 8

3.2 Controlling the Permeability of Concrete ........... 14
4 SULFATE ATTACK 18

4.1 Mechanism 20
4.2 Resistance of Concrete to Sulfate Attack ........... 22
4.3 Models and Service Life Predictions 24

5. CORROSION OF STEEL REINFORCEMENT IN CONCRETE 29

5 . 1 Mechanism 30
5.2 Resistance of Reinforced Concrete 33

5.2.1 Environment 33
5.2.2 Reinforcing Steel 34
5.2.3 Alkalinity of Pore Solution 35

5. 2. 3.1 Effect of pH on Corrosion Threshold .... 35

5.2. 3.2 Pozzolans 35
5. 2. 3. Carbonation 36

5.2.4 Domestic Chloride Ions 37

5.2.5 Concrete Cover and Quality 38
5.2.6 Stray Currents 40

5.3 Models and Service Life Predictions 40

5.3.1 Model by Tuutti 40

5.3.2 Model for Corrosion by Sea Water 46
5.3.3 Empirical Model 47

6. ALKALI -AGGREGATE REACTIONS 48

vi
 Table of Contents (Continued)
Page

6.1 Alkali-Aggregate Reaction 50

6.1.1 Mechanism 50
6.1.2 Determination of Aggregate Reactivity 52

6. 1.2.1 Petrographic Examination 53

6. 1.2. 2 Mortar-Bar Test for Potential Reactivity 53
6. 1.2. 3 Chemical Test for Potential Reactivity . 54
6.1.2.4 Combination of the Methods 54

6.1.3 Prevention of Alkali-Aggregate Expansive

Reactions . 55

6.2 Alkali-Carbonate Reaction 57

6.2.1 Mechanism 57
6.2.2 Determination of Potential Reactivity of
Aggregates 58
6.2.3 Prevention of Alkali-Carbonate Reaction 58

6.3 Models and Service Life Predictions 59

7. FREEZING AND THAWING 61

7 . 1 Mechanisms 62

7.1.1 Hardened Cement Paste 62

7.1.2 Freezing in Aggregates 65

7.2 Factors Controlling Resistance to Frost Damage 65

7.2.1 Exposure to Moisture 66

7.2.2 Water-Cement Ratio 66
7.2.3 Entrained Air 66
7 o 2 . 4 Materials and Tests 68
7.2.5 Curing 71

7.3 Models and Service Life Predictions 71

7.3.1 Service Life Prediction Model 71

7.3.2 Model of Frost Penetration 73
7.3.3 Modeling of Freezing and Thawing Processes .. 73

8. DETERIORATION OF CONCRETE BY LEACHING 74

8.1 Mechanism 75
8.2 Factors Influencing Leaching 76

vii
 Table of Contents (Continued)

Page

8.2.1 Permeability of Concrete 76

9. 8.2.2 Underground Water Movement 77
8.2.3 Chemistry of Underground Water 77
8.2.4 Chemistry of Hydrated Cement Paste 79

8.3 Models and Service Life Predictions 79

OTHER DETERIORATION PROCESSES 83

9.1 Microbiological Attack 83

9.2 Salt Crystallization 86
9.3 Attack by LLW 87

10. EMPIRICAL KINETIC MODELS 88

11. RECOMMENDED RESEARCH PROGRAMS 90

11.1 Candidate Concretes 93

11.2 Recommended Approach for Predicting Service Lives . 95

11.2.1 Accelerated Testing 100

11.2.2 In-Service Concrete Studies 102
11.2.3 Models for Service Life Prediction ......... 103
11.3 Research Priorities 107

12. SUMMARY and CONCLUSIONS 109

13 . ACKNOWLEDGMENTS Ill

14 . REFERENCES ... 112

APPENDIX A. Barrier Systems for Protecting Concrete

Surfaces A-1

APPENDIX B. Some Common Chemicals and Substances Aggressive

to Portland Cement Concrete B-1

viii
 LIST OF FIGURES
Page

Figure !• QA/QC activities in concrete construction 9

Figure 2. Effect of capillary porosity in hardened cement

paste on permeability (9) 12

Figure 3. Effect of water-to-cement (w/c) ratio on

permeability of concrete (14) «,o.. 13

Figure 4. Effect of water-to-cement (w/c) ratio on

permeability of hardended cement paste (19) 16

Figure 5. Major causes of cracks in concrete ( 22) «««oeeea»« 19

Figure 6. Schematic of corrosion of steel reinforcement

in concrete . 31

Figure 7. Effect of water-to-cement ratio and depth of

concrete cover reinforcement on chloride ion
penetration (38) 39

Figure 8. Schematic of conceptual model of corrosion of

steel reinforcement in concrete (37) 41
11.
Figure 9. Calculated induction period. Effects of
concrete cover, threshold concentration, threshold
concentration of chloride ions, concentration of
chloride ions, and concrete permeability on
calculated induction period (37) 44

Figure 10. Estimated fraction of calcium released per

year from concrete by groundwater (89) 84

Figure Estimated time dependence of pH within concrete

exposed to groundwater ( 89) ........ .9 0. .......... 8S

Figure 12. Typical rates of deterioration. Meanings of t*s

are given in table 5 92

Figure 13 . Methodology for developing and using accelerated

tests in predicting service lives (5) 97

Figure 14. Schematic of proposed conceptual model of sulfate

attack, by ettringite formation, of concrete 104

Figure 15. Method for predicting the advance of threshold

concentration of aggressive solution concrete at
in-service stresses, S^, based on elevated
stresses, S„m '
106
 LIST OF TABLES

Page

Table 1. Classification of Pore Sizes in Hydrated Cement

Pastes (13) 11

Table 2. Recommendations for Sulfate Resistance (23) 25

Table 3. Recommended Air Contents for Frost-Resistant

Concrete (23) 69

Table 4. Composition of Some Natural Groundwaters (95) .... 87

Table 5. Durability of Models (98) 91

X
1 . INTRODUCTION

1. 1 Background

The U.S. Nuclear Regulatory Commission has the responsibility for

developing a strategy for storing low-level radioactive wastes

(LLW) According to one approach, the radioactive wastes would be

stored in concrete vaults which are either buried in the earth or

constmcted above ground and covered with earth. A service life

of 500 years is required for the storage vaults as they may be

left unattended for much of their lives.

Several observations suggest that certain concretes could meet

the 500 year requirement. For example, some concretes placed

during The Roman Age are still intact (1) . These concretes,
however, may not be representative of the typical concrete of

their era and statistically be only a minute sample of the

population (e. g., one concrete structure out of a thousand or so

surviving) . Further, ancient concretes contain cements of

significantly different composition compared to modern cements.

Regarding the durability of modern concrete, there are some

structures, with concretes based on portland cement, which have

been in service for over one hundred years (2) Based on

analysis of the permeability of concrete, Atkinson (3) has

predicted that a high quality concrete could have a service life

greatly exceeding 500 years if the exposure conditions are not
severe. Therefore, based on consideration of the past

durabilities of concrete, it appears possible to design concrete

1
 . .

to give service lives of at least 500 years in favorable

environments

The present basis for selecting concretes and their constituents

needs to be further advanced to ensure that concretes with

expected service lives of at least 500 years are designed. At

least two approaches can be followed. In the first, the most

durable concrete feasible could be developed and assumed that it

will have the desired life. The other approach involves the

development of performance criteria based on considerations of

the factors controlling the service life of concrete. In the

latter approach the results of accelerated durability tests could

be coupled with the development and application of mathematical

models to form a basis for predicting service lives of concrete.

This approach has been described by Pommersheim and Clifton (4)

1.2 Definition and Concepts of Durability and Service Life.

Durability and service life are often erroneously interchanged.

The distinction between the two terms is evident when their

definitions as given in ASTM E 632 (5) are compared:

durability , the capability of maintaining the serviceability

of a product, component, assembly or construction over a

specified time. Serviceability is viewed as the capability

of the above to perform the function (s) for which they are

designed and constructed.

2
 service life (of building component or material) . the period

of time after installation during which all properties

exceed the minimum acceptable values when routinely

maintained.

Both durability and service life incorporate the concept of

design requirements being met or exceeded for a given time. They

can be used as a basis for measuring the adequacy of durability

analysis procedures. Frohnsdorff, et al., (6) suggest the

following procedures:

1) Make a quantitative assessment of time to failure of a

component or material when it is exposed to the expected in-

service conditions. This estimate can be derived from either in-

service performance tests or from accelerated aging tests.

2) Obtain test results which correlate with in-service

performance.

3) Identify the degradation mechanisms causing failures at high

and low stresses.

ASTM E 632 defines accelerated aging tests as ones in which the

degradation of building materials is intentionally accelerated
over that expected in-service. This is done in anticipation that
the effects of in service stress levels can be predicted from

performance at accelerated stress levels. Proper identification

of degradation mechanisms is important in extrapolating test

results obtained from accelerated aging to normal in-service

3
conditions. In the accelerated testing approach it is essential

that the degradation process at accelerated stress levels is the

same as that occurring under the expected service conditions.

I. 3 Scope and Outline of Report

This report examines the basis for making service life

predictions based on accelerated testing and mathematical

modeling of factors controlling the durability of concrete buried

in the ground. Deterioration processes are analyzed based on
considerations of their occurrence, extent of potential

destruction, and mechanisms. Approaches for preventing their

occurrence are also discussed.

Two of the important factors controlling the service life of

concrete are its quality and permeability as discussed in

Sections 2 and 3 . The performance of reinforced concrete

subjected to sulfate attack, corrosive environments, alkali-

aggregate reactions, freezing and thawing conditions, and

leaching by water are covered in Sections 4 through 8. Some

other possible deterioration processes are briefly covered in

Section 9. Several proposed empirically based kinetic models are

presented in Section 10 for major deterioration processes of

concrete. A recommended research plan for developing methods for
predicting the service life of concrete is presented in Section
II. This plan is ambitious and research priorities, therefore,

are included. The performance of coatings and joint materials

4
for concrete are outside the scope of deterioration of concrete

materials and, therefore, they are placed in Appendix A. Lists

of inorganic and organic chemicals and substances which are known

to attack concrete are presented in Appendix B.

The leaching of concrete by groundwater is also covered. It is

recognized that NRC has stipulated that concrete vaults for

underground storage of LLW are to be located above the
groundwater table. However, in some 500 years the groundwater

table could fluctuate. Also, precipitation which is lime-

deficient, or which becomes acidic during percolation through

soil, can be aggressive to concrete.

2. QUALITY OF CONCRETE
The performance of a concrete structure is usually no better than

the quality o^ its materials and construction practices. The

need for quality concrete increases directly with the severity of

requirements placed on its performance. Clearly, the requirement

that concrete to be used to construct storage vaults for LLW must

be durable for 500 years, is a severe demand necessitating that

close attention is given to the factors controlling its quality.

Throughout this report the need for quality concrete is stated.

Therefore, some concepts of quality and factors affecting quality

of concrete are briefly discussed.

5
2.1 Concepts of Quality

Quality of concrete can be broadly defined as the totality of

features and characteristics of concrete that bear on its ability

to satisfy an established requirement. A working definition of

quality often used is "fitness for use" (7) The level of

quality is dependent on proper quality assurance and quality

control practices. These terms have been defined by the Business

Roundtab 1 e ( 8 ) as

"Quality Assurance (QA) - a planned and systematic pattern

of all actions necessary to provide adequate confidence

that a product will conform to established requirements."

"Quality Control (QC) - implements the quality plan by

those actions necessary for conformance to established

requirements .

This is the definition of QC in a narrow sense. In a broader

sense, QC involves quality of conformance plus control of the

quality of design and evaluation of customer satisfaction

(quality plan) . Based on this broader scope, QC of concrete LLW

vaults includes the control of quality by testing of physical and

chemical properties of concrete materials and of fresh and

hardened concrete, inspection of construction processes and

practices, and evaluation of owner satisfaction.

6
Another aspect of an effective QA/QC program is quality audit.
Quality audit is the determination of the conformance to
established requirements. For example, testing of a LLW storage

vault, and its components and concrete, when it is accepted by

the owner, constitutes a quality audit.

2.2 Quality of Concrete

Many QA/QC activities and construction practices affect the

quality of in-place concrete as indicated in Fig. 1. These QA/QC

activities include checking the criteria for selecting concrete

materials and for designing concrete, establishing requirements

for quality, inspecting construction practices, testing of

quality, and quality auditing of completed structure. The

criteria for selecting materials and designing concrete should

include performance and service life criteria. Once performance

requirements are established in the preliminary design, ensuring

that the materials meet the performance requirements is also a QC

function. Construction practices which need inspection include

the proportioning of concrete at the ready-mix plant, mixing of

the constituents, placing and compaction of concrete, and curing

of concrete. It cannot be overemphasized that good construction

practices must be strictly followed in order to achieve concrete

with a service life of 500 years.

7
3 PERMEABILITY OF CONCRETE
3.1 Effect of Porosity on Permeability

Hardened concrete consists predominantly of mineral aggregates

and hardened cement paste. A considerable amount of void space is

distributed within the constituents of hardened concrete and also

in regions between the constituents. However, most of the void

space is located in the hardened cement paste. In cement

technology, pore space in the hardened cement paste can be

categorized into the following sequence of decreasing size ranges

(table 1) : entrained air, capillaries, and gel pores. The

minimum porosity of a completely hydrated cement paste is around

28 percent (9) Permeability of portland cement paste increases

with porosity (Fig. 2) and with the extent that the pores are

interconnected

The water to cement ratio (usually expressed as the w/c ratio) of

concrete is the most important factor in controlling the porosity

of the hydrated cement paste and thus the permeability of

properly prepared concrete. As the w/c ratio is decreased the

porosity and consequently the permeability of concrete is

decreased (Fig. 3) . Permeability is an indicator of the ease of

passage of liquids and gases in concrete. The permeability

coefficient of a liquid flowing through a porous solid can be

expressed by the following form of the D*Arcy equation (10)

Q = -kA//Lt (dP/dx) ( 1)
 • QA/QC planning

• Functional requirements
for concrete

• Performance criteria
for concrete

• Concrete proportions
• Concrete constituents

QC inspection of:
• Ready-mix plant procedures
• Mixing of concrete
QC tests:
• w/c ratio (slump)
• Air content
• Density
• Specimens taken for 28-day
strength measurements

Inspection of:
• Consolidation

• Finishing of concrete surface

Proper curing depends on:

• Temperature
• Relative humidity
• Wind

• Quality audit performed on concrete

Figure 1. QA/QC activities in concrete construction.

9
 ;

where Q is the volumetric flow rate (mVs) , A is the cross

sectional area (m^) (for one-dimensional flow in the x

direction) , and ii is the viscosity of liquid (Pa*s) . P is the

component of pressure capable of causing flow; k is the

permeability coefficient which is a characteristic of the porous

medium and independent of the fluid, provided that the fluid flow
is viscous.

When considering water flow in concrete it is common to calculate

the hydraulic conductivity, K, as given by the equation (10)

Q = AK(dh/dx) (2)

where dh/dx is the hydraulic gradient across a specimen and is

dimensionless. The hydraulic conductivity, K, is often referred

to as the permeability and has the units of m/s. The permeability

of hardened cement paste is around 10-^^ m/s for a w/c of 0.51,

while the permeability for a hardened cement paste in which all

capillary porosity has been eliminated is thought to be less than

10*^^ m/s. Nyame and Illston reported (11,12) that the total

porosity of hardened cement paste is not uniquely related to

permeability, whereas the maximum continuous pore radius is

uniquely related to permeability. The maximum continuous pore

radius is representative of the size of pores in which the water

flows. Depending on its w/c ratio, aggregate contents, and

curing conditions the permeability of mature concrete (concrete

10
 ble 1. Classification of Pore Sizes in Hydrated Cement Pastes (13)

signation Diameter Description Role of Water Paste Properties

Affected

Strapped 0. 1-1. 0mm Round cross- Strength

Air section
^

pillary 10-0.05 um Large Behaves as Strength

pores (50nm) capillaries bulk water permeability

50-10nm Medium Moderate Strength

capillaries surface permeability
tension shrinkage at
forces high
generated humidities

1 Pores 10-2 . 5nm Small (gel) Strong surface Shrinkage to

capillaries tension 50% RH
j
forces
generated

2 . 5-0 5nm
. Micropores Srongly absorped Shrinkage
water; creep
menisci form
< -0.5 nm Micropores Structural water Shrinkage;
"interlayer” involved in creep
bonding

11
 .

CAPILLARY POROSITY, percent

Figure 2. Effect of capillary porosity in hardened cement paste
on permeability (9)

12
o
I

E LU

CD O
< O
LU

tr
LU
a.

Figure 3. Effect of water-to-cement (w/c) ratio on

permeability of concrete (14)

13
 . .

in which cement hydration is essentially complete) can range from

10*^° to 10*^^ m/s (13) . The permeability of concrete is higher

than cement paste at the same w/c ratio for several reasons,

including the development of microcracks at the cement-aggregate

interface and of a low density transition zone in the cement

matrix near the interface, segregation of particles creating

bleeding channels, and lack of proper consolidation.

3.2 Controlling the Permeability of Concrete

Permeability of hardened cement paste can be reduced by
decreasing the porosity, maximum size of pores, and by reducing
the extent of inter-connectedness of pores. As illustrated in

Fig. 4 (13) , reducing the w/c ratio is an effective means of

decreasing the permeability of hardened cement paste. However,

significant decreases in the w/c ratio can result in difficulties

with, the placing and consolidation of concrete. These

difficulties can be overcome by the use of water reducers. Normal

water-reducing admixtures can reduce the w/c by around 15% (15)

If a significant reduction in w/c is desired (between 20 to 33%

for normal concrete (15)), then high-range water-reducing

admixtures (HRWRA) can be used. HRWRA are also known as

superplasticizers. In addition to making concrete more

**flowable, •*
HRWRAs can reduce the amount of bleeding. At a given
w/c ratio, HRWRAs do not appear to affect the pore size

distribution (16)

14
Another approach for reducing the permeability of concrete is to

add pozzolanic materials which react with the Ca( 0 H )2 produced by

the hydration of portland cement. This reaction produces more

calcium silicate hydrate. Silica fume is a particularly reactive

pozzolanic material because of its small particle (average

particle diameter of around 0.1 jum, which is about 1/100 of the

diameter of cement particles) and its high glassy silica content

(between 85 to 98 percent). Studies by Sellevold^ et al., (17)

and by Cheng-yl and Feldman (18) indicate that silica fume does

not significantly influence total porosity. It reduces the

volume of capillary pores, while increasing the gel porosity, and

significantly decreasing the volume of inter-connected pores. By

producing smaller and more discontinuous pores, the addition of

silica fume to concrete markedly decreases the permeability of

concrete (16) . Other pozzolans, which usually have finenesses

comparable to cement, do not demonstrate this fine pore effect.

Because of the high surface area of silica fume, the water

requirement for a flowable concrete is increased by silica fume

addition. The need for more water can be compensated for by the

addition of a HRWRA. The combination of HRWRA and silica

fumes have been used to produce concretes with w/c ratios of

0.24 and lower, and compressive strengths greater than 127 MPa

(18,400 psi) (20). Some uncertainty exists regarding the freeze-

thaw durability of such concretes if entrained air is not

present.

15
 .

O
X
(O H

LU

CC
LU
Q.

w/c RAi 10

Figure 4. Effect of water-to-cement (w/c) ratio on permeability

of hardened cement paste (19)

16
In addition to compositional and microstructural effects, other

factors can have a significant effect on permeability. Browne

and Domone (21) reviewed the permeability of concretes used in

the construction of submersible structures. They concluded that,

while permeability was a function of w/c ratio, aggregate

content, and curing conditions, the main variable influencing

permeability of concrete was the care with which it was placed.

Another factor which could significantly influence the
permeability of concrete is the occurrence of cracks. For

example, in hard crystalline rocks, the flow of water is

predominantly through long fissures, and, therefore, field

permeabilities can be much greater than permeabilities measured
in the laboratory on uncracked specimens. Alexander (10) showed

that for viscous flow in stone fissures of width b with average

spacing of s, the equivalent permeability (k^) is given by;

k^ = bVl2s. (3)

Presuming that this equation is valid for concrete, then cracks

with widths of 1 mm, spaced 1 m apart, would give an equivalent

permeability of 8 x 10*” m^ compared to an intact permeability of

around 10*^^ m^. The major types of cracks in concrete can be

divided into those occurring before hardening and those occurring

after hardening (Fig. 5) . Those occurring before hardening can

be caused by construction movement, plastic shrinkage or drying

shrinkage, and early frost damage. The cracks occurring after

17
 .

concrete has hardened involve physical, chemical, thermal and

structural processes. Cracks occurring before hardening and

most, if not all, of those occurring after hardening (especially

those caused by design loads, drying shrinkage, crazing, and

early thermal contraction) could be prevented by a proper QA/QC

program.

