Reproduction of delayed ettringite formation (DEF) in concrete and

Fracture Mechanics of Concrete and Concrete Structures Assessment, Durability, Monitoring and Retrofitting of Concrete Structures- B. H. Oh, et al. (eds)
ⓒ 2010 Korea Concrete Institute, Seoul, ISBN 978-89-5708-181-5
Reproduction of delayed ettringite formation (DEF) in concrete and
relationship between DEF and alkali silica reaction
S. Hanehara & T. Oyamada
Iwate University, Morioka, Japan
ABSTRACT: Delayed ettringite formation (DEF) is a phenomenon in which ettringite is generated and accumulates in the concrete after hardening, eventually leading to expansion and destruction of the concrete. In
this study, DEF expansion was reproduced in laboratory using a variety of cements, sulfate additions, steam
curing conditions and storing conditions. DEF was found to take place only when the three conditions of excessive sulfate, high-temperature steam curing and ample water supply are met simultaneously. The influence
of the kind of calcium sulfate (gypsum, hemihydrate or anhydrite CaSO4) and the heating rate of steam curing
was studied. As it is reported that both the alkali silica reaction (ASR) and DEF are sometimes observed in
the same deteriorated concrete, the relationship between the alkali silica reaction and DEF is also discussed.
1 INTRODUCTION
Delayed ettringite formation (DEF) is a phenomenon
in which ettringite is generated and accumulates in
concrete after hardening, eventually leading to expansion and destruction of the concrete. The deterioration of concrete due to this phenomenon was reported in 1990 and later in the United States and
Europe (Taylor et al. 2001, Shimada 2005). It is reported that this phenomenon is liable to appear in
concretes with a high sulfate content and subjected
to steam curing (Hanehara 2003, Shimada 2005).
Moreover, DEF takes place only when the three
conditions of excessive sulfate, high-temperature
steam curing and ample water supply coincide. In
their previous reports (Hanehara et al. 2006a, 2006b,
2008a, 2008b), the authors confirmed that a steam
curing temperature of not less than 70°C is required
for the occurrence of sulfate expansion due to DEF,
and that water curing at not less than 20°C is necessary after the steam curing.
This paper examines in greater detail the material conditions affecting expansion by DEF, further
clarifies the conditions for occurrence of DEF, and
explains the mechanism of occurrence thereof with
respect to the influences of cement type and aggregate type. The results of curing in the DEF acceleration test and mortar bar test method using alkali reactive aggregates were observed and the
relationships between the two compared. The relationship between the alkali silica reaction and DEF
is also discussed.
2 DEF REPRODUCTION CONDITIONS
A schematic chart summarizing the authors’ previous works (Hanehara et al. 2006a, 2006b, 2008a
2008b) is shown in Figure 1. It can be seen that DEF
takes place only when the three conditions of presence of excessive sulfate in the cement, hightemperature steam curing at 70°C or over, and ample
water supply at normal temperature coexist. Conversely, DEF does not take place if any of these
conditions is missing. Based on the conditions for
occurrence of DEF, we proceed to a detailed study
of material conditions below.
Figure 1. Occurrence condition ( DEF).
2.1 Influence of temperature of steam curing
To identify the influences of the steam curing conditions, we added 2% potassium sulfate as SO3, maintained the pre-curing duration of 4 hours, used various steam curing temperatures in the range of 50°C
= − Dquantity
( h, T )∇h of expansion diminaddition of 0.5%,J the
ishes to 0.04% at the curing age of 350 days.
The proportionality coefficient D(h,T)
moisture permeability and it is a nonlinea
of the relative humidity h andPlain
temperature
& Najjar 1972). The moisture+0.5%
mass balanc
+1.0%water mas
that the variation in time of the
+1.5%
volume of concrete (water content
+2.0% w) be eq
+2.5%
divergence of the moisture flux
J
+3.0%
6.0
5.0
Expansion(%)
to 90°C, and prepared the mortar samples using
high-early-strength cement. Figure 2 indicates the
changes in length of the mortar prisms.