4. SULFATE ATTACK
Probably the most widespread and common attack of concrete in

contact with soil involves sulfates. Sulfate attack of concrete

can be very deleterious resulting in cracking of concrete and in

severe cases its disintegration. Naturally-occurring sulfates of

sodium, potassium, calcium, and magnesium are sometimes found in

groundwaters and soils, especially when high amounts of clay are

present. If the sulfates are in groundwater and concrete is

subjected to wetting and drying cycles by the groundwater, then

upon evaporation of the groundwater from concrete surfaces the

deposited sulfates may accumulate in the concrete at a

concentration higher than that in the groundwater. Such

processes can significantly accelerate sulfate attack, or also

the attack by other aggressive salts such as chlorides. Sulfate

attack has occurred in several regions of the United States, and

is a particular problem in arid regions, such as the Northern

Plains area and in the southwestern states (23) Localized

sources of sulfates in groundwater include mine tailings, fills

of blast-furnace slag, and deposits of chemical wastes.

18
 <

0)

_ 2®
® w
c
O ®.E
c-
— g®™
f-

® ®^
® E ®
g m £ £
®

c C
® o
E -
® Q E Q
o CO c c
o ® ® ® D ® c ® (22).

aggregates
0) -I—
O) ® E
o ® c E ®
b ® C6
®
^ ® c
11
o o o
S
® E D) ®
® E
® >
> o
o > o E
$ ® ® o ® TD
shrinkage
O vZ I-
® ® _ •D 1= E ®
® E
concrete
•4-i
CO
c D)
O) o
^
^
»_
® o o
^® ®® ^ -o
o CO ^® « -c c 0
®
Shrinkable
O'—®C
OT I

N cb *D O)
.9 .9 D)

O ®
in
Crazing t ® E ® ® >.
m 4-« w. 0 (O
w ® 1 A
Drying
o ^ ® il X ® O w. ® o o
o <o li. UJ UJ < OQ m Q_ Q_ UL CO
cracks

“ —

of

causes

__
u.
o O
Physical
E E o Major

0 ® o
.c j:
O H- CO
5.

Figure

O)
E
2 ® o =
<£ ® TD
£

% i2
g;oS
t- 5

19
 .

The water used in irrigation can be a potential source of sulfate

attack because of the gradual accumulation of sulfates as the

water evaporates.

4 . 1 Mechanisms
Sulfate attack of concrete is a complex process and at least

three major deleterious reactions can take place depending on the

environment (24,25)

The first process considered involves the reaction of sulfate

ions with calcium aluminate hydrate to form the calcium

sulfoaluminate product, ettringite (aCaO'AlgOj'aCaSO^'SlH^O) z

2 (3Ca0*Al203-12H20) + 3 (Na2SO^-10H2O) —> 3Ca0-Al203*3CaS0^“31H20 +

2A1 ( 0 H )3 + 6 NaOH + I 7 H 2 O (4)

Ettringite has a considerably larger volume than the reactants.

At low concentrations of sulfate ions, ettringite decomposes to a

low sulfate form 3 Ca 0 'Al 203 ‘CaS 0 ^‘ 12 H 20 . The amount of ettringite

accumulated and the stress induced, therefore, depends on the

availability of sulfate ions. The formation of ettringite

appears to be the cause of most of the expansion and

deterioration of concrete caused by sulfate solutions (23)

Another process involves the replacement of calcium hydroxide

(Ca( 0 H) 2)
in concrete by gypsum (CaS 0 ^° 2 H 20 ) ;

20
 . ;

Ca( 0 H )2 + Na 2 SO^- 10 H 2 O —> CaS 0 ^' 2 H 20 + 2NaOH + 8 H2 O (5)

Formation of gypsum can lead to the deterioration of concrete by

two processes (25) In one process, because gypsum occupies more

volume than calcium hydroxide, expansive stresses are produced.

In another, gypsum is gradually leached, leaving a porous

concrete with a higher permeability. In flowing water and given

sufficient time, calcium hydroxide may be essentially completely

converted to gypsum; while, in quasi-stagnant water, equilibrium

will be attained and only a portion of the calcium hydroxide will

be converted.

Similar reactions can occur involving magnesium sulfate. In

addition, magnesium sulfate can attack the calcium silicate

hydrate (
’
3 Ca 0 2 Si 02 'nH 20 ) formed by the hydration of portland

cement. The reaction is of the pattern (25)

’ ’
3 Ca 0 2 Si 02 3 H 20 + 3MgS0^ + 6 H 2 O —> 3 (CaS 0 ^- 2 H 20 ) + 3 Mg(OH )2

+ 2Si02 (6)

This reaction has particularly severe effects. Because of the

very low solubility of magnesium hydroxide, the equilibrium is

almost completely to the right. The reaction, however, is slow

and is of greatest concern when concrete is exposed to sea water.

21
 .

4.2 Resistance of Concrete to Sulfate Attack

Extensive research has been carried out on the resistance of

concretes to sulfate attack. For example. Miller and Manson (26)

carried out a 25-year investigation of the sulfate resistance of

concrete using more than 75,000 specimens made from 122

commercial cements. Also, the Portland Cement Association (27)

has been carrying out a comprehensive study of the resistance of

concretes to sulfate-containing soils. The results of these and

numerous other studies have clearly indicated that the tricalcium

aluminate content of portland cement greatly affected its sulfate

resistance (9, 23, 28). Accordingly, ASTM C 150 includes a Type

V (sulfate resistant) cement which sets a maximum of 5%, by mass,

on C3A and a Type II (moderately sulfate-resisting) for which the

C3 A is limited to 8 %. There is also evidence that the

aluminoferrite (C^AF) phase of portland cement may participate in

delayed sulfate attack and, thus, limits have been placed on its

contents. The traditional practice of relying on such cements

for acceptable sulfate resistance is being replaced with the

recent practice of using low-permeability, low Ca(OH )2 containing

concrete (28) The sulfate resistance of concrete can be

increased either by adding pozzolans or granulated blast-furnace

slag to the concrete, or by using blended cements containing a

pozzalan or a granulated blast furnace slag. These mineral

admixtures react with the Ca( 0 H )2 produced by the cement

hydration to produce additional calcium silicate hydrate which

reduces the capillary porosity and thus permeability. Also,

22
 . .

decreasing the Ca( 0 H )2 contents will reduce the amount of

concrete expansion caused by the formation of gypsum (equation

5) . The US Bureau of Reclamation (29) carried out a 12 year

study on the effect of the composition of fly ash on the sulfate

durability of concrete. It was found that the effectiveness of a

fly ash on improving the sulfate resistance of concrete could be

predicted based on its CaO and Fe 203 contents. Based on the ASTM

C1012 test method, silica fume has been observed to improved the

sulfate resistance of mortar bars (Hooten) . Long-term durability

tests on the sulfate resistance of concrete containing silica

fume, however, have not been reported. The effectiveness of

mineral admixtures in preventing sulfate attack depends on their

composition and physical properties, and, therefore, they should

conform to the appropriate ASTM standards before being used.

Recommendations for selecting cements and mineral admixtures for

concretes exposed to sulfates have been given in the ACI Guide

for Durable Concrete (23) The recommendations are presented in

table 2. Also, they are presented in an expert system form in

Durcon (30) Although ASTM Type V portland cement is recommended

in the table for severe and very severe exposures, little Type V

cement is available in the United States (31) . In the case of a

severe environment, a Type II portland cement with blast furnace

slag or a pozzolan should be an adequate substitute for Type V.

23
 ,

A site with a very severe environment is, undoubtedly, unfit for

storing LLW in concrete vaults. Following these recgmmendations
should result in a concrete with acceptable durability for many

years, possibly for the 60 to 100 year design life of typical

concrete structure. However, they do not provide a basis for

predicting the service life of a concrete exposed to a sulfate

environment, especially in the case of concrete with a design

life of hundreds of years.

4.3 Models and Service Life Predictions

Two major studies have been reported (32, 33) on approaches for

estimating the service life of concrete exposed to sulfates. In

the first, the US Bureau of Reclamation (32) continously

immersed concrete specimens in a 2.1% Na 2 SO^ (sodium sulfate)

solution until failure (expansion of 0.5%) or until the

investigation was completed. The age of specimens at the

completion of the continuous-immersion study was between 18 to 24

years. Companion specimens were subjected to an accelerated test

in which they were immersed for 16 hours in a 2.1% NajSO^

solution and then forced air dried at 54 'C (130*F) for 8 hours.

From a comparison of the times for specimens to reach an

expansion of 0.5% in both the accelerated test and the continuous

immersion test, it was estimated that 1 year of accelerated
testing equalled 8 years of continuous immersion. This was

considered to be a conservative ratio as a 1:10 ratio was

24
 Table 2. Recommenations for Sulfate Resistance (23)

Exposure Water Soluble cement Water-

soluble (SO^) in cement
sulfate (a) water ,
ppm ratio,
soil, percent maximum (b)

Mild 0 . 00 - 0.10 0-150 — *

Moderate 0 , 10 - 0,20 150-1500 Type II 0,50

(Includes IP (MS)
Seawater) IS (MS)
(c)
Severe 0 . 20 - 2.00 1500-10,000 Type V 0,45
(d)
Very Over 2.00 Over 10,000 Type V 0,45
Severe + Pozzolan
or Slag (e)

(a) Sulfate expressed as SO^ is related to sulfate expressed as SO 3

as in reports of chemical analysis of cement as SO 3 X 1.2 = SO^.

(b) A lower water-cement ratio may be necessary to prevent corrosion

;of embedded steel.

(c) Or a blend of Type I cement and a ground granulated blast-furnace

slag or pozzolan that has been determined by tests to give equivalent
isulfate resistance.

(d) Or a blend of Type II cement and a ground granulated blast-

furnace slag or a pozzolan that has been determined by tests to give
equivalent sulfate resistance,

(e) Use a pozzolan or slag that has been determined by tests to

improve sulfate resistance when used in concrete containing Type V cement.

25
thought to be more realistic. In many cases, the time for the
concretes in the continuous immersion test to expand by 0.5% was
estimated by straight-line extrapolation. They concluded that
most of the concretes with Types II and V portland cements showed

a life expectancy of less than 50 years when exposed to a 2.1%

NagSO^ solution. Some concretes containing certain fly ash and

ground blast-furnace slag mineral admixtures had estimated

service lives of 150 years or more. A 2.1% solution of Na2S0^ is

a very severe environment (table 2) and if concrete was exposed

to a lower level of sulfates the life expectancy would be

expected to increase. In addition to providing an approach for

service life estimates, the accelerated test results clearly

demonstrate the aggressive effect of wetting and drying cycles

with sulfate solutions on the durability of concrete.

The above method could be used for predicting the service lives

of in-service concretes, in continuous contact with groundwater,

if the ratio of 1:8 for the time to attain a equivalent expansion

in the accelerated and continuous immersion tests using the 2 . 1%

sulfate solution holds for other sulfate concentrations. Based

on the ratio of 1:8, however, to predict if a concrete would have

a life expectancy of 500 years may necessitate an unacceptably

extended extrapolation of the accelerated test results as

follows. Assuming that a concrete with a life of 500 years was

actually being tested, then to predict that the concrete would

have a life expectancy of 500 years by testing until the 0.5%

26
 . .

expansion criterion was reached in the accelerated test would

require a testing period of at least 500/8 years, i.e., 63 years.

Unfortunately, few durability tests are carried out for 63 years

and a 5 to 10 year test is often considered long-termed. If the

accelerated test was terminated in 5 years, the expansion for the

500-year-life concrete would be no more than 0.04% and little

confidence could be given to predictions based on extrapolating

the expansion data to 0.5%.

Another approach to predicting the service life of concrete

exposed to groundwater containing sulfate salts was developed by

the Building Research Establishment in England (33) In the

laboratory, concrete specimens were immersed in a 0.19M sulfate

solution (a mixture of alkali and magnesium sulfates) for up to 5

years. The accelerated laboratory tests resulted in a visible

deterioration zone, X^, and the following empirical equation was

developed:

X^Ccm) = 0.55CA(%)-([Mg] + [SOJ)-t(y) (7)

where CA% is the percentage by weight of tricalcium aluminate

( 3 Ca 0 'Al 203 ) in the cement, [Mg] and [SO^] are the molar

concentrations of magnesium and sulfates, respectively, in the

test solution, and t(y) is the test time in years. The

variability in the depths of attack were around 30% of the

average

27
 ,

Equation no. 7 was found to give satisfactory correlations with

the results of field tests (3) in which the depth of

penetrations were in the range of 0.8 to 2 cm after 5 years. The

equation was used to calculate a range of lifetimes of concrete

exposed to groundwater of a known concentration of sulfate salts.

Concretes made with ordinary portland cements containing between

5 and 12% C^A, gave estimated lifetimes of between 180 to 800

years, with a probable lifetime of 400 years. When a sulfate

resisting portland cement with 1 . 2% C3A was used, the minimum and

probable lifetimes were estimated to be 700 years and 2500 years,

respectively. These times were estimated based on the loss of

one-half of the load-bearing capacity of a 1 m thick concrete

section, i.e., of 50 cm. The estimates involve an

extrapolation based only on an empirical equation that has been

shown to be reasonably predictive only during short-term testing.

Atkinson, et al., (3) attempted to verify the equation by

determining the extent of deterioration of concretes buried and

exposed to the groundwater of a clay for 43 years. An alteration

zone of about 1 cm was observed in the concretes which could be

caused by several processes. Based on the tricalcium aluminate

contents of the cements, equation no. 7 predicts that the

thickness of the deteriorated region should be between 1 and 9

cm. Therefore, Atkinson, et al., (3) concluded that the

equation either gives a correct estimate or an overestimate of

the rate of sulfate attack.

28
5. CORROSION OF STEEL REINFORCEMENT IN CONCRETE
Portland cement concrete normally provides an internal

environment which protects reinforcing steel from corrosion. The

high alkaline environment (pH > 12.5) in concrete results in the

formation of a tightly adhering film (gamma iron (III) oxide)

which passivates the steel and thereby protects it from

corrosion. However, the corrosion of steel embedded in concrete

has become a serious problem in recent years in certain types of

structures, e.g., bridges and parking garages. Chloride ions are

considered to be the major cause of premature corrosion of steel

reinforcement in many of these structures. Chloride ions are

common in nature and small amounts are usually unintentionally

contained in the mix ingredients of concrete. Chloride ions also

may be intentionally added, most often as a constituent of

accelerating admixtures. The sources of chloride ion which are

causing the most problems, however, are deicing salts (usually

calcium chloride) and sea water.

In addition to natural deposits of chloride salts, groundwater

may be contaminated with chloride ions coming from runoff water,

e.g., from bridges or pavements treated with deicing salts.

Corrosion can occur in some cases in the absence of chloride

ions. For example, carbonation of concrete results in reduction

of its alkalinity and, thereby, can depassivate the steel and

initiate corrosion. Carbonation, however, is usually not a

29
 . ,

problem with dense concrete and with a proper cover of concrete

over the reinforcement. Corrosion also can occur in concrete

through the formation of concentration cells (34,35).

Because of the predominant role that chloride ions plays in

corrosion of reinforcing steel, their effects on the durability

of concrete will be largely addressed herein. Also, protecting

reinforced concrete from chloride ions should also improve its

durability towards other corrosive substances.

5,1 Mechanism
The corrosion of steel in concrete is almost always

electrochemical in nature. Its process is similar to the action

occurring in a battery, involving the formation of cathodic and

anodic poles, an electrical conductor (the reinforcing bar) , and

an aqueous electrolyte (the pore solution in concrete) (Fig 6)

Reactions at the anodes and cathodes are usually referred to as

"half cell reactions" (36) At the anode, iron is oxidized to

iron (II) (ferrous) ions:

Fe Fe^"" + 2e'- (8)

At the cathode, reduction takes place. In the alkaline

environment of concrete the cathodic reaction is:

H 2 O + I/ 2 O 2 + 2e* 20H“° (9)

The overall (cell) reaction is:

30
 reinforcement

steel

of

corrosion

of

Schematic

6.

Figure

31
 . . .

Fe + I/ 2 O 2 + H 2 O Fe(0H)2. (10)

Other subsequent reactions occur with Fe(OH )3 being formed. It

precipitates as FeO'OH which gives the characteristic rust color.

Equation (9) indicates that both water and oxygen must be present

for the corrosion to occur. With only water and oxygen present,
however, the reaction proceeds very slowly and with good quality

concrete the resulting rate of corrosion may be very small under

normal environmental conditions (37) Chlorides can accelerate

the corrosion rate so that corrosion can be a serious problem

within a few years. For example, bridge decks which are treated

with chloride-ion-containing deicing salts have been known to

deteriorate a few years after construction because of corrosion
of reinforcing steel (36) The process by which chloride ions

cause the accelerated corrosion of reinforcing steel is not

unequivocally known, but chloride ions are thought to depassivate

steel by dissolution of or by a physical breakup of the

protective oxide layer on the steel (36) Chloride ions also can

reduce the electrical resistance of concrete which can increase

the corrosion current.

The corrosion products of reinforcing steel have a volume

several times (36, 37) that of the original iron, which can

result in the development of internal tensile stresses leading to

the cracking of the concrete. If corrosion is allowed to proceed

32
 .

unabated, it can ultimately result in the destruction of

reinforced concrete.

5.2 Resistance of Reinforced Concrete.

The resistance of reinforced concrete to corrosion is dependent

on several important factors including the environment, the

permeability of the concrete, the concrete constituents, the

depth of cover of concrete over the reinforcement, and the
quality of the concrete.

5.2.1 Environment

It is anticipated that concrete vaults buried in soils will not

be exposed to seawater nor will deicing salts be directly applied

to them. However if such severe environments are encountered,

the use of special protective measures must be considered

including the use of coatings or sealers for concrete, dense

concrete overlays, metallic and organic coatings for reinforcing

steel, corrosion inhibitors, and cathodic protection. These

practices are described in the report prepared by ACI Committee

222, Corrosion of Metals in Concrete (36)

The severity of the corrosive environment will be largely

dependent on the soluble content of the soil and groundwater.

Similar to the case with sulfate ions, wetting and drying cycles

with aqueous solutions of chloride ions can result in the

concentration of chloride ions deposited in the pores of the

33
 . . . .

concrete being much higher than the concentration in the

groundwater

By this process, the corrosion of reinforcing steel can be

significantly accelerated. This is the basis for an accelerated

corrosion test for reinforced concrete developed by the Federal

Highway Administration (38)

5.2.2 Reinforcing Steel

In dealing with severe corrosive environments, changes in the

metallurgy of reinforcing steel have been considered. However,

the susceptibility to corrosion of typical reinforcing steel

(mild steel) is not thought to be significantly affected by its

composition, grade, or level of stress (39) Stainless steel

reinforcement has been used in special applications, especially

as hardware for attaching panels in precast concrete

construction. Stainless steel, however, is too expensive as a

replacement for typical reinforcing steel in most cases.

Stainless-steel-clad bars have been found to reduce the frequency

of corrosion- induced cracking compared to mild steel, but did not

prevent it (36) Fusion-bonded epoxy powder coatings have been

found to be effective in protecting steel reinforcement from

corrosion (40) and they have been widely used in bridge decks.

34
5.2.3 Alkalinity of Pore Solution
5.2.3. 1 Effect of pH on Corrosion Threshold

The chemistry of the pore solution in concrete was shown by

Hausman (41) to influence the initiation of corrosion by chloride

ions. In studies carried out in sodium hydroxide solutions and

saturated calcium hydroxide solutions, he found evidence for a

relationship existing between the chloride ion concentration

necessary to induce corrosion (threshold concentration) and the

activity of the hydroxide; with the threshold level significantly

decreased as the pH is lowered. For example, at a pH of 13.2

more than 8000 ppm of chlorides ions were required to induce

corrosion, while at a pH of 11.6 around 71 ppm of chloride ions

were sufficient to induce corrosion.

5. 2. 3. 2 Pozzolans
The effect of pH on the threshold concentration of chloride ions

poses an interesting dilemma in deciding whether to use a

pozzolan. Pozzolans react with the hydroxide ions resulting from

the hydration of portland cement and thus can reduce the pH of

the pore solution. Thus it would appear that the use of

pozzalans would decrease the resistance of reinforced concrete to

corrosion. The effect of pozzolans, however, on reducing the

permeability of concrete has been reported (42) to more than

compensate for the lowering of pH. Also, the amount of pozzalan

added can be adjusted so that sufficient amounts of hydroxide

ions remain to maintain a high pH. Concrete usually has a high

35
 . . :

reserve basicity because most of the calcium hydroxide is not in

solution which can be dissolved as a pozzolan reacts with the

pore solution. While, it seems likely that concretes with

pozzolans should adequately protect reinforcing steel from

corrosion, this assumption has not been verified by long-term

studies

5.2 . 3 . 3 Carbonation
Carbonation of concrete by carbon dioxide reduces the pH of the
pore solution in the affected regions. If the depth of

carbonation extends to the reinforcing steel, then the chloride

ion threshold concentration can be significantly reduced.

Carbonation follows the general equation (43)

dx/dt = Kt^^^ (11)

in which the proportionality constant K depends on several

parameters relevent to the quality of concrete and to the

environment.

The rate and extent of carbonation depends on the environmental

relative humidity, reaching a maximum at 50% relative humidity

(43) Diffusion of gaseous carbon dioxide takes place several

orders of magnitude more rapidly though air than through water.

If the pores of concrete are saturated with water, the amount of

carbonation occurring will be negligible. In most cases, the time

36
 .

needed to carbonate 2 cm of high-quality concrete is reported

(44,45) to be of the order of tens of years, and the penetration

rate of carbon dioxide rapidly falls well below a mm/year soon

after construction. Carbonation can be a serious problem,

however, if the reinforcing steel is closer to the concrete

surface than recommended, e.g. in ACI 318 (46) . High

temperatures, such as those which occur in Middle East countries,

appear to increase the severity of corrosion attributed to

carbonation (43, 47, 48)

Another process by which the pH of concrete can be reduced is by

the leaching of soluble alkaline salts with water, especially if

the water contains high levels of dissolved carbon dioxide. This

process is discussed in Section 9 on groundwater.

5.2.4 Domestic Chloride Ions

In some cases sufficient chloride ions have been present in the

constituents (called "domestic" chloride ions (36)) of concrete

to cause corrosion. For example, aggregates may have entrapped

sea water during their formation which could be leached out by

the concrete mix water. Also, certain chemical admixtures,

especially set accelerators containing calcium chloride, can

contain corrosive amounts of chloride ions. Recognizing that

domestic chloride ions can induce corrosion, ACI Committee 222,

Corrosion of Metals in Concrete, recommends (36) the following

37
maximum acid-soluble chloride ions, expressed as percent by mass
of the cement, be permitted in reinforced concrete:

Prestressed concrete - 0.08%

Reinforced concrete - 0.20%

Domestic chloride ions have been found to be tied up by the

tricalcium aluminate of Portland cement to form calcium

chloroaluminates (39) . While using cements with high tricalcium
aluminate contents may be of benefit in reducing corrosion
problems, the potential for sulfate attack will be increased.

5.2.5 Concrete Cover and Quality

The results of numerous studies (37,38,39) have indicated that a

proper concrete cover over the reinforcing steel is vital in

preventing corrosion. Combination of a low w/c ratio and the

recommended cover (e.g., as given in ACI 318 (46)) can

significantly improve the resistance of reinforced concrete to

corrosion caused by chloride ions. The effect of w/c ratio and

the depth of concrete cover on the ingress of chloride ions is

shown in Fig. 7.

Good concrete construction practices are necessary for obtaining

corrosion-resistant reinforced concrete. Poor consolidation,

excessive bleeding, and presence of honeycombing can result in

non-uniform concrete with higher than anticipated permeability.

38
 ‘3ovdans M0i3a Hidaa
o o
in
CM in
in
N-
o
in
14.0 CO
d

12.0
o 0 •

CO 4-)

0 CXD

d ro
o —
S-l

c
/yd'^
0 C
o 0
c •H
<D 4-1 X
10.0 in O 0 fO

Cl i_ S-l
CM
0 X X
d CL X 0
CLc
lbs
0 0
T3 CL

T3 c
C 0
•H
8.0 o (T3

o 0
CM
d O
•H 'O
CONTENT,
4J -H
o 0 U
Sh O
r-H
LU -u x;

9 c u
0
6.0 in c e c
o 0 0
0
CHLORIDE
d X 1

0 c
4J

4J 0
o 1

L 0
g
4

LU 0 u
+J S-l
CD o
4.0 o < (d
5 4-(
c
tr
d LU 4-i
0 0
-H

AVERAGE

$ 4-)

U
0 0
4-1 >
4-4 O
in W O
q
d 1^

0
S-l

3
CO
•H
fa

q q q
CM cd

u! Monaa Hidaa
‘

3 ovdans

39
 .

as well as resulting in the development of concentration

corrosion cells. Also, proper curing is important for the

attainment of low permeability.

5.2.6 Stray Currents

Stray currents can be a potential problem with reinforced

concrete structures buried in the ground. Stray currents can be

defined as currents that follows paths other than the intended

circuit (36) .. They can greatly accelerate the corrosion of

reinforcing steel. The most common sources of stray currents are

electrical railways, electroplating plants, and cathodic

protection systems. Cathodic protection systems are often used

to protect underground metal pipes transporting petroleum

products and natural gas.

5.3 Models and Service Life Predictions.

Three models which can be used for predicting the service life of

reinforced concrete exposed to chloride ions were disclosed

during the course of the review.

5.3.1 Model by Tuutti

Tuutti (37) has developed a model for predicting the service life

of reinforcing steel which will be outlined in some detail

because it appears to provide a means for predicting the life of

concrete structures buried in the ground. His model is based on

the corrosion sequence schematically shown in Fig. 8. In this

40
 o
E
H

steel

of

corrosion

of

model

conceptual

of

Schematic

8.

Fiqure

41
 : ,

model, corrosion starts after the end of an initiation period

(dominant period of no corrosion) which is followed by a

propagation period (period of active corrosion) . The corrosion

process is initiated by the diffusion of chloride ions to the

depth of the reinforcing steel or by carbonation reducing the pH

of the pore liquid in contact with the steel, or by the

combination of chloride ions and carbonation.

In the following, only the effect of chloride ions on the

initiation period will be considered. The length of the

initiation period is largely controlled by the rate of diffusion

of the chloride ions in the concrete and by the threshold

concentration for the process. Because chloride ions may react

with the tricalcium aluminate of portland cement the

concentration has two components; concentration of bound chloride

ions (c,j) and concentration of free ions (c^) , related by the

constant R (37)

Cb = R'Cf (12)

The diffusion equation (D) in one dimension, can be expressed

as;

= [D/(R + 1) ] (13)

42
 ;

where D/(R + 1) = the effective diffusion coefficient, x is

distance, and t is time.

In addition to the pH of the pore solution, the threshold

concentration of chloride ions depends on the moisture content

and the potential of the steel. The chloride ion threshold for

water-saturated concrete should be in the range of 5,000 to

11,000 ppm. (49). The effects of concrete cover, threshold

concentration of chloride ions, concentration of chloride ions,

and concrete permeability on the calculated induction period are

shown in Fig 9.

The corrosion rate in the propagation period is controlled by the

rate of oxygen diffusion to the cathode, resistivity of the pore

solution, and temperature. The diffusion resistance to oxygen is

determined by the water saturation of the concrete cover and by

the cross-section of pores not filled with water, which are

dependent on the concrete's porosity, pore size distribution,

inter-connectiveness of pores, and ambient relative humidity.

The combined mass transfer coefficient, K^, is given by (37)

Kt = + (K„j(,)/m (14)

= (8-10'Vx,i,) + 10'V4-10*'X„2o.

43
 concentration

induction

threshold

ions,
cover,
calculated

chloride

concrete on

of
of

permeability

Effects

concentration

concrete

period.

threshold and

induction
concentration

ions,

)
37

(
chloride

concentration,

Calculated

period

Surface
of

9.

Figure

44
 .

where m is a temperature dependent constant and X is the distance

in either a air or water filled pore or pore network. Clearly,

the migration of oxygen through saturated concrete will be a

substantially slower process than through partially water filled

concrete and can become the rate controlling process. The

effects of water in the pores on corrosion is demonstrated by the

observation that corrosion at a relative humidity of 100% is

negligible (50)

In using the model for predicting service lives, the recommended

procedure by Tuutti (37) is given by the following steps:

Step 1. Compile the relevant material, structural, and

environmental data.

Step 2 . Calculate initiation time, based on experimental

data for the diffusion rates and threshold

concentration of corrosive substances.

Step 3 . Calculate the corrosion rate in the propagation stage.

Step 4 . Determine the amount of steel corrosion required to

cause failure.

Tuutti (37) has provided values for many of the model parameter

which can be used in estimating the service life of reinforced

concrete exposed to corrosive environments. Also, he has given

45
recommendations on methods for obtaining values for the other

parameters. The model appears to provide a means for making

semi-quantitative predictions of the minimum service life of

concrete structures.

5.3.2 Model for Corrosion by Sea Water

A model somewhat similar to that of Tuutti (37) was developed by

Bazant (51) for corrosion of reinforced concrete exposed to sea

water. This model is comprehensive and mathematically describes'

the following processes:

(1) diffusion of oxygen and chloride ions in pore water

(2) diffusion of iron (II) hydroxide near steel surface^

(3) depassivation of steel due to chloride ions,

(4) effects of oxygen and iron (II) hydroxide on cathodic

and anodic potentials and their polarizations,

(5) flow of electrical current through the electrolyte in

pores,

(6) mass sinks for oxygen, iron (II) hydroxide, and hydrated

red rust , and

(7) rust production rates based on reaction kinetics.

The complete equation system is quite complicated and its

complete solution has not been accomplished. Nevertheless, the

model is based on a logical approach for developing, from theory,

a method for calculating the service life of reinforced concrete.

Based on some simplifying assumptions, the use of the model to

46
predict the service life of concrete structures in sea water has

been demonstrated. For example, it predicts that a deeply

submerged reinforced concrete structure could have a service life

(considering only corrosion) of some 15,000 years (52). The

model, perhaps, could be modified to address the case of

reinforced concrete buried in soil. However, in the present form

it is not obvious that this model has any advantages over that

developed by Tuutti.

5.3.3 Empirical Model

Browne (53) developed an empirically based method for predicting

the remaining service life of in-service reinforced concrete

structures. The procedure for making predictions is:

(1) Samples are obtained from a concrete structure at different

depths from the concrete surface and their chloride contents are

determined.

(2) Use the equation

c(x,t) = c^[l - erf (x/2Dj.i-t)'^ (15)

(where c(x,t) is the chloride concentration at depth x after time

t, for an equilibrium chloride concentration of c^ at the

surface; is the chloride ion diffusion coefficient; and erf

47
is the error function (obtained from standard statistical

tables)), to obtain values of c^ and

(3) Once the values of c^ and are obtained, then the chloride

ion concentration at any distance from the surface can be

calculated using equation no, 15.

(4) A chloride ion concentration of 0.4%, based on weight of

cement, is used by Browne as the threshold value. The time to

reach the threshold concentration at the depth of the reinforcing

steel gives the remaining service life.

If the value of can be estimated for a concrete, based on its

porosity features, then equation no. 15 could be used to predict

the time of the initiation period in the model of Tuutti (37)

6. ALKALI -AGGREGATE REACTIONS

Alkali-aggregate reactions are different from those previously

discussed in that the reactive components are usually internally
contained in concrete and, therefore, are not dependent on the

diffusion of an aggressive solution into the concrete. For

appreciable amounts of swelling to occur, however, a source of
water is required.

48
 .

It appears that all aggregates react to some degree with alkalies

in cement. In the majority of cases the reactions are beneficial

(e.g., increasing the bond between aggregate and hydrated cement

paste) or innocuous. In some cases, however, the reactions

result in the formation of expansive products which can cause

serious cracking of the concrete. The expansion is attributed to

the imbibition of water by the reaction products. When cracking

occurs its effects can range from being cosmetic (in most cases)

to resulting in structural failure (uncommon) of concrete. As

discussed in section 3 , cracking of concrete even though not

causing structural collapse can significantly increase its

permeability. Alkali-aggregate reactions may proceed rapidly

(several months to a few years) or much slower (e.g., early sign

observed only after some 20 to 25 years (54,55,56)). Expansive

alkali-aggregate reactions are known to occur with certain

siligeous aggregates and certain dolomitic limestone aggregates.

Alkali-silica is the most common and well studied reaction and

has been found to occur in numerous places in the United States

(9) with the greatest abundance of alkali-silica rocks being

found in the western states (23) . Slow reacting siliceous

aggregates have been found in the eastern states. Alkali-
carbonate reactions have occurred in some midwestern and eastern

states (23)

Cracking of concrete occurs because of the internal development

of localized expansive stresses. If the reactive particles are

49
numerous or of a fine size, more uniformly distributed expansive

forces may developed which may result in uniform expansion rather

than cracking.

Problems with alkali-aggregate reactions may grow in the future

because of large increases in the amount of cement being

imported. The alkali contents of the imported cement may not be

tightly controlled. Also, in some locations the supply of high

quality aggregates has become scarce.

6.1 Alkali-Aggregate Reaction

6.1.1 Mechanism. While the reaction is not thoroughly understood

its general mechanistic path appears to be:

(1) Cleavage of silica-oxygen bonds of the silicate network

(depolymerization) by hydroxide ions and the dissolution of the

depolymerized species. The dissolution of the silica is a

function of the pH of the pore solution. A small amount of

silica dissolution occurs at the pH (12.5) of a saturated

solution of calcium hydroxide but increases by over a thousand-

fold at a pH above 13 (13) . Sodium and potassium hydroxides are

very soluble and the pH of the pore solution of high alkali

cements could be as high as 13.6 (57) which would greatly

increase the amount of silica dissolved.

 .

(2) Formation of a hydrous alkali silicate gel. This gel is

thought by some researchers to act as an semi-permeable membrane

(43) which promotes the imbibition of water, leading to its

swelling. The swelling further destroys the rigidity of the

aggregate and permits the deeper diffusion of hydroxide ions into

the aggregate.

(3) As the reaction proceeds, the amount of water imbibed

increases as a dilute suspension of colloidal particles is formed

which can move from the aggregate into the surrounding matrix.

If sufficient amount of expansive products are formed, the

resulting internal pressures can cause cracking of the concrete.

The extent of the expansion will be limited by the availability

of water. However, even in arid regions the internal relative

humidity of concrete may be sufficiently high so that reaction

products are formed around aggregates (43) If the concrete is

sometime in the future exposed to moisture, expansion can occur

rapidly.

In addition to the amount of alkalies in concrete and the

availability of water, the rate of expansive reaction is also

influenced by the properties of the aggregate. Concrete

aggregate which are susceptible to expansive alkali reactions are

often polycrystal linic and polymineralitic. The reactive mineral

may be present in small amounts and its distribution may vary

51
from aggregate to aggregate. The permeability as well as the

size of aggregates can affect the rate of reaction.

6.1.2 Determination of Aggregate Reactivity

Decisions on the potential alkali-silica reactivity of an

aggregate are sometimes based on its service records performance.

If service records are not available or they indicate that the

aggregate may be reactive, laboratory examination and test need

to be performed.

The source of aggregates have changed over the years and it is

doubtful that sufficient long-term service records exist to form

a base for predicting that silica-containing aggregates will be

free from expansive reactions for some 500 years.

Laboratory investigation of the potential reactivity of silica-

containing aggregates usually consist of petrographic

examination, mortar bar test, and chemical test. These methods

are briefly analyzed in the following to determine if the

individual methods or their combination can be used in predicting

the potential expansion of concretes made with silica-containing

aggregates. The ACI Guide on Durable Concrete (23) gives a more

complete discussion and cites numerous references to detailed

investigations

52
 .

6. 1.2.1 Petrographic Examination. ASTM C 295 (58) provides a

standard practice for the petrographic examination of aggregates.

It can give information on the potential reactivity of an

aggregate based on identification of the minerals present. If

present in detectable quantities, presence of the well-known

reactive minerals of aggregates may be fairly easily identified

by an experienced petrographer . However, some minerals can be

present in small quantities which escape detection but which are

sufficient to result in concrete cracking.

6. 1.2. 2 Mortar-bar Test for Potential Reactivity. The ASTM C

227 (59) mortar bar test is the test most generally relied upon

to indicate potential alkali reactivity. The test can require up

to six months to complete. However, certain metamorphic

siliceous rock could require longer test times (possibly up to 3

years) and elevated temperatures to give evidence of expansion

(60) If carried out for a sufficient testing period, the test

is usually reliable in predicting if an aggregate is potentially

reactive. However, studies by Oberholster and Davies (61)

suggest that the test does not always give a reliable estimate of

the extent of expansive reaction that could occur in field

concrete with a potentially reactive aggregate. It is not

surprising that the test would not reliably predict field

performance because aggregates from even a single source can vary

in the type, amount, and distribution of reactive minerals, the

degree of crystallinity, size of crystal grains, and the pore

53
characteristic. Also, the test conditions do not appear to

reasonably simulate the range of conditions encountered by field

concrete

6. 1.2. 3 Chemical Test for Potential Reactivity. This test (ASTM

C 289 (62)) is used primarily for a quick evaluation of the

potential reactivity of an aggregate. The method measures the

amount of silica dissolved in a concentrated sodium hydroxide

solution and the reduction in alkalinity. The data is then

correlated with aggregate reactivity using standard curves.

Results can be obtained in a few days. The results usually

correlate with aggregate behavior in the mortar bar expansion

test, but it is not suitable for predicting the degree of

expansion occurring in concrete (43) The chemical test can be

considered to be an aggregate reactivity acceptance test (43)

6. 1.2. 4 Combination of the Methods. The U.S. Corps of Engineers'

Engineering Manual (63) concludes "that none of the available

means of evaluation can be relied upon independently, or

collectively, to provide an unquestionably positive answer to the

problem of potentially harmful reactivity of an aggregate whose

constituents include materials capable of reacting with the

alkalies in Portland cement." The results of the petrographic

examination with corroborating evidence from the mortar-bar test

are considered to be the most reliable indicators of potential

reactivity. It is further stated that "Spurious results are

54
 .

possible with the chemical test when certain substances are

present in the aggregate being tested." Thus the chemical test

should be used in conjunction with the petrographic examination.

Based on the analyses of the predictive capability of the above

standard test methods, it appears they do not provide a basis for

predicting the long-term expansive characteristics of aggregate

containing silica minerals.

6.1.3 Prevention of Alkali-Aggregate Expansive Reactions

Methods for preventing alkali-aggregate expansive reactions are
given in detail in the ACI Guide to Durable Concrete (23) They

will be briefly discussed herein to indicate the approach for

dealing with the problem. Obviously, the best method for

preventing the reaction is not to use a reactive aggregate.

However, the identification of all reactive aggregates is not, at

present, possible. Also, in some cases the only local aggregates

which are available are potentially reactive. Another obvious

method is to keep concrete dry, which may be virtually impossible

to accomplish for 500 years when the concrete is buried in the

ground. The use of coatings for concrete to reduce the ingress

of moisture is discussed in Appendix A.

Since the alkali-aggregate reaction is sensitive to the pH of the

pore solution, designing a low-alkali containing concrete is an

effective measure. Often the alkali content is kept low by the

55
use of "low-alkali" cement (maximum of 0.60% equivalent sodium

oxide according to ASTM C 150 (64)). A higher alkali containing

cement can be used if the proportions of the concrete

constituents are adjusted to give a low alkali content in the

concrete. External sources of alkalies must be considered. For

example, penetrating groundwater containing large amounts of

alkali sulfates could react with calcium hydroxide forming alkali

hydroxides. Then, because of the higher solubility of the alkali

hydroxides compared to calcium hydroxide, the alkali content of

buried concrete will increase. In a few cases (65,66) aggregate

has contained appreciable amounts of alkalies and alkali-

aggregate reactions occurred even though a low-alkali cement was

used.

Many pozzolans have been found effective in reducing expansion

due to alkali-aggregate reactions to an acceptable level.

Several mechanisms have been proposed to explain the effects of

pozzolans including (43)

(1) because of their fines, pozzolans react with alkaline

hydroxides more rapidly than aggregates and reduce the alkalinity
of the pore solution. Also, alkali-pozzolan reactions do not

yield deleterious expansive products.

(2) reduced permeability in the hydrated cement pastes.

56
 . . . .

(3) partial replacement of a portland cement with a pozzolan

containing a smaller proportion of alkalies will result in a

reduction of the concrete's alkali content.