When steam curing is done at a temperature no
less than 70°C, expansion due to DEF occurs from
around the curing age of 30 days to about 110 days.
The higher the curing temperature, the earlier the
expansion takes place, but the ultimate expansion
amount is comparable, at about 2.5%. On the other
hand, with steam curing at a temperature below
60°C, there is minimal expansion due to DEF as is
the case when curing takes place at normal temperatures without steam curing. The boundary temperature conditions for occurrence of DEF are between
60°C and 70°C.
4.0
3.0
2.0
1.0
0.0
t
0
1.5
1.0
0.5
5.0
Expansion(%)
Expansion(%)
90℃
80℃
70℃
60℃
55℃
50℃
20℃
4.0
3.0
2.0
1.0
0.0
0
50
100
150
Age(days)
200
250
Figure 2. Influence of steam curing temperature on DEF expansion.
2.2 Influence of sulfate content
Potassium sulfate was added in the range of 0% to
4.0% and calcium sulfate in the range of 0% to
3.0%. The changes in length of the mortar prisms are
given in Figure 3 and Figure 4, respectively.
Although the cement does not show any expansion when no sulfate is added, the greater the added
quantity, the earlier the onset of expansion during
curing and the larger the amount of expansion are.
This is true for any addition of sulfate, regardless of
its type. In the case where potassium sulfate is
added, expansion in excess of 0.1%, which is the criterion of the alkali-silica reaction after 6 months curing was observed at a curing age of around 100 days,
even with the minimum sulfate addition of 0.5%.
The final quantity of expansion was greater than
1.0%. With potassium sulfate additions of 1.5% or
over, expansion starts at a curing age of around 30
days. With potassium sulfate addition of 4%, an expansion of 6% was observed at the curing age of 150
days and the mortar was fractured. Thus DEF may
occur even with a small quantity of sulfate.
In the case of calcium sulfate addition of 1.5% or
more, the mortar shows a large expansion of no less
than 1.5% at the curing age of 300 days. Even with a
calcium sulfate addition of 1%, expansion was 0.4%
at the curing age of 350 days. With a calcium sulfate
100
200
Age(days)
300
The water content w can be expressed a
wa
vapor, and adsorbed water) and the non-e
(chemically bound) water wn (Mil
Pantazopoulo & Mills 1995). It is reas
assume that the evaporable Plain
water is a fu
+0.5%
relative humidity, h, degree+1.0%
of hydration
degree of silica fume reaction,+1.5%
αs, i.e. we=w
+2.0%
= age-dependent sorption/desorption
+2.5%
+3.0%this assum
(Norling Mjonell 1997). Under
by substituting Equation 1 +4.0%
into Equati
obtains
Figure 3. Influence of amount of potassium sulfate on DEF exof the evaporable water we (capillary
pansion.
6.0
2.0
+4.0%
w
3.0
2.5
− ∂ = ∇•J
∂
0.0
∂w
∂wAge(days)
∂h
e α ∂we
+ ∇ • ( D ∇h ) =
− e
Figure 4. Influence of amount
α c ∂on
α
∂h ∂tof calciumhsulfate ∂(anhydrite)
c
s
DEF expansion.
0
100
200
300
&+
α&s + w
/∂h is the ofslope
of thethe
sorption/
where
∂wequantity
The larger the
added
sulfate,
isotherm
(alsoMoreover,
called even
moisture
greater the expansion
volume.
with capac
governing
(Equation
must be
similar added quantities
of equation
sulfate, the
onset of3)exby
appropriate
boundary
and
initial
pansion occurs at a later curing age in the case of the conditi
The compared
relation between
thetheamount
calcium sulfate addition
to that of
po- of e
water and
called ‘‘
tassium sulfate addition,
and relative
the finalhumidity
amount ofis exisotherm”
if
measured
with
increasing
pansion is also smaller. This is probably because the
and ‘‘desorption
isotherm”
alkali ions of the humidity
alkaline sulfate
decrease the
stabil- in th
Neglecting
their
difference
(Xi et al.