Certain fly ashes, silica fumes, and ground blast furnace slags

have been found to reduce the expansive reactions. Their

effectiveness does depend on their alkali contents. For example,

it has been observed that use of a high alkali containing silica

fume material increased the mortar-bar expansion compared to the

control Portland cement (67)

6.2 Alkali-Carbonate Reaction

6.2.1 Mechanism. Although the reaction mechanism is not well
understood, it is thought that the reactive carbonate rocks are

dolomitic limestone of the general formula, MgCOj-CaCOj and

contain clay (23) In their reaction with alkali hydroxides,

dedolomization occurs (i.e., magnesium carbonate, MgCOj, is

liberated) and magnesium hydroxide (called brucite) is formed

(ACI 23) The reaction appears to break down the structure of

the rock permitting the ingress of water. Then the absorption of

water by the clay minerals is thought to take place, which

becomes one of the causes of the swelling characteristics of the

reaction (68) The affected concrete is most severely cracked in

areas where an abundant supply of moisture is available.

57
6.2.2 Determination of Potential Reactivity of Aggregates

Potentially reactive dolomitic limestones are identified by

petrographic examination (ASTM 295 (58)), and the rock cylinder

test (ASTM C 586 (69)). In addition, a test method is being

developed by ASTM to measure the length change of concrete due to

alkali-carbonate rock reaction and a Canadian standard, CSA-
A23.2-14A, is available (23).

Definite correlations between laboratory tests and field

performance need to be established. Apparently, while the

laboratory tests can be useful in identifying potentially

reactive aggregates, they do not provide a basis for predicting

the rate or extent of expansion in field concrete.

6.2.3 Prevention of Alkali-Carbonate Reaction

ACI Guide to Durable Concrete (23) recommends that the best way

to avoid alkali-carbonate reactions is not to use such

aggregates. Other ways are to dilute reactive aggregates with

nonreactive aggregates (70,71) or to use a cement with equivalent

alkali content below 0.40% (23). Unlike with alkali-silica
reactions, pozzolans are not effective in mitigating alkali-

carbonate reactions except for diluting the alkali content of the

concrete

58
 .

6.3 Models and Service Life Predictions

Neither a model which mathematically describes the mechanism of
alkali-aggregate reactions nor a means for predicting the service
life of concrete with potentially reactive aggregates was found

in the literature review. Correlations between accelerated

laboratory expansion or chemical reactivity tests and field

expansions have not been adequately established. Thus there is

not a sound basis for predicting the rate or the amount of

expansion to be expected using a potentially reactive aggregate

in field concrete.

The problem, therefore, cannot be treated as one of kinetics

(e.g. , reaction rate) but must be treated as a thermodynamic

problem. That is, will an aggregate be thermodynamically stable

for 500 years in the presence of alkali hydroxides at the

concentrations to be expected for concrete buried in the ground?

The answer will provide a basis for predicting the service life

in the sense that concrete made with a thermodynamically non-

reactive aggregate will not experience alkali-aggregate expansive

reactions. However, it will not provide a basis for predicting

the service life of concrete made with a potentially reactive

aggregate

A test for the reactivity of siliceous aggregates could involve

determination of their chemical reactivity with alkali

hydroxides. The mechanism of alkali-silica reactions involves

59
the dissolution of silica and, if a chemical reaction takes place

between a siliceous aggregate and alkalies, then, in the absence

of calcium ions, the silica concentration of the aqueous solution

should increase (57) Determinations of the dissolved silica

concentration, therefore, could be use to indicate if the

aggregate is thermodynamically stable in the presence of

alkalies. The testing procedures would need to be worked out.

Also, such questions as what temperature should be used, should

the aggregate be reduced to a fine size to expose possible

reactive constituents, and what is an appropriate testing time,

would need to be answered. Another approach would be to prepare

concrete with a higher alkali content than that of the concrete

to be used in actual construction, expose it to elevated

temperatures and moisture and, after a specified testing time,

microscopically examine the aggregates for the development of any
reaction products. The presence of reaction products would

indicate that the aggregate is reactive. Because almost all

aggregates containing silicious minerals may show some reaction

in this test, aggregates of concrete which have been in service

for many years without expansion occurring due to alkali-

aggregate reactions could be subjected to the test and the extent

of reaction determined. This would prove baseline data for

evaluating the results of testing potentially reactive

aggregates

60
 .

7. FREEZING AND THAWING

Freezing and thawing damage (also called frost damage) occurs

when a damp concrete is exposed to cycles of freezing and

thawing. The concrete does not need to be fully saturated with

water for damage to occur as the critical level of saturation is

around 85 percent for most concretes (72) Freezing and thawing

damage is a serious problem for exposed concrete in the northern

states, resulting in its cracking and reduced service life.

Horizontal concrete surfaces such as pavements, bridge decks,

roofs, sills, and other surfaces where water can remain in

contact with the concrete for sometime are especially susceptible

to freezing and thawing damage. Several precautions must be

taken to avoid freezing and thawing damage, including the

incorporation of entrained air in concrete, and proper placement,

consolidation, and curing of the concrete.

The concrete structure storing low-level radioactive wastes will

be safe from cyclic freezing when they are covered with soil as

they will be located below the freeze line. Prior to being

covered, however, they could be susceptible to cyclic freezing

damage as follows. Immature air-entrained concrete can be

damaged if it has not developed adequate strength before being

exposed to freezing conditions. Also, if air-entrained concrete

is kept in a state of nearly complete saturation so that the

entrained air bubbles are extensively filled with water, it may

suffer damage from cyclic freezing. This could happened to the

61
LLW storage structures if water is allowed to pond against them

during the construction phase or while they are being loaded with
LLW.

Because a potential for freezing and thawing damage does exist

for the LLW concrete storage vaults before they are buried, the

mechanisms of frost action, factors to be considered in making

durable concrete, and applicable models are reviewed.

7 . 1 Mechanisms
Both the hardened cement paste and the aggregates in concrete can

undergo frost damage. The hardened cement paste can be made more

resilient to frost damage while little can be done to improve the

resistance of an aggregate except to select another aggregate.

Therefore, the mechanisms in the hardened cement paste will be

discussed in more detail than that for aggregates.

7.1.1 Hardened Cement Paste

Several mechanisms of the freezing process in hardened cement

paste have been proposed and the widely accepted mechanisms are

discussed herein.

When water freezes in the pore system of hardened cement paste

expansive stresses are developed which, if they exceed the

tensile strength of the cement, can cause severe cracking. An

early explanation developed by Powers (73,74) was that hydraulic

62
pressures are developed in the pores as water expands in volume

by 9% when converted to ice. The resulting expansion causes

excess water to move from the site of freezing to adjacent voids.

The hydraulic pressures are developed by resistance to the

movement of water from the site of freezing. If the voids are

not close to the sites of freezing or are not able to accommodate

the water, then the hydraulic pressures exceed the tensile

strength of the hardened cement paste. This hypothesis indicates

that the magnitude of the developed pressure depends on the rate

of freezing, extent of saturation and permeability of the paste,

and the distance from the freezing site to the nearest air-filled

void. Entrained air improves the frost resistance of the paste

by decreasing the average distance of the flow-path. The

hydraulic pressure mechanism does not account for many of the

processes observed when concrete with a predominant amount of

small pores freezes. However, it may be active in porous

concretes of large interconnected pore systems.

Another explanation developed by Powers (75) and Helmuth (76) is

analogous to the mechanism that causes frost heave in soils (see

Fragelund for ref.). In this hypothesis, osmotic pressures are

developed. Ice crystals formed in the capillary pores are able to

attract unfrozen water from gel pores because of differences in

free energy between ice and unfrozen water, resulting in the

movement of water to the site of freezing. The free energy of

the ice (measured by its vapor pressure) is lower than that of

63
the unfrozen water in the gel pores. The free energy of the ice

crystal increases as the free energy of the unfrozen water in the

gel pores decreases as the result of the internal desiccation.

The ice crystal continues to grow until the free energies are

equal. Sufficient osmotic pressures can develop to cause

cracking of the hardened cement paste. When the paste contains
entrained air bubbles, and the average distance between them is
less than a critical size, the unfrozen water preferentially

migrates to the air bubbles rather than to the capillaries. By

this process the osmotic pressure is reduced to well below the

tensile strength of the hardened cement paste.

The relative importance of the two damage mechanisms, development

of hydraulic pressure and osmotic pressure, is not unambiguously

known. However, it appears that the osmotic pressure mechanism

should be most effective in concretes with small pores, such as

concretes with low w/c ratios or dense concretes containing

pozzolanic admixtures (77) . Conversely, the hydraulic pressure

mechanisms would probably be active in porous concrete.

A later hypothesis by Litvin (78) suggests that the desorption of

water from gel pores is solely responsible for creating hydraulic

pressures in the hardened cement pastes and ice formation does

not occur in the capillaries.

64
 . . .

Regardless of which proposed mechanism is correct, the resistance

of hardened cement paste to frost action is known (79) to depend

on the spacing of entrained air bubbles, the amount of water in

the pores of the paste, the permeability of the paste, the

freezing rate, and the strength of the paste.

7.1.2 Freezing in Aggregates. Most rocks have pore sizes much

larger than those in cement paste and they expel water during

freezing (73) Some aggregates have an appreciable amount of

fine pores and if they contain large amounts of absorbed water,

severe cracking can occur through ice formation (80) The size

of an aggregate can control its resistance to frost damage with

frost resistance increasing as its size decreases. A critical

size exists below which aggregates can be exposed to freezing

temperatures while saturated with water without being damaged

(81). A detailed discussion of the effects of aggregates on the

frost resistance of concrete is provided by Cordon (79)

7.2 Factors Controlling Resistance to Frost Damage

Factors which should be considered in making concrete resistant

to frost damage are briefly discussed herein for concrete

structures storing LLW that will be buried underground below the

freeze line. The factors covered include exposure conditions,

w/c ratio, air entrainment, selection of materials, and curing.

A more complete set of recommendations dealing with various types

65
of concrete structures and exposures are given in the ACI Guide

to Durable Concrete (23) , and in Durcon (30)

7.2.1 Exposure to Moisture. As previously discussed, if air-

entrained concrete is kept in a state of nearly complete

saturation so that the entrained air bubbles are extensively

filled with water, it may suffer damage from cyclic freezing.

The critical saturation level could be exceeded if the concrete

is exposed to flowing or ponding water for a lengthy period of

time. Thus measures should be implemented during the

construction and filling stages to minimize the uptake of water.

For example, the geometry of the construction site should promote

drainage away from the structure,

7.2.2 Water-Cement Ratio. Frost-resistant normal weight concrete

should have a w/c ratio not to exceed the following:

1) Thin sections and sections with less than 25 mm of cover over

the reinforcement a w/c ratio of 0.40

2) All other structures a w/c ratio of 0.45

7.2.3 Entrained Air. The recommendations for air entrainment

are for normal strength concretes, with compressive strengths

below 55 MPa (8,000 psi) , Some problems have occurred with

entraining the recommended amount of air and controlling the air

66
 .

bubble spacing in higher strength concretes, as will be discussed

later.

Recommended air contents are given in table 3 for severe and

moderate exposures. A severe exposure is defined as a cold

climate where the concrete may be in almost continuous contact
with moisture prior to freezing, or where deicing salts are used.

If the exposure to moisture prior to freezing only occasionally

occurs, then the exposure is classified as moderate. The values

maybe slightly lowered if the performance of the concrete in a

known environment is known. The distance between the air bubbles

in the hardened cement paste is very important in obtaining a

frost resistent concrete. The air bubbles should have a maximum

spacing of 0.4 mm (0.016 in) or a spacing factor of 0.2 mm (0.008

in) .

The need for entraining air in high-strength concretes

(compressive strengths greater than 55 MPa (8000 psi) ) is a

controversial issue which is the subject of numerous studies as

follows. The resistance to freezing is directly dependent on the

capacity of concrete for freezable water and because of the

reduced porosity and permeability concretes of low w/c ratios may

not have a capacity for freezable water. Therefore, high-

strength concretes, should not need entrained air. However, some

studies have indicated that non-air entrained high-strength

concretes do not perform satisfactorily in laboratory tests (82)

67
Philleo (16) suggested that low w/c ratio concretes may have no

capacity for freezable water if all the available space in the

hardened cement paste is filled with hydration products; or the
paste becomes so impermeable that saturation by water will not

occur. This implies that non-air entrained high-strength

concretes could undergo frost damage if they are exposed

prematurely to freezing conditions. The review by Philleo (16)

is an excellent source of information on the freezing and thawing

resistance of high-strength concrete.

7,2.4 Materials and Tests. Whenever a durable concrete is

desired, the constituents, (cement, aggregates, and mineral and

chemical admixtures) should meet the requirements of the

appropriate ASTM specifications. The different types of portland

and blended cements, when used in properly proportioned and

prepared air-entrained concrete, will provide similar resistance

to frost damage.

Most fly ashes and natural pozzolans when used as admixtures

produce durable concrete provided that the recommended air

content and strength are attained.

The performance of an aggregate should be determined by tests of

concrete containing the aggregate. The most widely used

laboratory test on the freeze-thaw resistance of concrete is

68
[(Table 3. Recommended Air Contents for Frost-Resistant Concrete (23)

Average air content percent®

eNomial Maximum
aggregate Severe Moderate
<
mm (size, in.) Exposure** Exposure‘s

9.5 (3/8) 7.5 6

'12.5 (1/2) 7 5.5

il9 (3/4) 6 5

i38 (1.5) 5.5 4.5

75 (3) 4.5 3.5

‘:i50 (6) 4 3

1
® A resonable tolerance for air content in a field construction is =
1-1/2 percent.
^
Outdoor exposure in a cold climate where the concrete may be in
almost continuous contact with moisture prior to freezing, or where
deicing salts are used.
ii

Outdoor exposure in a cold climate where the concrete will be only

'
occasionally exposed to moisture prior to freezing, and where no deicing
;|
salts will be used.

69
ASTM C 666 (83) . It involves subjecting concrete, which is

initially saturated with water, to rapid cycles of freezing and

thawing. The durability of the concrete is determined by

measuring the reduction in its dynamic modulus. The test does

not realistically simulate in service conditions. However, the

relative performance of aggregates in the test will be predictive

of their relative performance in the field, which is of great

assistance in selecting aggregates for durable concrete. Some

aggregates have been found to only have fair performance in the

test but still the concretes made with them have adequate frost

durability. If a concrete has a high durability factor after 300

cycles it should be able to withstand a severe freezing and

thawing environment. At present C 666 is regarded as the

standard test and is most widely used.

Another test, ASTM C 671 (84) has been developed to more closely

simulate the moisture and freezing rates encountered in field

exposures and is less severe than ASTM C 666. It requires more

expensive equipment, which is not commercially available, and may

require more time to perform than C 666. The results obtained

with C 671 are sensitive to the moisture content of the concrete

prior to testing. However, it has been suggested that C 671 is a

more realistic test for high strength concrete than C 666 (16)
 ,

7.2.5 Curing. Proper curing can be very important because

immature air-entrained concrete can be damaged if it has not

developed adequate strength before being exposed to freezing

conditions. Normal strength air-entrained concrete should be

able to resist a few cycles of freezing and thawing when it

attains a compressive strength of about 3.45 MPa (500 psi)

provided it is not exposed to an external source of moisture.

Concrete should attain a compressive strength of 28 MPa (4000
psi) before being subjected to severe freezing and thawing

exposure. If the exposure is moderate, then a compressive

strength of 21 MPa (3000 psi) should be adequate.

7.3 Models and Service Life Predictions

Three models mathematically describing the mechanisms of frost

damage were found in the literature search and are discussed

herein. Only one of the models provides a basis for making

service life predictions and it is presented first.

7.3.1 Service Life Prediction Model. A model for predicting the

service life of concrete exposed to freezing and thawing cycles

has been developed by Fagerlund (85) . The main aspects of it are

outlined in the following. In the model air bubbles are termed

"air voids.” The spacing of unfilled air voids is treated as a

variable because it changes when water is filling the air void
system (air void absorption) . Frost damage will not occur until
the air void absorption reaches a critical value (i.e., the

71
critical air void absorption) and freezing temperatures are

encountered. Methods for obtaining the critical air absorption

have been described by Fagerlund based on the air void volume and

distribution, and the spacing of unfilled air voids.

The model considers that the complete volume of concrete of a

structure rarely fails during one freezing and thawing cycle but

occurs progressively, i.e., starting from the outer regions of a

volume of concrete, damage spreads inward through many freezing

and thawing cycles. The period of service of an element of

concrete is the period of time during which the amount of air

void absorption in the element is always lower than its critical

air void absorption, during periods of freezing temperatures.

The service life of the structure is the period of time after

which the structure can no longer fulfill its purpose because of

frost damage. A structure in constant contact with water and

exposed to freezing temperatures will have a finite service life

because eventually the critical air void absorption will be

exceeded. However, a structure that goes through cycled wetting

and drying phases could have a finite or infinite life, with

regard to frost damage, depending if the critical air absorption

is exceeded while the freezing temperature is reached.

The model is intended to give a means for making a reasonable

estimate of the service life of concrete. To make a reasonable

service life prediction, relationships for the long-term

72
absorption process in the air void system must be known as a

function of the properties of the concrete, such as permeability,

air void volume, and air void distribution. Several examples of

service life predictions made by using the model are given by

Fagerlund for concrete with non-porous aggregate.

7.3.2 Modeling of Frost Penetration. A two dimensional finite

element computer model for predicting frost penetration in

saturated aggregates and concrete has been developed by Olsen

(86) . Based on differential equations for predicting temperature

and moisture content, and equations relating freezing

temperatures and water potential, the model predicts temperature,

depth of frost penetration, water and ice content of aggregates

or concrete, and the pore water pressure. The model appears to

give predicted values which are in reasonable agreement with

experimental data.

The model can be used to analyze the influence of its parameters

on the freezing process. For example, the effect of selecting

aggregates with different desorption properties on frost

durability can be evaluated. However, the model cannot be used

for making service life predictions.

7.3.3 Modeling of Freezing and Thawing Processes. Bazant, et

al., (87) developed a model which involves the formulation of

relationships for sorption isotherms for concrete exposed to

73
freezing temperatures, diffusion of water, temperature

distribution, and internal stresses in the concrete resulting

from the freezing of water. The model essentially reduces the

freezing and thawing durability problem to the calculation of

stresses and strains. Solution of the model would require

finite-element computer analysis. To apply the model,

information would be needed on the effect of mechanical loading

on the freezing and thawing process and on the resulting

stresses. In its present stage of development the model cannot

be used to make service life predictions.

8. DETERIORATION OF CONCRETE BY LEACHING

Buried concrete in contact with water percolating through soils,

e.g., from precipitation, or groundwater, can undergo

deterioration by the dissolution of the soluble materials in the

hardened cement paste. In this report, the effects of damp soils

are taken to be the same as those of groundwater as the

mechanisms would be the same and over 500 years the effects may

be similar. The concrete vaults storing LLW are required to be

constructed above the groundwater table. However, the

groundwater table may change over 500 years and, therefore, the
effects of leaching by groundwater are covered. Considering the

common constituents of portland cement paste, alkali salts are

the most soluble followed by calcium hydroxide. The solubility

of silica oxide is very low so that the decomposition rate,

through dissolution, of the calcium silicate hydrate is very

74
 .

slow. Leaching can reduce the pH of the concrete as well as

making it more porous. Recently, the destruction of a concrete

conduit by calcium absorptive acidic groundwater was reported

(88) Such cases of complete deterioration are rare because

usually the water becomes sufficiently concentrated with the

dissolved salt so that a dissolution equilibrium is reached, or
the concrete is of sufficient high quality that the rate of

leaching is very slow and confined to the surface area. It is

doubtful that in the sites selected for the LLW storage

structures, that the underground water, by leaching alone, will

be able to endanger the structural integrity of the concrete.

However, if concrete is exposed to intermediately or continuously

flowing underground water for 500 years, other deterioration

process could be activated. For example, the amount of

dissolution of calcium hydroxide may be sufficient to lower the

pH at the depth of reinforcing steel and, thereby, promote

corrosion and cracking of the concrete. Thus, the leaching of

concrete is addressed in this section. In the following,

leaching is confined to direct dissolution processes as opposed

to reactions between aqueous solutions with cement reaction

products to cause volume changes, such as sulfate attack. Also,

it is presumed that the aggregates are inert to leaching.

8 . 1 Mechanisms

Leaching by water involves the penetration of concrete by water

or aqueous solutions, the dissolution of soluble constituents of

75
the hydrated cement paste, and the transportation of the
dissolved species to the underground water reservoir or its flow
system.

8.2 Factors Influencing Leaching

The depth of penetration of underground water into the concrete

is largely controlled by the permeability of the concrete and the

hydrostatic pressure of the water. Assuming a constant

temperature, the extent of dissolution is dependent on the

chemistry of the water and the amount of soluble constituents in

the hardened cement paste, and the mobility of the underground

water. If the water reservoir is small and stagnant then it is

likely that the concentration of dissolved salts will come in

equilibrium with the concrete and the dissolution process will be

limited. If, however, the underground water is flowing, then

dissolution equilibrium may not be attained and leaching can then

be a continuing process. Transportation of dissolved species in

concrete is dependent on their migration velocity through the

pore system and the flow rate of the ground water.

8.2.1 Permeability of Concrete. The permeability of concrete is

clearly one of the most important factors controlling the rate

and extent of leaching. Means for decreasing the permeability of

concrete have been discussed in section 3

76
 . .

8.2.2 Underground Water Movement. The flow rate of underground

water has been reported (89) to be a major factor in the leaching

of a concrete repository for radioactive waste disposal and

should be considered in selecting a site. A model has been

developed by Atkinson (89) by which the effects of the

underground water flow rate on leaching can be predicted. This

is discussed later.

8.2.3 Chemistry of Underground Water. Acidic waters will attack

concrete through the dissolution and neutralization of the alkali

and calcium hydroxides. The rate and extent of the acid leaching

is dependent on the acidity of the water, i.e., leaching

increases as the pH decreases. Calcium-absorptive acidic soils

are especially aggressive to concrete (88) The potential

leaching capabilities of ground waters can be related to their

Langelier Indexes (also called Saturation Index) . The Langelier

Index is related to the total solids, total alkalinity, pH, and

calcium content of the water (90) A positive index indicates

that calcium carbonate will be precipitated; a negative index

indicates lime-deficient water (soft water) capable of dissolving

calcium from hardened cement paste.

Carbonic acid can be found in some underground waters as a result

of the absorption of carbon dioxide by rain, which is most likely

to occur around industrial areas, or from the decaying of

vegetation. While such waters are usually acidic, they may be

77
neutralized by reacting with calcareous aggregates but still

contain significant amounts of the bicarbonate ion (91) . The

bicarbonate reacts with calcium hydroxide to precipitate calcium

carbonate;

Ca(HC 03)2 + Ca(0H)2 2 CaC 03 + 2 H2 O (16)

The resulting carbonate- free water can then dissolve calcium

carbonate to form more bicarbonate:

CaC 03 + CO 2 + H 2 O Ca(HC 03)2 (17)

so that the process is repeated (92) Through this process, the

bicarbonate ion can react with the calcium of the calcium
silicate hydrate resulting in its decomposition (92) It appears

that water containing more than 20 ppm of dissolved carbon

dioxide can result in deterioration of hydrated cement paste

(23) .

Some soils, such as peat, clay and alum shale may contain iron

sulfide (pyrites) which, upon oxidation and in the presence of

moisture, produces sulfuric acid (23) The oxidation could occur

in excavated soil sites which are left exposed to the atmosphere

for lengthy periods. Also, underground water may become

acidified with sulfuric acid by drainage through mines or mining

tailings, or by coming in contact with industrial waste.

78
A dense concrete with a low w/c ratio should be able to resist
most underground waters. If the underground water is acidic but

stagnant, the use of limestone aggregate should be considered.

In the unlikely case that a site is selected with flowing

underground water that is highly acidic or highly calcium-

sdDsorptive, a low permeability concrete with limestone aggregate

may not be sufficiently durable. In such a case, an appropriate

surface coating must be used (Appendix A) . The analyses of soils

and groundwaters would indicate the need for special precautions.

8.2.4 Chemistry of Hydrated Cement Paste. The constituent of

hydrated portland cement paste that causes the most severe

problems through leaching is calcium hydroxide. A typical
hardened portland cement paste could contain between 20 and 25%

calcium hydroxide by volume (93) . The leaching of a portion of

the calcium hydroxide would result in a porous matrix with an

increase in permeability. If a severe leaching condition exists,

the amount of calcium hydroxide in the paste could be decreased

by the addition of a pozzolan. However, the effect of the

reduction in calcium hydroxide on the stability of the calcium

silicate hydrate would need to be determined.

8.3 Models and Service Life Predictions

Atkinson (89) developed a model to estimate the pH change

occurring in the concrete of nuclear waste repositories by

leaching the constituents of the hydrated cement paste by

79
groundwater. This model deals with the chemical state of the

hydrated cement paste and does not consider the structural

integrity of the concrete structure. In this model the

repository is approximated as a sphere of equal volume to the

actual repository. The fractional release rate (dF/dt) of any

component of the concrete into the groundwater is related to the

flux density of the groundwater (J^) , and the concentration of

the component in the pore solution (C^) and its initial solid

concentration in concrete (C^) t

dF/dt = 3/4 (J^ q/R CJ (18)

where R is the radius of the concrete sphere. At very low flow

rates ^ the loss is controlled by diffusion to the surrounding

geology characterized by an intrinsic diffusion coefficient.

The groundwater flux density is determined by the hydraulic

gradient (h/L) and the hydraulic conductivity (K) of the

geological environment in which the repository is located:

J^=K(h/L). (19)

The flux density, is the volume of water crossing unit area

of ground in unit time and has units of m s°\ It is often

referred to as the groundwater flow rate (m^ s‘^ per m^ section)

(89) Ranges of hydraulic conductivity and corresponding

80
flux for a 1% hydraulic gradient, indicate that an "average"

value for groundwater flow rate could be taken to be 10'^^ m s*^

(89). According to Chapman, et al., (94) groundwater around a

"typical repository" could have a flow rate of up to 10 ms,

whereas groundwater in a fast aquifer flows at 10"^ m s'^ . Based

on the model, the effects of groundwater flow rates on the

fraction of calcium released from concrete is shown in Fig. 10.

Considering the flow rate and chemistry of the ground water, and
the chemistry and the permeability of the concrete, the model

predicts that the groundwater flow rate will have the largest

effect on the service life of concrete.

The pH of the concrete is an indicator of its chemical state as

it is subjected to leaching by groundwater, as schematically

represented in Fig. 11. Initially the pH is above 13 due to the

alkali hydroxides in the concrete pore solution. After the
alkali hydroxides have been removed by leaching, the pH is

controlled at 12.5 by calcium hydroxide. When the calcium

hydroxide is depleted, the Ca/Si ratio of the calcium silicate

hydrate (CSH) begins to decrease; when the ratio of 0.85 is

reached, the pH is 10.5. During the final dissolution stage, the
low Ca/Si calcium silicate hydrate dissolves congruently at a pH

of 10.5 until it is consumed. When the calcium silicate hydrate

is completely dissolved the pH will be either that of the

groundwater or of any residual salts in the original concrete

(89, 91).

81
The reduction of pH of a structure, with a radius of 20 m, and a

reference concrete with a portland cement content of 185 kg m°^,

caused by a groundwater flux density of 10'^° m s‘\ as a function

of time is shown in Fig 11. According to the model predictions,

the pH will be above 10.5 for about 4x10^ years in the mass of

concrete . In this calculation, the permeability of the concrete

was taken to be the same as the ground. Also, the model predicts

that a pH front of 10.5 will move towards the center of the

concrete at a rate of 1 m for every 1.5x10^ years.

The effects of groundwater containing 10*^ molar of "reactive

species" was estimated by the model to decrease the time for the

pH of the concrete to be reduced to 10.5, by a factor of four

compared to "nonreactive" groundwater. The major contributor to

the concentration of the reactive species was the bicarbonate ion

(95). Therefore, the model predicts a pH of 10.5 could be

experienced by concrete lying at a depth of 1 m from the surface

in about 3.5 x 10^ years, when exposed to the above conditions of

groundwater flow and salt concentration. Although these

estimates are rough they do suggest that good quality concrete
should not be directly degraded by groundwater in less than 500

years if the groundwater is flowing at a rate not exceeding 10*^®

m s‘\

An analysis of relationships between groundwater flow and pH of

concrete indicates that under certain conditions the passivation

82
film on reinforcing steel can be destroyed. According to Fig.

11, the calcium hydroxide of the mass concrete (20 m sphere) can

be completely depleted in about 10^ years, which translates into

a reduction of the pH to below 12.5 at the depth of 100 mm in

around 500 years. If the groundwater contains 10’^ M of reactive

species, then the estimated time is only 125 years. Therefore,

under the above conditions, reinforcing steel placed at 100 mm

below the surface of concrete may be exposed to a pH of less than
12.5 within 500 years of construction.

9. OTHER DETERIORATION PROCESSES

Deterioration processes which usually are not a problem during
the normal 50 to 100-year service life of concrete structures may

become more serious as concrete exceeds this life span. Two

examples of such possible processes are presented in this

section. In addition, deleterious substances may be present in

the LLW which can attack concrete and such materials are briefly

discussed.

9 . 1 Microbiological Attack

Sulfate producing bacteria, such as Thiobacillus thiooxidants,

are capable of oxidizing elemental sulfur and sulfides (e.g.,

iron sulfate, pyrite) to sulfuric acid in a relatively short

83
rate

Leach in
I

0) fd

E 0
>1
)

u
0)
Cu
'3
'O
c 0)

0) to

o td
0)
I— )
.

X M
(1) CTi
CO
3
s
3 S-i

•H OJ
ki U -M
0) rH (d
rG5
« U
C
T3

% iw 3
O O
“D Li

C C
O
Cn

3 •r-l >i
O 4J
O
o fd
L
u-i
(U
+j
(u
Li
'3 U
(U C
+J O
D) Id U
O g
•H 6
4J O
cn Lj
W 4-1

0)
Li
3
tr*
"H
in fa

T
I

00

repository)

(U
4J
o
u
c
o
volume
o
X!
CO
P
•H
CO k-
5
CO
unit (D X
a,

/ 0)
M-J
o
(89).

E (D
O
C
(1)

'V
groundwater
c
0)
in D) a
O T
0)
3
groundwater

(U
e
•H
+J

to
T3
<V
(Volume
-P T3
(0 (U
g tn
•H o
+j a,
cn X
w OJ

log^Q

o
p
3
Cn
•H

CO

Q.

85
period under aerobic conditions (92) Aerobic conditions could
exist during the period when the LLW storage site is excavated.

This process can acidify the groundwater which can deteriorate

concrete by an acid reaction as well as by sulfate attack. Some

bacteria can attack concrete by transforming ammonia into

nitrites and nitrates or by producing lactic acid and butyric
acid. In the normal design life span of concrete,

bacterial action does not appear to be a major cause of concrete

deterioration by itself, but can tend to make groundwater more

deleterious to concrete (92) It is difficult to predict whether

bacterial activity over 500 years will have much effect on buried
LLW concrete vaults. If necessary, the examination of buried

concrete vaults possibly could give useful information on the

long-term potential effects of bacteria activity.

9.2 Salt Crystallization

Dissolved soluble salts can be precipitated in the pores of

porous materials as a result of water evaporation. If the pores

are filled with salt crystals, further crystal growth can result

in the development of internal stresses. If these stresses

exceed the tensile strength of the material, cracking can occur.

Salt crystallization damage is usually a slow process occurring

near surfaces where evaporation takes place. It is a major cause

of stone deterioration (96) Recently, Sayward (97) suggested

that salt crystallization can make a significant contribution to

the deterioration of concrete. Salt damage would most likely be

86
 Table 4. Composition of Some Natural Groundwaters (95)

Granite Sandstone Shale Limestone Dolomite

pH 7.0 8.0 7.3 7.0 7,9

pNa 3.4 3.3 2.6 3.0 3,5

pK 4.0 4.0 4.2 3.7 »«

pCa 3.5 3.0 2.5 2.7 2.8

pMg 3.8 3.5 2.5 3.4 2.8

pH^SiO^ 3.2 3.9 3.5 3.7 3.4

pHCOj 2.9 2.6 2.1 2.3 2.2

pCl 4.0 3.7 4.0 3.2 3.3

pSO^ 4.2 3.2 2.2 3.4 4,7

“log 2,8 2.4 1.7 2.2 2.2

(ionic
strentgh

pX denotes minus logarithm (to base 10) of the

concentration of X in mole litre'^.

found on the inner walls of buried concrete structures, where

evaporation of a salt solution would be expected to take place.

Significant salt damage could be anticipated if permeable

concrete is exposed to rapidly flowing groundwater which is laden

with salts.

9,3 Attack by LLW

The most likely constituents of LLW to attack concrete are those

which are acids or which can form acids upon hydrolysis or

87
oxidation. Both inorganic and organic acids will react with the

calcium hydroxide of hydrated cement pastes and thus reduce the

pH of the pore solution and, if present in sufficient amounts,

destroy concrete. Organic substances such as alcohols,

aldehydes, and phenols can produce acidic solutions which can

attack concrete. The stronger acids will attack concrete the

most rapidly. However, over a long period of time the extent of

attack will be primarily dependent on the amount of acid present.

If acid attack is anticipated, the use of a calcarious aggregate

should be considered because, in principle, calcium carbonate

will neutralize an equivalent amount of acids

CaCOg + 2H3O* Ca^" + SH^O + 00^ (20)

Lists of a variety of inorganic and organic substances which are

known to attack concrete are given in Appendix B.

10. EMPIRICAL KINETIC MODELS

During the 1986 CEB/RILEM Workshop on Durability of Concrete,

kinetic models for many of the common deterioration processes of

concrete were proposed (98) These models are essentially

empirical models which need confirmation. Sub-models need to be

developed to solve these models similar to the sub-models

developed by Tuutti (37) for solving his corrosion model.
Nevertheless, they demonstrate what is generally known about the

88
differences in the reaction patterns of the deterioration

processes.

Mathematical models for carbonation, chloride penetration,

corrosion, frost attack, sulfate attack, and alkali-aggregate

reactions are given in table 5. The progress of deterioration

(in arbitrary units) are represented in Fig. 12. This figure is

intended to show only the general shape of the plots of

deterioration versus time and do not indicate relative

deterioration rates.

The curves for sulfate attack, alkali-silica reaction, corrosion

of reinforcing steel, and frost attack all exhibit an induction

period which is controlled by the rate of penetration or

diffusion of water or aqueous solutions in concrete. This

observation is consistent with the mechanisms described in the

sections dealing with a more detailed discussion of the

respective deterioration processes. Information on the relative

time of the induction period compared to the time for failure

after the initiation (t^^ in Fig. 12) is very desirable. It is

possible that if the induction period controls the deterioration

rate that data on penetration or diffusion rates could provide a

sufficient basis for estimating service lives.

89
11. RECOMMENDED RESEARCH PROGRAM

The recommended research program is based on the analysis of the

deterioration processes that are considered to be serious

problems for concrete buried in soils. Of the major processes

covered in the review part of this report, only freezing and

thawing does not appear to need further exploration in the

LLW underground storage program. Freezing and thawing is not

likely to be a problem if the concrete structures is covered with

soil and is below the freeze line. If the recommendations given

in Section 6 are followed, then the likelihood of freezing and

thawing damage occurring is minimal. At the present level of the

knowledge of alkali-aggregate reactions, it does not appear

likely that the service life of concrete containing potentially

reactive aggregates can be predicted with any confidence. It is

recommended that a test be developed, as outlined in Section 5.3,

to determine the potential alkali reactivity of an aggregate.

None of the presently available test methods nor the present

level of knowledge of alkali-aggregate reactions are adequate to

provide a basis for giving a reasonable level of assurance that a

potentially reactive aggregate will not undergo a severely

deleterious reaction within 500 years after construction.

Candidate concretes for evaluation are discussed first, followed

by a recommended approach for predicting the service life of

concrete buried underground. All forseeable deterioration

90
 Table 5. Durability of Models (98)

Form of Attack Model

Carbonation X = kt [1]

where k = (c D)/ b

Chloride penetration X = kt [2]

where k = (c D)/ b

Corrosion X = tg + .kt^® [3]

where t^ = kt in Eq. [1] or [2]

a = 0.5 for diffusion control

a = 1.0 for reaction control

Frost attack factors N = t„0 + K,R

6 [4]

where N is the number of cycles,

and Kg and R defines environmental
and resistance respectively

Sulfate attack X = tg + kt^® a>l [5]

Alkali-silica reaction X = tg + kt^® a>l [6]

a The undefined parameters are?

K is the rate constant,

t is time,
R is a constant
b is the reserve pasicity,
D is the diffusion constant, and
t^ is the initiation time

91
 of

Meanings

reaction

attack

TIME

deterioration.

Sulfate
Alkali-silica
*-0 5.

table
of

in
rates

given

Typical

*-0
are

^crit
12.

Figure

92
processes are addressed, including sulfate attack, corrosion of

reinforcing steel, leaching by groundwaters with various mineral

contents, and the synergistic effects of combinations of

chloride, sulfate, carbonate salts, and acids.

11.1 Candidate Concretes

Two different approaches (they can be combined) can be followed

in the design of concrete for constructing underground vaults for

storing LLW; 1) select cements and aggregates which have

mineralogical compositions that minimize aggressive chemical

attack; and 2) design dense concretes with low porosities and
permeabilities. Often the first approach, combined with some

reduction in permeability, is followed in designing durable

concrete with normal lifetimes (less than 100 years) . This

approach is often satisfactory if high quality concrete is

produced and it the service conditions are the same as those

considered in designing the concrete.

The second approach is made more feasible by recent developments

in concrete technology. As discussed in Section 3, very low

porosity concretes can be produced by combinations of high-range

water reducers (superplasticizers) , suitable pozzolans, silica

fume, and dense impermeable aggregates. A standard test does not

exist for measuring permeabilities of concrete and their

compressive strengths are often used as a measure of relative

permeabilities; both the permeability and compressive strength of

93
concrete are improved by reduction in porosity, thus high-
strength concretes tend to have lower permeabilities than normal

strength concretes. It is recommended that although concretes

are classified in the following on the basis of strength,

research on relationships between moisture movement in concrete

and service life should be performed, which would involve

developing an acceptable method for measuring factors controlling

moisture movement in concrete, e.g., permeability. Based on the

differences in materials and technologies for preparing high-

strength concretes, three permeability levels, based on

compressive strength classifications, are proposed for study? 1)

Class I, compressive strength below 70 MPa (10,000 psi) ? Class

II, compressive strength roughly between 70 MPa (10,000 psi) and

103 MPa (15,000 psi); and Class III, with compressive strengths

above Class II, Class I includes conventional concrete and would

be tested to provide a basis for determining the improvements

obtained with the more impervious concretes. Also, this would

facilitate comparisons of the large amount of data for

conventional concretes in the literature with the performances of

Classes II and III concretes.

Class II concretes can be readily prepared by ready-mixed

concrete companies. Class III concretes are becoming

commercially available. For example, a structure is being

constructed in Seattle, Washington (99) using ready-mixed

94
 . .

concrete with a compressive strength of approximately 138 MPa

(20,000 psi)

11.2 Recommended Approach for Predicting Service Lives

The principle methods for predicting the service life of concrete

that were disclosed in the review and presented in this report

are based on experience, accelerated testing, mathematical

modeling based on understanding the chemistry and physics of

deterioration processes, and empirically-derived correlations.

It is recommended that a combination of accelerated testing and

modeling be used to predict the service life of concrete selected

for underground storage of LLW.

A recommended practice for the development of accelerated short-

term tests which can be used for the prediction of service life

is given in ASTM E632 (5) The following discussion of the

application of the practice is based on reference no. 6. The

practice consists of four main parts (Fig 13) : 1) Problem

Definition; 2) Pre-Testing; 3) Testing; and 4) Interpretation and

Reporting of Data.

In Part 1, referring to the numbered boxes in the chart, the

first step (Step 1) is to define the performance requirements for

the material or component in service and to set performance

criteria which must be met. These criteria provide an objective

basis for selecting materials. In Step 2, the material should

95
be characterized in terms of its constituents and interfaces

between the constituents. Concrete should be characterized in

terms of its composition and mixture proportions, mechanical

properties, permeability, the chemistry and microstructure of the

hardened cement paste, and the mineralogy of the aggregates.

Critical performance characteristics are specified in Step 3

these characteristics will be used in delineating the limiting

condition below which the material or component is deemed

unserviceable. For example, failure criteria for the concrete of

LLW storage vaults may be 1) the corrosion of reinforcing steel,

2) the deterioration of 50% of the concrete, or 3) the movement of

sufficient groundwater through the concrete to attack the LLW

containers. In Step 4, the expected ranges of degradation

factors to which the material or component is likely to be

exposed in service are identified. For example, to develop an

accelerated test for sulfate attack for underground buried

concrete, the range of the concentrations of sulfates in the

groundwater, at the different sites, needs to be known. Possible

synergistic effects between degradation factors must be

identified. An example of a synergistic effect is the combined

influence of groundwater flow and groundwater chemistry on the

service life of concrete (discussed in Section 8) . The

identification of the possible deterioration processes. Step 5,

is needed to develop simulative accelerated tests. With the

knowledge obtained in Steps 1 to 5, it may be postulated how the

degradation process can be accelerated. If degradation can be

96
 ,

^MT I . PWiLEH KFINITIOI

Figure 13 Methodology for developing and using

accelerated tests in predicting service lives (5)

97
accelerated without changing the mechanism of failure, then

laboratory test time can be reduced. Once the experimental

procedure has been determined, the performance requirements for

the test specimens can be stated (Step 7) . It must be recognized

that much of the knowledge may not always be available. In such

cases, assumptions based on the best available opinions and

experience should be made. Part 1 is largely completed for

concrete storing LLW underground by the information contained in

the present report. However, further improvments in the

understanding of relevant deterioration processes is urgently

needed

When Part 1 is completed, Part 2, Pre-testing, can be started

(Step 8) . Its purpose is to demonstrate that rapid failure can

be caused by intensifying the factors specified in Part 1. If it

is previously known that rapid failure can be induced, then this

step can be deleted. Increases in the concentration of sulfates

in groundwater is known to increase the rate of sulfate attack on

concrete and thus does not need to be explored.