ity of ettringite atcase.
the time
of steam
curing,
decomthe following,
‘‘sorption
isotherm”
posing a larger quantity
of ettringite
and,
for thatwill be
reference
to bothofsorption
desorption c
reason, subsequent
regeneration
a largeand
quantity
By
the
way,
if
the
hysteresis
of ettringite (DEF) results in a large expansion at anof the
would
takensteep
into account,
two
early time. The isotherm
heating rate
wasbevery
from
evaporable
room temperaturerelation,
to maximum
steam water
curing.vs relative humi
be used according to the sign of the varia
relativity humidity. The shape of the
2.3 Influence of environmental
andisstoring
isotherm for HPC
influenced by many p
conditions ofespecially
concrete those that influence extent and
and, in turn,
To investigate thechemical
effects ofreactions
storing conditions
after determ
structure
and
pore
size
distribution
steam curing, mortar samples were prepared using (waterratio,
cement
composition, SF
high-early-strength
cement
withchemical
2% of potassium
curing
andcured
method,
temperature,
sulfate as SO3 and
then time
steam
at 90°C.
After mix
etc.).
In
the
literature
various
steam curing, the samples were stored in waterformulatio
at
foundand
to indescribe
theatsorption
isotherm
20°C, 40°C and 60°C
humid air
40°C. Figet al. 1994).
ure 5 shows theconcrete
changes(Xi
in length
of theHowever,
mortar in th
paper
the semi-empirical
expression
pro
samples. In the case
of storage
in water at 20°C
, exNorling
Mjornell
(1997)
is
adopted
b
pansion started from around the curing age of 30
Proceedings of FraMCoS-7, May 23-28, 2010
J = −D
( h, Treached
) ∇h
days
and
2.5%. No expansion took place
(1)
during water storage at 40°C or 60°C, or in humid
air The
at 40°C.
For the production
of ettringite,
temproportionality
coefficient
D(h,T) is acalled
perature
of
around
20°C,
which
is
a
low
temperamoisture permeability and it is a nonlinear function
ture,
suitable.humidity
Furthermore,
cured Tin(Bažant
humid
of theisrelative
h and mortar
temperature
air
at
20°C
shows
rather
little
expansion
even
after
& Najjar 1972). The moisture mass balance requires
100
days.
In
the
production
of
ettringite,
which
is a
that the variation in time of the water mass per unit
dissolution-precipitation
reaction,
the
reaction
may
volume of concrete (water content w) be equal to the
be
considered
through
a liquid phase.
divergence
of as
theprogressing
moisture flux
J
An attempt will be made in this paper to put the
conditions
for occurrence of DEF in order, based on
(2)
− ∂wprevious
= ∇ • J reports.
our
∂t
Expansion(%)
3.0 water content w can be expressed as the sum
The
of the
2.5 evaporable water we (capillary water, water
vapor,
and adsorbed water) and the non-evaporable
2.0
20℃ in water
1966,
(chemically
bound) water wn (Mills
40℃ in water to
Pantazopoulo
& Mills 1995). It is reasonable
1.5
in water
assume
that the evaporable water is a 60℃
function
40℃
in wet air of
1.0
relative humidity, h, degree of hydration, αc, and
0.5 of silica fume reaction, αs, i.e. we=we(h,αc,αs)
degree
= 0.0age-dependent sorption/desorption isotherm
(Norling0 Mjonell
1997).
and
50
100
150 Under
200 this
250 assumption
300
Age(days)
by substituting Equation 1 into Equation 2 one
Figure
obtains5. Influence of storing condition of haredened cement
mortar after steam curing on DEF expansion.
∂w ∂h
e + ∇ • ( D ∇h) = ∂we α ∂we α
−
c and conditions
s
n
2.4 ∂hInfluence
of hother factors
∂α
∂α
∂t
DEF
c
&+
s
& + w&
on
(3)
The
type
cement has a great influence on DEF. In
where
∂wof
e/∂h is the slope of the sorption/desorption
the
case
of
Portland
cement,
the curing
age at which
isotherm (also called
moisture
capacity).