Accelerated durability testing is carried out in Part 3 (Step 9)

At the same time (Step 10) ,

long-term tests under in-service

conditions should be initiated. The results of long-term tests

provide the most convincing evidence that the results of

accelerated tests can be extrapolated to in-service conditions.

In the case of developing methods for predicting the service life

98
of concrete for LLW underground vaults, the performance of in-

service concrete structures exposed to groundwater or to flowing

surface waters could be investigated and the results compared

with the results of the accelerated tests. If the data from in-

service concrete is consistent with the predictions from

accelerated tests then confidence in the extrapolations is

significantly increased.

If the results of the accelerated tests and long-term tests are

consistent with each other, Part 4, Interpretation and Reporting

of Data, should be undertaken. This includes use of experimental

data to predict the course of deterioration under expected in-

service conditions (Step 13) and to predict the time at which

failure, as defined by the performance criteria, will occur

(Steps 14 and 15) . Mechanistic mathematical models are developed

in this Part to describe deterioration processes and to relate

the rates in accelerated tests with those of long-term in-service

tests. The final step (Step 16) involves reporting the data with

clear and explicit statements of the assumptions used in

developing the models.

It is recommended that the methodology given in ASTM E 632 be

followed in developing accelerated tests and procedures for

predicting the service life of concrete.

99
11.2.1 Accelerated Testing

Concrete deterioration processes which are likely to be important

in LLW storage are largely controlled by the ingress of

aggressive solutions into the concrete. For this reason^

laboratory experiments should be aimed at developing data on

penetration rates which are needed for estimating the
concentrations of agrressive species accumulating with time in

any volume element of the concrete. Diffusion and permeability

data will be needed for predicting the service lives of concrete

exposed to sulfates, chlorides, and carbonates, and to ground

waters with a negative Langelier Index and/or with an acidic pH.

Because the permeability coefficients in good quality concrete

are low (<10‘^° m s’’) , acceleration will be required to obtain

penetration rates which can be used in the present study.

Possible methods for accelerating the penetration of aqueous

solutions into concrete include increasing the temperature, using

a series of concentrations of attacking species which range from

expected in-service concentrations to concentrations

substantially above the "worst case", forcing solutions into

concrete by application of hydrostatic pressure, and cycles of

wetting and drying.

Relationships between the depth of penetration at in-service

conditions and penetration results from accelerated testing can

be evaluated by application of a modified form of the time

100
 :

transformation function, p,. (t), which is normally expressed in

the equation (100)

F,(t) = F^(p,.(t)) (21)

where t is time, Fj(t) is the life distribution at the i'th

accelerated stress level, and FQ(t) is the life distribution at

the in-service stress level. In the modified form, P|(t) is the

distribution of the depths of penetration with accelerated stress

i, and Pg(t) is the distribution of depths of penetration under

in service conditions;

P.(t) = P^(p,(t)) . (22)

The penetration-time transformation function can be obtained by

using a range of elevated stresses, where dissolved salts in

groundwater, temperature, hydrostatic pressure, and wetting and

drying cycles are examples of environmental stresses. An example

of the use of the penetration-time transformation function to

predict the depth of penetration at service stress Sq is shown in

Fig. 14. In the figure, the depth of penetration for a threshold

concentration is plotted versus time for stresses above the

service stress. When the threshold concentration has penetrated
to a critical depth, either failure has occurred or the failure

process has been initiated.

101
The time transformation function approach is applicable if the

deterioration mechanism under all tested stress levels is the

same as that taking place at in-service conditions (100) . Also,

it is assumed that deterioration begins at the instant of stress

application and that deterioration is an irreversible cumulative

process. Once a well-behaved penetration-time transformation

function is obtained, it can be used to predict the depth of

penetration into concrete as a function of time at LLW storage

sites. To be of general use, transformation functions will need

to be obtained for the range of concretes used in construction of

the storage structures. If a relationship can be derived between

the function and some concrete parameter or property (e.g., w/c

ratio, compressive strength, porosity, or permeability) then

values for the time transformation function can be extrapolated

from a few concretes to a wider range of concretes.

Synergistic effects may develop between deterioration processes,

e.g., sulfate attack and corrosion (101). It is recommended that

the importance of synergistic effects be assessed by the

combination of time transformation functions for different

processes with a time synergistic function, and using a

multivariable analysis (100) .

11.2.2 In-Service Concrete Studies

Adequate long-term testing can require many years of exposure,

especially when attempting to predict the service life of

102
concrete which is intended to remain in service for 500 years.
Investigation of the performance of in-service concrete can
supplement or possibly provide an alternative to long-term

testing that is based on companion specimens to those of the

accelerated testing program. It is recommended that specimens be

taken from in-service concrete structures exposed to groundwater

or to flowing surface water. Depth of deterioration and the depth

of alteration in the microstructure and chemistry of specimens

can be determined by light and scanning electron microscopy.

Even though the concretes from the in-service structures may

differ from those in the accelerated testing program, they could

provide valuable information on the type and extent of concrete

deterioration which can be anticipated. If the concrete and

service environment are adequately characterized, the data on

depth of penetration and extent of deterioration could be used to

test if the models give at least qualitatively reasonable

predictions.

11.2.3 Models for Service Life Prediction

Once the values for the inputs are known, models can be used to

predict the service life of concrete. The type of model

developed by Tuutti (37) for corrosion predictions is well-suited

for the present study. By coupling it with penetration data

obtained from accelerated testing, estimates of the service life

of reinforced concrete could be obtained. It is recommended that

103
 C
o
Concrete

sulfate

of concrete.

model

of

conceptual
formation,

LJJ

s
H-
proposed

ettringite

of

by

Schematic

attack,

14,

Figure

104
the development of a similar type of model be undertaken for
sulfate attack, which could be coupled to penetration data

obtained from accelerated testing.

A proposed model for predicting the service life of concrete

subjected to sulfate attack is presented in Fig. 14. It consists

of two major mechanistic steps: Step 1 , diffusion of sulfate ions

to an unit element in a concrete component; and Step II, the

expansion and cracking of the unit element of concrete due to the

formation of ettringite. Step I is completed when the threshold

concentration (concentration of sulfate ions required

to cause concrete expansion through ettringite formation) is

reached at the unit element of concrete. Thus in Step I,

ettringite formation is initiated. When no further ettringite

crystals can be accommodated in the pore volume, expansion will

start (beginning of Step II) . Cracking will occur when the

internal stresses (clT^j) exerted by ettringite formation exceed

the tensile strength (f^) of the concrete. Failure is

considered to occur when the unit element cracks. If the time

for the completion of Step I is much greater than that for Step

II, then the prediction of service life is reduced to only

considering Step I. Even if only Step I needs to be modeled,

determination or estimation of penetration rates is not
sufficient, by itself, to predict the service life. The model
must also be capable of predicting if a sufficient amount of

105
 CO
CO
0
-C ^^
D.
0 0)
D o
”>
0 CO

concentration

stresses,

threshold

in-service

of

at
S
advance

concrete

stresses,

the

solution

T
elevated

predicting

LU
1
on

aggressive
for

based

Method ,

of
S

15.

Figure

N011Vyi3N3cJ dO Hld3a

106
ettringite will be formed by the sulfate - tricalcium aluminate
reaction to cause cracking. Otherwise the model may falsely

predict failure when the tricalcium aluminate content of the

cement is too low to induce cracking.

The model developed by Atkinson (89) for leaching by groundwater

should be further evaluated to determine if it gives, at least,

reasonable service life predictions. Also, recommendations

should be developed on the acceptability of sites based on their

groundwater flow and chemistry, including their Langelier Index.

11.3 Research Priorities

Based on analyses of the likely deterioration processes of

I

concrete buried underground and the factors controlling their

rate, the following research priorities, in order of decreasing

importance are recommended;

1) Determination of penetration rates of sulfate, chloride, and

carbonate ions and acids in candidate concretes. By using a

modified time transformation function estimate the penetration

rate at in-service stresses from penetration data obtained at

elevated stresses. Predict service lives based on time required

for specific concentrations to reach critical (failure) depth.

2) Develop mathematical model to predict service life of concrete

exposed to sulfate ions, based on conceptual model schematically

presented in Fig. 14. Model considers both rate of sulfate

penetration and the rate of expansive reactions. Use the

107
penetration data obtained in Item 1) to solve the sulfate model.

Validate model by predicting depth of sulfate attack of in-

service concrete.

3) Develop method for determining potential alkali reactivity of

aggregates. Validate method by testing aggregates of in-service

concrete showing reactive and apparent unreactive aggregates.

4) Determine reliability of mathematical models for predicting

service lives of concrete subjected to corrosive environments and

concrete exposed to leaching conditions. If necessary, modify

existing models or undertake development of new models.

5) Carry out service life evaluation of barrier coating systems

for concrete using accelerated testing methodologies.

6) Investigate joint systems. The possibility of either avoiding

or reducing the amount of joints should be evaluated by a stress

analysis. Also, the use of prestressing tendons should be

considered in the analysis. Carry out service life evaluation of

the selected joint system or its alternative.

The first four research priorities are aimed at identifying

concretes which will have reasonable probabilities of having

service lives of at least 500 years. The fifth research priority

reflects the reality that even good quality concrete can crack.

However, the use of protective coatings does not convert a poor

quality concrete into a high-performance concrete. Joints may be

a weak link in a structure once high quality concrete is

108
obtained. It appears that joint materials have short service

lives.

12. SUMMARY AND CONCLUSIONS

In this report, the major deterioration processes that concrete

of underground LLW storage vaults will need to resist are

identified and analyzed. These processes are sulfate attack,

corrosion of reinforcing steel, alkali-aggregate reactions, and

leaching by underground water. Freezing and thawing damage could

occur before the vaults are covered with soil. Other

deterioration processes which could occur are microbiological

attack, salt crystallization, and attack by LLW forms, especially

acids.

All of these processes involve the penetration of the concrete by

water or aqueous solutions. Therefore, it is concluded that

concrete with low permeabilities are most likely to achieve

service lives of around 500 years.

Some research has been carried out on predicting the service life

of concrete based on the development of models. Models have been

developed for predicting service lives of concrete exposed to

corrosive conditions, freezing and thawing, and leaching by

groundwater. These models appear to give at least semi-

quantitative estimates of service lives. Further work on
developing models is recommended. For example, a modeling

109
concept for predicting the service life of concrete exposed to
sulfates is described.

It is concluded that, in developing an approach for predicting

the service lives of concretes, with desired lives of 500 years,

that accelerated testing should be combined with mathematical

models. Accelerated tests should be developed following the

procedures outlined in ASTM E 632 (5) .

Although alkali-aggregate expansive reactions do not often cause

structural failure, they pose a difficult problem. Present test

methods are not adequate to predict the behavior of potentially

reactive aggregates in field concrete. Furthermore, the present

knowledge on alkali-aggregate reactions is inadequate for making

service life predictions. It is recommended that a test method

be developed for determining if an aggregate is potentially

reactive.

The long-term performance of concrete is often controlled by the

level of its quality. Based on available knowledge, it seems

likely that concrete can be formulated which should have service

lives of 500 years in the expected environments. However, such

lives will be obtained only if the concrete materials meet

carefully defined specifications, and good construction practices

are followed. To assess this, an adequate QA/QC program must be

established and faithfully implemented.

110
13 . ACKNOWLEDGMENTS
The authors gratefully acknowledge the sponsorship of the U.S.

Nuclear Regulatory Commission (NRC) for this research. The

helpful suggestions by Mr. Jacob Philip, NRC, and Dr. Jon Martin,

National Institute of Standards and Technology are greatly

appreciated.

Ill
14. REFERENCES

1. P.E. Halstead, "Causes of Concrete Decay," The Consulting

Engineer (Supplement) April -May (1971)
,

2. B.M Idorn and N. Thaulow, "Examination of 136 Year Old

Portland Cement Concrete," Cement and Concrete Research,
vol. 13, 739-743 (1983).

3. A. Atkinson, D.J. Goult, and J.A. Hearne, "An Assessment of

the Long-Term Durability of Concrete in Radioactive Waste
Repositories," Materials Research Society Symposium
Proceedings: Scientific Basis for Nuclear Waste Management
IX, Vol. 50, 239-246 (1986)

4. J. Pommersheim and J. Clifton, "Prediction of Concrete

Service-Life," Materials and Construction, Vol. 18 (103),
21-30 (1985)

5. "Standard Practice for Developing Accelerated Tests to Aid

Prediction of the Service Life of Building Components and
Materials," ASTM Designations E 632-81, American Society for
Testing and Materials (1981)
6. G. Frohnsdorff, L.W. Masters, and J.W. Martin, "An Approach
to Improved Durability Tests for Building Materials and
Components," NBS Technical Note 1120, National Bureau of
Standards, July (1980)

7. "Nondestructive Inspection and Quality Control," Metals

Handbook, Vol. II, American Society for Metals (1976).

8. "Quality Assurance/ Quality Control Survey," Construction

Industry Cost Effectiveness Project, Business Round Table.

9. H. Woods, "Durability of Concrete Construction," ACI

Monograph No. 6, American Concrete Institute (1968)
10. Ac Atkinson, "The Influence of Wasteform Permeability on the
Release of Radionuclides from a Repository," Nuclear and
Chemical Management, Vol 5, 203-214 (1985)

11. BcK. Name and J.M. Illston, "Capillary Pore Structure of

Hardened Cement Paste," 7th International Symposium on
Chemistry of Cement, Vol. Ill, pp. VII-181 -VII-186 (Paris,
1980)

12. BcK. Name and J.M. Illston, "Relationships between

Permeability and Pore Structure of Hardened Cement Paste,"
Magazine of Concrete Research, Vol. 33 (116), 139-146
(1981)

112
 . . . . . . . . . , .

13. S. Mindness and J.F. Young, "Concrete,” Prentice-Hall, inc.

(Englewood, NJ, 1981)

14. "Concrete Manual," 8th ed. , U.S. Bureau of Reclamation

(1975)

15. P.C. Hewlett, "The Concept of Superplasticized Concrete,"

ACI SP-62, Superplasticizers in Concrete (1979)

16. R.E. Philleo, "Freezing and Thawing Resistance of High-

Strength Concrete," National Cooperative Highway Research
Program Synthesis of Highway Practice 129 (1986)
17. E.J. Sellevold, D.H. Eager, E.K. Jenses, and T. Knudsen,
"Silica Fume-Cement Pastes Hydration and Pore Structure," in
:

Condensed Silica Fume in Concrete, Report BML 82.610,

Institut for Bygningsmateriallaere, pp 19-50 (Trondheim,
Norway, 1982)

18. H. Cheng-yi and R.F. Feldman, "Influence of Silica Fume on

the Microstructural Development in Cement Mortars " Cement ,

and Concrete Research, Vol 15 (5), 283-294 (1985).

19. T.C. Powers, L.E. Copeland, J.C. Hayes, and H.M. Mann,
Journal of the American Chemical Society, Vol 51 (3) 285-
298 (1954).

20. K.L. Saucier, "High-Strength Concrete for Peacekeeper

Facilities," Miscellaneous Paper SL-84-3, U.S. Army
Waterways Experiment Station, Vicksburg, Miss. (1984)
21. R.D. Browne and P.L. Domone, "Permeability and Fatigue
Properties of Structural Marine Concrete at Continental
Shelf Depths," International Conference on Underwater
Construction Technology, Cardiff (1975)
22. "Non-Structural Cracks in Concrete," Concrete Society
Technical Report No. 22 (London, 1982)

23. "Guide to Durable Concrete," ACI 201.2R-77, American

Concrete Institute (1977)

24. W.C. Hanson, The Chemistry of Sulphate-Resisting Portland

Cements, E.G. Swenson (ed.). University of Toronto Press
(1968)

25. A.M. Neville, "Properties of Concrete," (3rd ed.). Pitman

Publishing Ltd. (London, 1981)

113
26. D.C. Miller and P.W. Manson, "Long-Term Tests of Concretes
and Mortars Exposed to Sulfate Waters," Technical Bulletin
No. 184, Agricultural Experimental Station, University of
Minnesota (1951)
27. Advisory Committee, Long-Term Study, "Ten- Year Report on the
Long-Term Study of Cement Performance in Concrete," ACI
Journal, Vol 49 (7), 601-616 (March, 1953).

28. D.M. Roy, "Mechanisms of Cement Paste Degradation Due to

Chemical and Physical Factors," 8th International Symposium
on Chemistry of Cement, Vol. 1, 362-380 (Rio de Janeiro,
1986)

29. E. —
Dunston, "Sulfate Resistance of Fly Ash Concretes -The R-
Value," in Concrete Durability: Katherine and Bryant Mather
International Conference, ACI SP-100, Vol. 2, pp 2027-2040
(1987)

30. J.R. Clifton and B. C. Oltikar, "Expert System for Selecting

Concrete Constituents," in Computer Applications in Concrete
Technology, ACI SP-98, pp 1-24 (1987).

31. P.W. Brown, J.R. Clifton, G. Frohnsdorff, and R.L. Berger,

"The Utilization of Industrial By-Products in Blended
Cement," Proceedings of the Fifth Mineral Waste Utilization
Symposium, pp 278- 284 (Chicago, 1976)

32. G.L. Kalousek, L.C. Porter, and E.J. Benton, "Concrete for
Long-Term service in Sulfate Environment," Cement and
Concrete Research, Vol. 2 (1), 79-90 (1972).

33. W.H. Harrison and D.C. Teychenne, "Sulphate Resistance of

Buried Concrete: Second Interim Report on Long Term
Investigation at Northwick Park, Building Research
Establishment (Garston, 1981)

34. C.N. Scott, "Corrosion Protection Properties of Portland

Cement Concrete," Journal of American Water Works
Association, Vol. 57 (8), 1038-1052 (1965),

35. GeE. Monfore and G.J. Verbeck, "Corrosion of Prestressed

Wire in Concrete," Journal of American Concrete Institute,
Vol. 32 (9), 491-515 (1960).

36. "Corrosion of Metals in Concrete," ACI Journal, 3-32 (Jan-

Feb, 1985)

37. K. Tuutti, "Corrosion of Steel in Concrete," Swedish Cement

and Concrete Research Institute, Stockholm (1982)

114
 . . . . . . .

38. K.C. Clear, "Time-to-Corrosion of Reinforcing Steel in

Concrete Slabs. Vol 3; Performance After 830 Daily Salt
Applications,” Report No. FHWA=RD-76-70, Federal Highway
Administration (1976)
39. G.J. Verbeck, "Mechanisms of Corrosion of Steel in
Concrete," in Corrosion of Metals in Concrete, ACI SP--49,
American Concrete Institute, pp 21-38 (1978)
40. J.R. Clifton, H.F. Beeghly, and R.G. Mathey, "Nonmetallic
Coatings for Concrete Reinforcing Bars," Report No. FHWA-RD
74-18, Federal Highway Adminstration (1974)

41. D.A. Hausmann, "Steel Corrosion in Concrete," Materials

Protection, November, 19-23 (1967)

42. C.L. Page, N.R. Short, and W.R. Holden, "The Influence of
Different Cements on Chloride-Induced Corrosion of
Reinforced Steel," Cement and Concrete Research, Vol. 16,
79-86 (1986).

43. R. Turriziani, "Internal Degradation of Concretes Alkali-

Aggregate Reaction. Reinforcement Steel Corrosion," 8th
International Symposium on Chemistry of Cement, Vol 1, 388-
442 (Rio de Janeiro, 1986)

44. C.L. Page and K.W.J. Treadaway, "Aspects of the

Electrochemistry of Steel in Concrete," Nature, Vol. 297,
109-115 (1982)

45. . C.E. Locke, "Mechanism of Corrosion of Steel in Concrete,"

Proceedings of Solving Rebar Corrosion Problems in Concrete
Chicago NACE Convention (Sept. 1982)

46. "Building Code Requirements for Reinforced Concrete," ACI

318-83, ACI Manual of Concrete Practice 1985, Part 3,
American Concrete Institute.
47. C. Dehghanian and C.E. Locke, "Electrochemical Behavior of
Steel in Salt Contaminated Concrete; Part 1," Corrosion,
Vol. 39 (8), 299-305 (1983).