The
expansion
starts
is
earliest,
and
the
expansion
ratio
governing equation (Equation 3) must be completed
for
high-early-strength
(HPC) is
by appropriate
boundary Portland
and initialcement
conditions.
largest,
followedbetween
by whitethePortland
(WPC)
The relation
amountcement
of evaporable
and
ordinary
Portland
cement
(OPC).
A
larger
exwater and relative humidity is called ‘‘adsorption
pansion
ratio
is
shown
by
cements
with
a
large
volisotherm” if measured with increasing relativity
A and
a high level
of ettringite
ume
of C3and
humidity
‘‘desorption
isotherm”
in thegeneration
opposite
before
steam
curing,
while
blended
cements
with in
a
case. Neglecting their difference (Xi et al. 1994),
A,
such
as
moderate
heat
cement,
small
volume
of
C
3
the following, ‘‘sorption isotherm” will be used with
fly-ash
blast-furnace
cement,
have a
referencecement,
to both and
sorption
and desorption
conditions.
small
expansion
ratio
due
to
DEF.
By the way, if the hysteresis of the moisture
Expansion
duebetotaken
DEFinto
is affected
thedifferent
type of
isotherm
would
account,bytwo
aggregate.
Expansion
starts
earlier
and
the
expansion
relation, evaporable water vs relative humidity, must
volume
greater when
aggregate
used,
be used isaccording
to thequartzite
sign of the
variationis of
the
compared
to
when
reactive
aggregates
are
used.
relativity humidity. The shape of the sorption
Limestone
starts expanding
at parameters,
a still later
isotherm foraggregate
HPC is influenced
by many
curing
age
and
its
expansion
volume
is
small.
especially those that influence extent and rate of the
chemical reactions and, in turn, determine pore
structure and pore size distribution (water-to-cement
3ratio,
EXPERIMENTS
cement chemical composition, SF content,
curing time and method, temperature, mix additives,
etc.).Materials
In the literature various formulations can be
3.1
found to describe the sorption isotherm of normal
High-early-strength
cement (HPC)
concrete (Xi et al. Portland
1994). However,
in theobtained
present
on
the
Japanese
market
was
used.
It
has
a density by
of
paper the3 semi-empirical expression proposed
3.14
g/cm
and
Blaine
specific
surface
area
of
4450
Norling
Mjornell (1997) is adopted because it
2
cm /g. As sulfate, potassium sulfate, gypsum, hemiProceedings of FraMCoS-7, May 23-28, 2010
explicitlyoraccounts
the evolution
hydration
hydrates
anhydritefor
(calcium
sulfate) of
were
added.
reaction
andcontent
SF content.
This sorption
The
sulfate
was further
boosted isotherm
by addreads
ing
2% of SO3 to the SO3 in the cement (2.8% in the
case of high-early-strength cement). As fine aggregate, we used three different
types, i.e. quartzite
ag⎡
⎤
1 limestone,
gregate, alkali reactive aggregate,
and
and
⎢
⎥
w (h, α c , α s )size
,α ) 1 −
= G1of(α call
+
⎥ or less.
sthe⎢ types
thee particle
was
2.5
mm
∞
10(g α
− α c )h ⎥ aggre⎢
1 c reactive
The Rc and Sc values of
e alkali
⎣ the
⎦
(4)
gate, as evaluated per JIS A1145 Method of test for
∞ − α )by
⎤
(g α
h chemical
alkali-silica reactivity of⎡⎢ 10
aggregates
c
1 c
(α , α ) e
− 1⎥ respecmethod, are 45K1mmol/L
c s ⎢ and 109 mmol/L,
⎥
tively. Thus Sc is larger ⎣than Rc. Therefore⎦ this alkali reactive aggregate is categorized as reactive and
where
term (gel isotherm) represents the
not
inertthe
JISfirst
A1145.
physically bound (adsorbed) water and the second
term (capillary isotherm) represents the capillary
3.2
Preparation
of mortar
water.