48. C. Dehghanian and C.E. Locke, "Electrochemical Behavior of

Steel in Salt Contaminated Concrete; Part 2," Vol. 38 (9),
494-499 (1982)

49. K. Tuutti, Op. Cit. 295

50. K. Tuutti, Op. Cit. 143

115
51. Z.P. Bazant, "Physical Model for Steel Corrosion in Concrete
Sea Structures-Theory, " ASCE Journal of the Structures
Division, Vol. 105 (ST6),
1137- 1153 (June 1979).

52. Z.P. Bazant, "Physical Model for Steel Corrosion in Concrete

Sea Structures-Application, " ASCE Journal of the Structures
Division, Vol. 105 (ST6) 1155-1166 (June 1979).

53. R.D. Browne, "Mechanisms of Corrosion of Steel in Concrete,"

in Concrete in Relation to Design, Inspection, and Repair of
Offshore and Coastal Structures, ACI SP-65

54. L.S. Brown, "Some Observations on the Mechanics of Alkali-

Aggregate Reaction," ASTM Bulletin No. 205, p. 40 (1955).
55. A.L. Buck and K. Mather, "Concrete Cores from Dry Dock No.
2, Charleston Naval Shipyard, S.C.," Miscellaneous Paper No.
C-69-6, U.S. Army Waterways Experimental Station, Vicksburg,
Miss. (1969).

56. K. Mather, "Examination of Cores from Four Highway Bridges

in Georgia," Miscellaneous Paper No. C-73-11, U.S, Army
Waterways Experimental Station, Vicksburg, Miss. (1973)
57. L.J. Struble, "The Influence of Cement Pore Solution on
Alkali-Silica Reaction," Ph.D. Dissertation, Purdue
University, published as NBSIR 87-3632, National Bureau of
Standards (1987)

58. "Standard Practice for Petrographic Examination of

Aggregates for Concrete," ASTM Designation: C 295.
59. "Standard Test Method for Potential Alkali-Aggregate
Reactivity of Cement-Aggregate Combinations," ASTM
Designation C 227.

60. M.A.G. Duncan, E.G. Swenson, J.E. Gillott, and M.R. Foran,
"Alkali-Aggregate Reaction in Nova Scotia: I. Summary of a
Five Year Study," Cement and Concrete Research, Vol. 3 (1),
55-69 (1973).

61. R.E Oberholster and G. Davies, "Accelerated Method for

Testing the Potential Reactivity of Siliceous Aggregate,"
Cement and Concrete Research, Vol. 16 (2) 181-189 (1986)

62. "Standard Test Method for Potential Reactivity of Aggregates

(Chemical Method)", ASTM Designation C 289.

63. "Standard Practice for Concrete," U.S. Corps of Engineers

Engineering Manual EM 1110-2-2000 (1985)

116
 , . . . .

64. "Standard Specification for Portland Cement," ASTM

Designation: C 150.

65. D. Stark, "Alkali-Aggregate Reactivity in the Rocky Mountain

Region," Proceedings of 4th International Conference on
Effects of Alkalies in Cement and Concrete, pp. 235-243,
Purdue University (June 1978)

66. D. Stark, "Alkali-Aggregate Reactivity: Some

Reconsiderations," Cement, Concrete and Aggregates, Vol. 2,
92-94 (1980)

67. M. Kawamura, K. Takemoto, and S Hasaba, "Effectiveness of

Various Silica Fumes in Preventing Alkali-Silica Expansion,"
in Concrete Durability: Katherine and Bryant Mather
International Conference, ACI SP-100, Vol. 2, pp. 1809- 1819
(1987) .

68. J.E. Gilliot and E.G. Swenson, "Mechanism of the Alkali-

Carbonate Reaction," Journal of Engineering Geology, Vol. 2,
7-23 (1969).

69. "Standard Test Method for Potential Reactivity of Carbonate

Rocks for Concrete Aggregate (Rock Cylinder Method)," ASTM
Designation C 586.
70. H.H. Newlon and W.C. Sherwood, "Methods for Reducing
Expansion of Concrete Caused by Alkali-Carbonate Reactions,"
Highway Research Record, No. 45, 134-150 (1964)

71. H.H. Newlon, W.C. Sherwood, and M.A. Ozol, "Potentially

Reactive Carbonate Rock, Progress Report No. 8, A Strategy
for Use and Control of Potentially Reactive Carbonate Rocks
Including an Annotated Bibliography of Virginia research,"
Virginia Highway Council (1972)
72. A.M. Neville, Op. Cit. 407.

73. T.C. Powers, "A Working Hypothesis for Further Studies of

Frost Resistance of Concrete," ACI Journal Proceedings, Vol.
41 (4) 245-272 (1945) .

74. T.C. Powers, "Basic Considerations Relating to Freezing and

Thawing Tests," Proceedings ASTM, Vol. 55, 1132-1155 (1955).
75. T.C. Powers, "Resistance of Concrete to Frost at Early
Ages," Proceedings, RILEM Symposium on Winter Concreting
(Copenhagen, 1956) ,Research Bulletin No. 71, Portland
Cement Association.

76. R.A. Helmuth, "Capillary Size Restrictions on Ice Formation

in Hardened Portland Cement Pastes," 4th International

117
 Symposium on Chemistry of Cement, Vol. 2, 855-869 (published
as Monograph 43, National Bureau of Standards, 1962).

77. G. Fagerlund, "Prediction of the Service Life of Concrete

Exposed to Frost Action," in Studies on Concrete Technology,
Swedish Cement and Concrete Research Institute (1978)

78. G.G. Litvin "Phase Transitions of Adsorbates: IV. Mechanism

of Frost Action in Hardened Cement Paste," Journal of the
American Ceramic Society, Vol. 55 (1), 38-42 (1972).
79. W.A. Cordon, "Freezing and Thawing of Concrete-Mechanisms
and Control," ACI Monograph No. 3, American Concrete
Institute (1966)

80. R.A. Helmuth, "Dimensional Changes of Hardened Cement Pastes

Caused by Temperature Changes," Proceedings Highway Research
Board, Vol. 40, 315-336 (1961).

81. G« Verbeck and R« Landgren, "Influence of Physical

Characteristics of Aggregates on the Frost Resistance of
Concrete," Proceedings ASTM, Vol. 60, 1063-1079 (1960).

82. V.M Malhotra, "Mechanical Properties, and Freezing-and-

Thawing Resistance of Non-Air-Entrained Condensed Silica-
Fume Concrete using ASTM C 666, Procedures A and B," in Fly
Ash, Silica Fume, Slag, and Natural Pozzolans in Concrete,
ACI SP-91, pp. 1069-1094, American Concrete Institute
(1986)

83. "Standard Test Method for Resistance of Concrete to Rapid

Freezing and Thawing," ASTM Designation: C 666.

84. "Standard Test Method for Critical Dilation of Concrete

Specimens Subjected to Freezing," ASTM Designation C 671.

85. G. Fagerlund, "Service Life of Concrete Structure,"

Proceeding of Contemporary European Concrete Research,
Stockholm (1981)
86. M.P.J. Olsen, "Mathematical Modeling of the Freezing Process
of Concrete and Aggregates," Cement and Concrete Research,
Vol. 14, 113-122 (1984).

87. Z.P. Bazant, J.C. Chern, A.M. Rosenberg, and J.M. Gaidis,
"Mathematical Model for Freeze-Thaw Durability of Concrete,"
Report No. 86-ll/694m, Center for Concrete and Geomaterials,
The Technological Institute, Northwestern Univ. (1962)

88. WoG. Hime, B. Erl in, and R.R. McOrmond, "Concrete

Deterioration Through Leaching with Soil Purified Water,"
Cement, Concrete, and Aggregates, Vol. 8 (1), 50-51 (1986).

118
 . . . . .

89. A. Atkinson, "The Time Dependence of pH within a Repository

for Radioactive Waste Disposal," Report AERE - R 1177,
Materials Development Division, Harwell Laboratory,
Oxfordshire (1985)

90. W.F. Langelier, Journal of the American Water Works

Association, Vol. 23, 1500 (1936).

91. F.M. Lea, "The Chemistry of Cement and Concrete," 3rd ed..
Chemical Publishing Co., Inc., New York (1971).

92. I. Biczok, "Concrete Corrosion and Concrete Protection,"

Chemical Publishing Co., Inc., N.Y. (1967).

93. S. Mindness and J.F. Young, Op. Cit. , 97,

94. N.A. Chapman, I.G. McKinley, and D. Savage, Proceedings

OECD/NEA Workshop on Radionuclide Release Scenarios for
Geologic Repositories, 91 (1981)

95. W. Strum and J. Morgan, "Aquatic Chemistry," 2nd ed. , John

Wiley Publishers, Inc., N.Y. (1981),
96. E.M. Winkler, "Stone: Properties, Durability in Man's
99. Environment, Springer-Verlag, Vienna (1973)

97. J.M. Sayward, "Salt Action on Concrete," Special Report 84-

25, U.S. Army Cold Regions Research and Engineering
Laboratory, Hanover, NH (1984)

98 CEB/RILEM Workshop on Durability of Concrete Structure,

Report of Group 1 on Material Data for Service Life
Prediction, Bologna (1986)

K.A Godfrey, "Concrete Strength Record Jumps 36%, " Civil

Engineering, 84-86 (Oct., 1987).

100, J.W, Martin, "Time Transformation Functions Commonly

Used in Life Testing Analysis," Durability of Building
Materials, Vol. 1, 175-194 (1982).

101. J.J. Shideler, "Calcium Chloride in Concrete," Journal

of ACI, vol 23, 537-559, 1952).

119
 *
'^
VU '

|-
:A"’ 5
4 'i''

’ '
> S .'. . I*.
f

,i 5i 1;lli'iifjca' .S.M .Oes-:®

'
i? V.

..
•
-^r*(-. •;. -«i ir’'Lfl^ie^i V,^
.i'O
,.

/''if^i-riicilAi
,

i'sJbi is^fD V
•’'*'^

:'
k'Sr'
'

iiti I * ;; 0 .
%m
< '•i-C > \ . 1.

•
\' hM--
'-V <'•? V
..C.'-.'V

C5 ?
J,. :'.ii:".?in; .-.i- . --.v qotlari3oM AaH\at>aOte':,!(

“
.-..r imi,

• ^ » . . i A'L * '* ---

n. i: . i 7 C' .
-i'c C ^ ‘

'•'•p..
)
!'''f ^’^ijs-' r''io^^’xodAJ.'*

> v‘ •I ' * .-v;::”' » .lb r (J'‘Jo‘j''0 ict ^oq&S'j0

•
» (

'

^
» :
4 ^<’ 1 Ic:
i"i»iL'i'c '
!; '. -.:
i

,
f. /j?r. 'i
i..
^ ni

i C >; .
•
'; '

,
cS, lov ,

' r'M.'"'
'

. y v: i *> r '
‘

.
t. "Xf i
.'
» t-4
-
..'
.;^ ' pi* * *
>; ( ,! ilr, of,
’
.,«vr r r, Ovn'-' t:!<;

Vv
lS?' ' ... ifc . . .. .'_
i,'*

•7%
-•s*;'i'^'

.« ..{-?^'
'HvC' --,1 .' .* fc

o"4h* '.,'» \yiu io.li 'V .t.,.ftJSdV

‘}‘j:M
 ,

Appendix A: Barrier Systems for Protecting Concrete Surfaces

A1 . BACKGROUND

Most of the major degradation processes of concrete involve the

ingress of water or aqueous solutions into concrete. An approach

for restricting the ingress of water is to cover the exterior

concrete surfaces of a below-ground vault with a barrier

material. Also, a barrier material applied to the inside

concrete surfaces will assist in preventing nuclear waste

materials from escaping a below-ground vault. An additional

reason for applying barrier materials is to prevent movement of

unwanted materials through cracks in the concrete (Al) or through

joints, A detailed discussion of nonstructural cracks is given
in reference Al.

Recommendations for general design criteria and specific design

review criteria for the design, construction and operation of the

below-ground vault alternative method of low level radioactive

waste (LLW) disposal have been given by Denson et. al, (A2)

Moisture barriers are also treated, including applicable tests,

specifications, standards, and codes; coatings and sealers; sheet

membranes; waterstops; joint sealants; bentonite panels; and a

rationale for the moisture barrier recommendations.

Al
The information in this appendix is considered supplementary to

that by Denson et. al. (A2) . For example, waterstops,

geosynthetics, and bentonite panels are covered in reference A2,

but not in this appendix. This appendix provides information

primarily on asphalt, built-up membrane barriers for exterior

below-ground vault surfaces and protective barrier systems for
interior below-ground vault surfaces. Joint sealant materials

are also briefly covered.

Barrier materials and joint sealants are intended to prevent the

movement of liquid water and aqueous solutions and significantly

reduce the passage of vapor from water or aqueous solutions into

a building or through building elements, such as walls, roofs,

and floors.

A2 . EFFECTS OF RADIATION

The potentially detrimental effects of radiation on barrier

materials, including asphalt and its reinforcement, sealants, and

protection board may be an important factor in the selection of

barrier materials and systems. For example, asphalt, a major

component in conventional waterproofing barriers, can be

degraded by gamma irradiation (A3) The chairman of ACI

Committee 515 (Coatings for Concrete (A4)) expressed concern that

A2
(i) conventional (i.e., not designed to resist radiation)

waterproofing barrier systems and sealants may perform poorly

when exposed to radiation and, (ii) that the effects of radiation

on conventional barrier materials and sealants need to be taken

into account in selecting barrier materials and sealants.

According to the chairman, there is no published information on

the performance of waterproofing barrier systems for concrete

exposed to radiation. An improved basis for selecting barrier

materials and systems needs to be developed based on:

(i) an interpretation of published information on the effects of

radiation on individual components of the barrier systems or

materials similar to barrier materials,

(ii) analyses of field experience of the performance of barrier

materials and systems at existing facilities exposed to

radiation. Similarly, an improved basis for selecting sealants

needs to be developed based on an interpretation of published

information and field experience on the effects of radiation on

sealants.

The authors recommend, in addition to the above, that accelerated

durability testing should be carried out on (a) barrier materials

and systems and, (b) sealants to form a basis for predicting
their service lives.

A3
Another factor which should be considered is the possibility of
"self-heating" (A3) of the LLW resulting in unwanted thermal flow

of the barrier materials or sealants.

Because of the lack of published information on the performance

and service life of barrier systems and sealants for concrete

exposed to radiation and a lack of knowledge of the dosage of

radiation expected, the information presented herein is based on

the performance of conventional barrier systems and sealants in

the absence of radiation effects.

Therefore, if shown to be important, the effects of radiation, in

addition to conventional degradation factors (moisture, acid,

etc . ) treated in this report , need to be taken into account when

selecting barrier materials and sealants.

A3. TYPES OF CONVENTIONAL BARRIER SYSTEMS

Two types of conventional barrier systems for concrete surfaces

(A5) appear applicable to the 500 year design-life LLW concrete

burial vaults: waterproofing barrier systems and protective

barrier systems. A waterproofing barrier system consists of a

materials system applied to a concrete surface or structure and

is intended to exclude liquid water and aqueous solutions and to

significantly reduce the passage of vapor from water or aqueous

A4
 .

solutions. Examples of driving forces which can cause liquid or

vapor movement include hydrostatic heads, vapor gradients, and

capillary action. Porous concrete, cracks or structural defects,

or joints that are improperly designed or constructed will

increase the passage of liquids and vapors. A protective barrier

system consists of a materials system applied to a concrete

surface to protect the concrete from degradation by aqueous

solutions or chemicals and a subsequent loss in its structural

integrity.

A3.1 Waterproofing Barrier Systems

Reference A5 states that "Membrane waterproofing is the most

reliable type of barrier to prevent liquid water under a

hydrostatic head from entering an underground structure."

Waterproofing barriers have traditionally consisted of multiple
layers of bituminous-saturated felt or fabric cemented together

with hot applications of coal tar pitch or asphalt. There are

also cold-applied systems, which use multiple applications of

asphaltic mastics and glass fabrics. Other cold-applied systems

use liquids, sheet elastomeric materials, and preformed

"rubberized bituminous" sheets (A5) References A5 and A6 give

the details for the various hot-applied and cold-applied systems,

including materials used, and factors relating to the

interaction of the waterproofing barrier and the concrete surface

A5
(surface preparation, adhesion, water accumulation at the

interface, etc.)* it is essential that there be good adhesion

between the barrier membrane and the concrete to prevent water

migration and leakage at the barrier-concrete interface due to
unwanted cuts or openings in the barrier material.

The bitumens (asphalt or coal-tar pitch) used in hot-applied

systems have very little strength themselves. They, therefore,

need to be reinforced with fabrics or felts to withstand the

stresses caused by temperature changes, expansion, contraction,

vibration, and movement of the structure. Similarly, the mastic

and emulsions used in the cold-applied systems do not have the

necessary strength by themselves and need to be reinforced with

fabric.

A3, 2. Protective Barrier Systems

Protective barrier systems can be used to protect the concrete in

the below-ground vault from degradation by aqueous solutions and

chemicals from the LLW, which could permit the passage of LLW

through the concrete. At the ambient service temperature,

protective barrier materials should have the following basic

properties to be effective in protecting the concrete from
chemicals (A5)

(a) the protective barrier material should not swell, dissolve,

crack or embrittle upon contact with pertinent liquids or vapors;

A6
 .

(b) the barrier material should prevent the permeation or

diffusion of chemicals which are able to cause a loss of adhesion

between the barrier and the concrete; and

(c) the barrier material should have sufficient abrasion

resistance to prevent it from being damaged during placement and

maneuvering of the LLW containers in the vault. Brick sheathing
applied over the barrier material (e.g,^ built-up asphalt

membrane) is considered necessary to help prevent damage to the

barrier material in those areas where barrier contact and

potential damage is anticipated. For example, damage to the

floor and perhaps the walls could occur due to cutting forces

from mechanical action or cold-flow from sustained loading.

Reference A5 provides additional information, including factors

which affect the adhesion of the protective barrier to the
concrete, and also the effects of the concrete structure (e.g.,

cracks in the concrete reflecting through the barrier) and

foundation movements on the performance of the protective

barriers

A4. DEGRADATION FACTORS AND SERVICE LIFE PREDICTION OF BARRIER

MATERIALS

Various degradation factors can affect the ability of the barrier

(waterproofing and protective) materials and sealant materials to

A7
perform properly over the required design life of 500 years. The

factors anticipated include chemical attack; oxygen, if it is

available underground; moisture; alkalies; acids; biological

species (e.g. microorganisms, insects, and plant and tree roots);
,

radiation (from LLW) ; structural movement, including settlement,

which induces strains; sustained loading from the LLW, which can
cause cold-flow of the barrier material beneath the LLW; and

cutting forces occurring during placement of the LLW into the

below-ground vault. The effects of these degradation factors

would depend on their severity and their duration of occurrence

during the life of the structures.

In addition, prior to burial, the barrier materials and sealant

materials may be exposed to ultraviolet light, ozone, temperature

cycling, increased temperature (e.g., caused by sunlight, which

could result in unwanted flow in an asphalt barrier) and other

forms of weathering. The effect of these degradation factors

would depend on the severity and duration of exposure prior to

burial

There are a number of reports which provide information on the

effects of various conventional degradation factors, as well as

radiation, on the properties of polymers and bituminous and

asphaltic materials (see, for example, references A3 and A7 to

A17) ,

A8
ASTM E 632 (A18) , can be used as a guide in developing a

methodology for predicting the service life of the barrier

materials and sealant materials exposed to various degradation
factors as described above. Accelerated tests may need to be

used but care must be taken to insure that the degradation

mechanisms induced by the tests and the in-service environment

are the same. Appropriate physical and mechanical properties

which are sensitive to changes in the performance of the barrier

materials and sealant materials when exposed to the anticipated
environment need to be identified. For example, moisture
permeability could be one of the most important properties.

Consideration should be given to moisture permeability tests

which simulate, as closely as possible, the in-service

conditions. For example, with a built-up membrane barrier, the

configuration of the membrane (plies etc.) and how it would be

applied to the concrete surface should be simulated by the tests,

as close as possible.