This expression
is valid only for low content
theand
amount
of
of SF. The
G1 represents
Mortar
with coefficient
a water-cement
ratio of 0.5
a sandwater perratio
unitofvolume
held
in theaccording
gel porestoat JIS
100%
cement
1.5 was
mixed
R
relativeand
humidity,
can be expressed
5201,
formed and
intoit 4x4x16cm
prisms (Norling
using a
Mjornell
mould
in 1997)
which as
a plug for measuring the length can
be set according to JIS A1146 (Method of test for
alkali-silica
by mortar-bar
c c + k s ofα aggregates
G1 (α c , α s ) = kreactivity
(5)
vg αcuring
c vg(heat
s s curing) was performed
method).
Steam
as 4 hours of pre-curing at 20°C, steam curing at
90°C
andmaterial
gradual parameters.
cooling for From
8 hours.
ksvg are
the
wherefor
kcvg12andhours
Two
heating
methods
were
used
for
curing.
One
maximum amount of water per unit volume that can
method
was (both
quick capillary
heating by
placing
specimens
difill all pores
pores
and gel
pores), one
rectly
in an oven
90°C
after 4 hours of pre-curing,
obtains
can calculate
K1 asat one
and the other was gradual heating at 20°C/hour up to
90°C after 4 hours of pre-curing.
After
⎡
⎛
⎞ ⎤
∞ measuring
⎜ g α − α ⎟h ⎥
⎢
c
c
their length,
the
mortar
specimens
were
stored
in
⎝
⎠
w −
α s + α s −G ⎢ −e
⎥
c
s
⎥ acwater at 20°C or also stored in⎢ a vessel at 40°C
⎦ (6)
⎣
K (α c α s )to
= JIS A1146. Three prisms of each mortar
cording
⎛
⎞
∞
α − α ⎟measured
were prepared and their⎜⎝ glengths
over 400
c c ⎠h −
e
days. The changes in length were measured by
means of a dial gauge and cslide calipers.
Table 1
and ksfor
The material parameters k method
vg and g1 can
lists the materials and heating vg
curing.
10
0
0.188
0.22
1
1
1
,
1
10
1
1
be calibrated by fitting experimental data relevant to
free 1.(evaporable)
water Method
content
in concrete at
Table
Material and heating
for curing.
various
ages
(Di
Luzio
&
Cusatis
2009b).
Cement
High-early-strength cement (HPC)
Aggregates
Quartzite, ASR reactive aggregates
2.2 Temperature
evolution
Quick heating in 90°C oven
Heating
method
Note that,
at early
age, since
chemical reactions
or gradual
heatingthe
at 20°C/hr.
associated
with
hydration
and SF
SO4), gypsum,
hemihydrates
, reaction
(K2cement
Sulfate type
CaSO4
are exothermic,anhydrite
the temperature
field is not uniform
Addition
for non-adiabatic
systems
even if the environmental
K2SO
2.0%
4
of
sulfate
temperature
is constant. Heat conduction can be
CaSO4
2.0%
to cement
described in concrete,
at least for temperature not
exceeding 100°C (Bažant & Kaplan 1996), by
Fourier’s law, which reads
4 RESULTS AND DISCUSSIONS
q
= − λ ∇T
(7)
4.1 Influence of sulfate and heating condition
where
q is the
heat flux,of Ttheiskind
the ofabsolute
In
this study,
the influence
sulfate
temperature,
and λasisshown
the heat
conductivity;
in this
added
was studied
in Figure
6. The heating
rate of steam curing was 20°C/hr up to 90°C after 4
hours of pre-curing. Other conditions were the same
as shown in Table 1. The expansion due to DEF occurred from around the curing age of 70 days in the
case of the addition of 2% of K2SO4. The other mortars prepared with calcium sulfate had less expansion. The mortar to which 2% of anhydrite CaSO4
and hemihydrate were added exhibited no DEF expansion at all.