A possible source of information on the performance and

durability of barrier materials and sealant materials would be to

examine the condition of waterproofing barriers and sealants of

older buried structures for which records are available on their

age and on the types of materials used. Examples of potential

structures include facilities exposed to radiation, cemetery

burial vaults, basement walls of residences, foundations of

buildings and structures, underground storage structures, and

A9
transportation facilities. Similarly, the performance and

durability of protective barrier materials and sealant materials

in existing structures may provide information useful in

predicting their service lives,

A4.1 Conventional Waterproofing and Protective Barrier Materials

Potential conventional barrier materials are discussed based on

limited existing information, excluding the potentially

detrimental effects of radiation (see Sections A2 and A5)

Long term performance is considered to be the most important

criterion for selecting candidate materials for below-ground

vaults. Hence, materials with a history of performance of 500

years or more are preferred. The historical use of bitiiminous

materials has been documented (see below). Before World War II,

most of the commercial polymeric materials occurred naturally

(A19) . Hence, the use of synthetic polymers has no long-term

performance records relative to the required 500 year design

life. Even the bituminous materials do not have detailed

performance records concerning their physical and mechanical

properties, which would verify that they would perform properly

under specified degradation factors for 500 years. None-the-

less, because they have been doAUuaented historically, asphalt

built-up membranes should be considered for possible use as a

waterproofing barrier on the exterior of the below-ground vault.

AlO
A protective barrier system applied to the inside of the below-
ground vault, consisting of an asphalt built-up membrane covered
by brick sheathing, should also be considered, provided the

asphalt membrane has adequate chemical resistance o (Regardless

of the membrane, brick sheathing applied over the protective

barrier membrane material is considered necessary to help prevent

damage to the barrier material in those areas where barrier

contact and potential damage is anticipated - see Section A3. 2),

If an asphalt membrane (applied on the interior of the below-

ground vault) does not have adequate resistance against chemical

attack resulting from the LLW inside the below-ground vault, then
other types of protective barrier materials and systems need to

be considered. For example, if concentrated acids or

acid/solvent combinations are anticipated from the LLW, use of an

asphalt membrane (other membranes may be used depending on the

chemical environment) combined with masonry construction

consisting of "acid-proof" brick using a chemical-resistant

mortar (A5) should be considered.

Additional discussion of the design and selection of chemically

resistant masonry materials is given in references A20 and A21.

For conventional barriers, fiberglass fabric or felt

reinforcement in the asphalt membrane is believed to be more

durable than other reinforcement, provided the fiberglass binders

are water resistant.

All
Barrier systems containing asphalt are believed to be candidate

conventional systems because of the documented longevity of

asphalt for a variety of ancient applications (A22) . For

example, Abraham (A22) states with regard to a Pleistocene

Cypress tree trunk recovered from the Rancho-la-Brea asphalt pits

in California and preserved by asphalt for 25000 years? "This

represents one of the oldest specimens of wood in existence and

bears mute evidence of the remarkable preservation properties of

asphalt." Abraham documents a number of cases of the use of

asphalt mortar or asphalt mastic in ancient times. For example,

with regard to the use of asphalt with bricks (about 2200 B.C.),
Abraihcun quotes: "Robert Koldewey reports that when excavating in

Babylon, he found that ' it was exceedingly difficult to separate

the brick courses from each other. ' This was also confirmed by

Ao H. Layard, who states that 'Bricks bonded with asphalt have

remained immovably in place for thousands of years.'" Abraham

also states, with regard to King Nebuchadnezzar (604 to 561

B.C.), that burnt bricks bearing the name of Nebuchadnezzar "were

laid in asphalt, and are still so firmly jointed together today,

that it is impossible to part them without destroying their

integrity.

A4.2 Conventional Sealant Materials

Use of joints should be minimized as the performance of joint-

sealant materials is often a troublesome problem (A23, A24, A25)

A12
 . , . .

Joint design, sealant selection, and installation are treated in

references A23 and A24. Conventional, synthetic, polymeric

sealants have an estimated maximum service life of about 20 to 40

years (about 20 to 30 years for urethanes and polysulfides and

30 to 40 years for high-modulus silicones (A26) ) High-modulus

silicones have been used since the 1950 's and may have service

lives in excess of 40 years. These high-modulus silicones have

very little plasticizer to migrate out and cause hardening (A26)

Medium-modulus silicones have been used for about 25 years and

are currently replacing the high-modulus silicones because of the

greater movement capability of the medium-modulus silicones

(A26) The effect of radiation from LLW on the service life of

the sealants also needs to be considered (Sections A2 and A5)

Thus, it appears that the sealant materials could be at least as

critical as the waterproofing barrier materials, since the

synthetic, polymeric sealants have a documented service life of

only about 20 to 40 years compared to the required service life

of 500 years.

A4c3 Design Considerations for Conventional Materials

To prevent their failure, the interior and exterior barriers of

the below-ground vault must be designed to accommodate movement

at the joints. Furthermore, if for any reason an exterior or

interior barrier fails, liquid or vapor could migrate at the

A13
barrier-concrete interface until reaching a point of low
resistance to passage, which most likely would be at joints (and
their sealant materials ) or at unwanted cracks in the concrete.

Thus, if a barrier system fails, the sealant materials could then

become a critical element in the prevention of unwanted liquids

and vapors moving through the vault. Therefore, (i) the use of

joints should be minimized, (ii) penetrations through the

concrete or the barrier systems or both should be eliminated,

(iii) proper design is needed to insure that the barriers are
able to move (rather than fail) as the joints move and, (iv)

systems with redundant waterproofing barrier systems or redundant

concrete elements or both should be considered in designing an

impervious structure.

The use of protection board covering should be considered to

protect the exterior barrier from being damaged (e.g., when

performing work on, near, or over the barrier, and when

backfilling the barrier; see Part 1, IV, A. "Protection Courses,"

reference A6) In addition, the use of brick sheathing should

also be considered to protect the interior protective barrier

system in those areas where barrier contact and damage is

anticipated (see Section A3. 2).

A14
A5 . RESEARCH NEEDS

Very little is known about the long-term service life of barrier

materials and systems and sealant materials, including the
effects of radiation. A methodology, using ASTM E 632 (A18) as a
guide, needs to be developed to predict the service lives of

potential barrier systems and sealant materials, when exposed to

conventional degradation factors (moisture, acids, etc,) as well

as radiation.

The extent of potential degradation by radiation of barrier

systems and sealant materials depends on the dosage of the

radiation. The dosage of radiation to which barrier materials

are likely to be exposed should be determined,

A6. ACKNOWLEDGEMENTS

In preparing this appendix, the authors are grateful to W. C,

Cullen, K. Heffner, R. Janicki, D. Lawrence, R. G, Mathey, J.

Panek, W, R. Rossiter, and J, Von Wagoner who provided invaluable

information and assistance.

A7. REFERENCES

Al. "Non-Structural Cracks in Concrete,” Concrete Society

Technical Report No. 22, The Concrete Society, Wexham
Springs, Slough, U.K,, 1982.

A15
A2. R. H. Denson et. al., "Recommendations to the NRC for Review
Criteria for Alternative Methods of Low-Level Radioactive
Waste Disposal" Task 2a: Below-ground vaults, U. S. Army,
WES, NUREG/CR-5041, Vol. 1, Dec. 1987.

A3. A. J. Hoiberg (editor). Bituminous Materials: Asphalts,

Tars, and Pitches, Volume 1, Ch. 6 "Irradiation of
Bituminous Materials," by C. D. Watson and W. W. Parkinson,"
Interscience Publishers, N.Y., 1964.

A4. Private communication with D. Lawrence, Chairman of ACI 515,

Coatings for Concrete, December, 1988.
A5. ACI Report 515.1R-79, "A Guide to the Use of Waterproofing,
Dampproofing, Protective, and Decorative Barrier Systems for
Concrete," American Concrete Institute, Detroit, MI, 1979.

A6c NRCA Roofing and Waterproofing Manual, "Waterproofing and

Dampproofing," The National Roofing Contractors Association,
Rosemont, IL, Third Edition, 1989.

A7, A.Jc Hoiberg (editor). Bituminous Materials: Asphalts, Tars,

and Pitches, Volume 2, Part 1, Ch. 8, "Weathering:
Theoretical and Practical Aspects of Asphalt Durability," by
J. R. Wright, Interscience Publishers, N.Y., 1965.

A8. A. J. Hoiberg (editor). Bituminous Materials: Asphalts,

Tars, and Pitches, Volume 1, Ch. 7 "Microbial Action on
Bituminous Materials," by R. W. Traxler, Interscience
Publishers, N.Y., 1964.

A9. R. G. Mathey and W. R. Rossiter, Jr., "Properties of

Weathered Uncoated and Resaturant -Coated Bituminous Built-
'
*

Up Roofing Membranes," Durability of Building Materials, V.

2, 1983, pp. 59-78.

AlO. M. C. Baker, Roofs, Design, Application and Maintenance,

National Research Council of Canada, Multiscience
Publications Limited, Montreal, Quebec, Canada, 1980.

All. R. W. King , N. J. Broadway, and S. Palinchak, "The Effect

of Nuclear Radiation on Elastomeric and Plastic Components
and Materials," REIC Report No. 21, Radiation Effects
Information Center, Battelle Memorial Institute, Columbus
OH, September 1, 1961.

A12. A. D. Jenkins (editor). Polymer Science, Vol. 2, Ch. 23, A.

Charlesby, "Radiation Effects in Polymers," North Holland
Publishing Company, N.Y,, 1972.

A16
A13. H. H. G. Jellinek (editor), Degradation and Stabilization of
Polymers, Voliime 1, Ch. 8, "High Energy Degradation and
Stabilization of Polymers," Elsevier, N.Y.,N.Y., 1983.

A14. J. F. Kircher and R. E. Bowman (editors). Effects of

Radiation on Materials and Components, Reinhold Publishing
Corporation, N.Y., N.Y., 1964.

A15. K. T. Gillen and R. L. Clough, "Occurrence and Implications

of Radiation Dose-Rate Effects for Material Aging Studies,"
Sandia National Laboratories, prepared for U. S. Nuclear
Regulatory Commission, Report NUREG/CR-2157 , SAND80-1796,
RV, June 1981.

A16. W. Schnabel, Polymer Degradation, Macmillan Piiblishing Co.,

Inc., N.Y., 1981.

A17. A. Davis and D. Sims, Weathering of Polymers, Applied

Science Publishers, N.Y., 1983.

A18. ASTM E 632-82, Standard Practice for Developing

"
Accelerated Tests to Aid Prediction of the Service Life of
Building Components and Materials", ASTM, Philadelphia, PA,
1988.

A19. R, B. Seymour, Modem Plastics Technology, Reston Publishing

Co., Reston VA, 1975.

A20. W. L. Sheppard Jr., Chemically Resistant Masonry, (Second

Edition), Marcel Dekker, Inc., N.Y., 1982.

A21. W. L. Sheppard, Jr. (Editor), Corrosion and Chemical

Resistant Masonry Materials Handbook, Noyes Publications,
Park Ridge, N. J. 1986.
,

A22. H. Abraham, Asphalts and Allied Substances, Volume One,

"Historical Review and Natural Raw Materials", D. Van
Nostrand Company, Princeton N.J., 1960.

A23. ACI 504R-77, "Guide to Joint Sealants for Concrete

Structures," American Concrete Institute, 1977.

A24. J. R. Panek and J. P. Cook, Construction Sealants and

Adhesives, Second Edition, John Wiley and Sons, N.Y., 1984.

A25. L, I. Knab, "Factors Related to the Performance of

Concrete," U. S. Army Corps of Engineers, WES Technical
Report REMR-CS-12, March, 1988.
A26, Private communication with Jules Panek, January, 1989,

A17
 n,

I;

'
V )' V.- !,c V * t
.
^ ' ? -’'j
A''

:
,.' '.c'j^' 't,'*.' ' ^ ''‘^^:'5 ’’"d ^

•' * • '
•
•^r liS •.
.:rj?‘',4
<

y i.

. ,
--. J uaAV?*"^. xp.
yaX
©nTO
'

*r c.-.: o’.;..'/-;cj'V7.4 v’C^t/; ;.i<'.- 'iiw'''

"’WJ^

' -'V'V • .«' iy»

-
:'‘C

^
.
, .
f
. e A- 1 :-;,
'
VI ^ X jTal;^ m
«

V
T *•

#
A

‘V- I
.,' ,-i«aj!aa , (noiiifiS
-cj

JK^I
i> .. :•;!>.

'i.'iW'S'
4
l' |P-

• l'T, -yr^viv^:"

'
jt,,
’
'
y- ,;,r: i.r.oi^ 44 i

’‘ 1 - ' '
,
•'
'
vv 4 a«>.J^ o:*

H-
¥ ... a <1001: ,,ao.r:Jl 5 ;f BfiooSn i

.yi.'.' V .'•njM-‘;;7 y., H/y...,

>?'-?! .sifinsg a)»XuL- noi/?«ioinii?i^ato» :^il^

'tjf;
? 4:Vr
A 16 •.'tt.'j

CXA

'
M
 :

APPENDIX B. SOME COMMON CHEMICALS AND SUBSTANCES AGGRESSIVE TO

PORTLAND CEMENT CONCRETE

Table Bl. Inorganic Chemicals and Substances^

Distilled water

Soft water

Acids: Bromic acid Carbonic acid

Hydrochloric acid Chromic acid
Hydroflouric acid Hydrogen sulfide
Nitric acid Phosphoric acid
Sulfuric acid Sulfurous acid

Bases: Sodium hydroxide Potassium hydroxide

Salts: Chlorides (Ca, Na, NH^, Mg)^

Sulfates (Al, Ca, Fe, Cu, K, Mg, Na, NH^)
Nitrates and nitrites (Na, K, NH^)
Sulfides (Fe, Na)
Sulfites and bisulfites (Na)
Acetates (Ca, Al, NH^)

Gases: Carbon dioxide

Chlorine
Sulfur dioxide

Others Sodium thiosulfate

Sodium borate (borax) - small effect.

^Durability of Concrete Structures, CEM-RILEM International

Workshop, May 1983.

^FoM. Lea, The Chemistry of Cement and Concrete, 3rd ed.

(Chemical Publishing Co., N.Y,, 1970).

^Common cations.

Bl
 Table B2 . Organic Chemicals and Substances 1,2

Acids: Acetic acid Butyric acid

Citric acid Formic acid
Humic acid Lactic acid
Malic acid Tannic acid
Uric acid
Alcohols: Ethanol Methanol
Phenol Propanol
Glycerol (glycerine)
Amines
Esters

Stearates: Glyceryl sterate (stearin)

Sugar solutions

Vegetable oil and animal oil “ contain fatty acids:

Oleic acid
Stearic acid
Palmitic acid

^Durability of Concrete Structures, CEB-RILEM International

Workshop, May 1983.

^F.M. Lea, The Chemistry of Cement and Concrete, 3rd ed.

(Chemical Publishing Co., N.Y., 1970).
 NBS-n4A (REV. 2-8C1
U.S. DEPT. OF COMM. 1. PUBLICATION OR 2. Performing Organ. Report No. 3. Publication Date
REPORT NO.
BIBLIOGRAPHIC DATA
SHEET (See instructions) NISTIR 89-4086 JUNE 1989
4. TITLE AND SUBTITLE

SERVICE LIFE OF CONCRETE

5. AUTHOR(S)
James R. Clifton, Lawrence I. Knab
9.

6. PERFORMING ORGANIZATION (If joint or other than NBS, see instructions) 7. Contract/Grant No.

NATIONAL BUREAU OF STANDARDS

8. Type of Report & Period Covered
U.S. DEPARTMENT OF COMMERCE
GAITHERSBURG, MD 20899
10.
SPONSORING ORGANIZATION NAME AND COMPLETE ADDRESS (Street, City, Store, ZIP)
U.S. Nuclear Regulatory Commission
Washington, DC 20555

SUPPLEMENTARY NOTES

I i
Document describes a computer program; SF-185, FIPS Software Summary, is attached.
11. ABSTRACT (A 200-word or less factual summary of most significant information. If document includes a significant
bibliography or literature survey, mention it here)
U.S. Nuclear Regulatory Commission (NRC)
has
the responsibility for developing a strategy for the disposal of low-level radioactive
waste (LLW) An approach being considered for their disposal is to place the waste forms
.

in concrete vaults buried in the earth. A service life of 500 years is required for the
concrete vaults as they may be left unattended for much of their lives.
This report examines the basis for making service life predictions based on acceleratec
testing and mathematical modeling of factors controlling the durability of concrete
buried in. the ground. Degradation processes are analyzed based on considerations of
their occurrence, extent of potential damage, and mechanisms. A recommended research
plan for developing methods for predicting the service life of concrete is presented.
The major degradation processes that concrete of underground vaults will likely
encounter are sulfate attack, corrosion of reinforcing steel, alkali-aggregate reactions
and leaching by ground water. Freezing and thawing damage could occur before the vaults
4re covered with soil and therefore are addressed. Other degradation processes which may
occur are microbiological attack, salt crystallization, and attack by LLW, especially
by acidic materials. Two important factors controlling the resistance of concrete to
these degradation processes are its quality and permeability. Concepts of quality and
factors affecting quality of concrete are discussed. Permeability is discussed in terms
of the water-to-cement ratio, the pore structure of concrete, and the effects of cracks.

12. KEY WORDS (S/x to twelve entries; alphabetical order; capitalize only proper names; and separate key words by semicolons)
Accelerated testing; alkali-aggregate reactions; concrete; corrosion; durability;
groundwater; mathematical modeling; permeability; service life; sulfate attack

13. AVAILABILITY
14. NO. OF
PRINTED PAGES
[^U nl imited
I
I
For Official Distribution. Do Not Release to NTIS
151
I I
Order From Superintendent of Documents, U.S. Government Printing Office, Washington, D.C.
20402. 15. Price

[X] Order From National Technical Information Service (NTIS). Springfield, VA. 22161 $19.95

USCOMM-OC 8043-P80
 •ill
'
' ^>l'*A<l)>( fi^»«»|iii W im*»>i^ « ,jL. UK*(!»lrawMtni» mZ'i
- u-i'. a f;'. or

-
y'f^r^OTW-^rnTlS^TSf^i
^ ' ;' '

.'•r ' »A«»4-n »^< —— •mmrn - -T 1l TT --if

•'
:
-
iVv

r, 'Ti^Ts m jmmA
”
'
- ' V- ,.,,i? '-is,. VjO"> j rmmuMmi 4*^
'f '*‘
55?' ir?r"
, 1.1 V-- r' 4 . ' r<

r '/ X c ^ '
-itaoD -Y '.; r.injfaP
'

.'S>'tfAi^

aiiJi jO ,aos. 3MimMl

17 •, 5;i)T 0 <^ YtATi

I
.• }.'. i-. ',••/ 1 1 ' fr ",.< ''.'Jj. \
'
•*Q U>jpW*^tlS. A, .
^

^ • - ' " ^''= (' -O'ft ?ir »tJ

‘.-f; 1
X
’*••*
••.'.,1 .
,5 n uj-' jq!| 73 , 1 IJ: di
*-'''' - -I . s'7 ;t,’,j7i:.'> 'i^', ..^j<j 5j:i«07Wjs. na .

ij-.t'CS . T-/ j <:rfj ^

"’'f? e:S,Lii£V A.

f'o
.'
2 '' 1
:: . .xq .xioxJ&.-vv, 1 riQ ,i>fjyo 35 #fl .7

.' 'jr xa 'jT.ci^ ^ !.u'.i3 ’rfi4;i:;sy , ' .

noijPib/BT^a-to
*•'. 3’. '
!}•. j
v'.CaXJ'O . -i.jiirrJc,

;' .>’>,'Ij ,
i
^^si.'i-.
.Y<f
.

't-L.'. -. . : ., i<j ,' :ij 2 jS Lioi?’ rfjlW

- ••• : 3 .-i^y
'

I ? ". .
i 2 ' qsd- c»wT . ei & 1
•'.: ;?ii j.ji.uD zix 1
'
~
,

i ; 2ay "^ i-m •.•.> 30 v-fX ],«*»;•

• . < \ 'A*.
; 'I '-U.,/ 'i:?.e j-ioq '>r|7 f0.l.-ji-,' tXT‘-Kr!ai»*-a3’*'t^ '

'OI'W ., 5 nT -
<*'’« 'sUw,- '-. !(>&•••. ' ;. r'Afjvrvil* Sr .;r>s

'07. .-'i' i lii'o.L e3*3!&35gfc-3i.fcjUet xgftXiisis'li;

•*'> ‘
T
•
5 i>..= VX -, t
J V • |j:i.„j>lkGiftr

'
^»ti» i»

> : .. , ''dM
B?. ^

' •'•'
-v
''
.

A
- V

v"jȣ
•™^i
''itSX '
'ijs, teno*|*»*
'••'
'5*1^

'• -

.
''"'
jL ,. '
jAl-i