2.50
2.00
Expansion(%)
anhydrite CaSO4
1.50
J = − D ( h,(ASR)
T )∇h and DEF
4.2 Alkali-silica reaction
Test specimens were prepared with alkali reactive
Theaggregate
proportionality
coefficient D(h,T)
aggregate or quartzite
using high-earlymoisture
permeability
and
it is a nonlinea
strength cement with a 2% addition of potassium
of
the
relative
humidity
h
and
temperature
sulfate as SO3 and cured under the same steam
Theinmoisture
condition. Figure&8Najjar
shows1972).
changes
length inmass
the balanc
that
the
variation
in
time
of
the
water mas
case where the storage conditions after steam curvolume
of
concrete
(water
content
ing were set for humid air curing at 40°C, identicalw) be eq
divergence
of themethod
moisture
to those of the mortar
bar test
forflux
theJalkali-silica reaction, while Figure 9 indicates
w the case of water curing at
changes in length
− ∂in
= ∇•J
∂t
20°C.
hemihydrates
gypsum
1.00
1.8
K2SO4
0.50
1.6
1.4
0
100
200
300
400
Age(days)
Figure 6. Influence of kind of sulfate added heating to 90°C at
rate of 20°C/hr.
The mortar with the addition of 2% gypsum expanded after 200 days up to 0.3%. Figure 7 shows
the influences of the kind of sulfate in mortar prepared by quick heating directly in an oven at 90°C after
4 hours of pre-curing. Mortar to which K2SO4 was
added and quick-heated in an oven at 90°C showed
the same expansion as the same type of mortar
gradually heated at the rate of 20°C/hr to 90°C. Mortar to which 2% of anhydrite CaSO4 was added also
expanded, to 1.4%. Gypsum mortar showed expansion of 0.3% after 280 days. Hemihydrate mortar did
not expand at all. Expansion by DEF was affected
by the heating rate of the mortar to the maximum
temperature. The heating rate influences the hydration of cement and the formation and deformation of
ettringite at an early age with calcium sulfate added.
Since K2SO4 resolves quickly to water, the influence
of K2SO4 on the heating rate of mortar diminishes.
The addition of anhydrite CaSO4 to a concrete
subjected to high-temperature steam curing increases
the risk of occurrence of expansion due to DEF. It is
necessary in more research to explain due to DEF affected by heating rate in case of calcium sulfate.
2.50
Expansion(%)
2.00
anhydrite CaSO4
1.50
hemihydrates
1.00
gypsum
0.50
K2SO4
0.00
-0.50
0
100
200
300
400
Age(days)
Figure 7. Influence of kind of sulfate added quick heating in
90°C oven.
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0
Figure 8. Influence of steam curing and aggregates (ASR ag∂wat ∂40℃.
gregate, quartz) in wet air
h + ∇ • ( D ∇h) = ∂we α ∂we
− e
c
h
∂α
∂α
∂h ∂t
c
&+
s
3.5
α&s + w
where ∂we/∂h is the slope of the sorption/
isotherm (also called moisture
capac
90℃2.5
governing equation (Equation
ASR3) must be
2.0
by appropriate boundary and90℃initial conditi
limestone
The relation between the90℃amount of e
1.5
quartz
water
and
relative
humidity
is called ‘‘
1.0
20℃isotherm”
if
measured
with
ASR increasing
0.5
humidity and ‘‘desorption isotherm” in th
0.0
case. Neglecting their difference (Xi et al.
0
50
100
150
200
250
300
the
following,
‘‘sorption
isotherm” will be
Age(days)
reference to both sorption and desorption c
Figure 9. Influence of
(ASR ag-of the
Bysteam
the curing
way,andif aggregates
the hysteresis
gregate, quartz) in water
at
20℃.
isotherm would be taken into account, two
relation, evaporable water vs relative humi
With humid airbecuring
at 40°C, expansion
used according
to the signoccurs
of the varia
in the case of relativity
reactive aggregate,
humidity. regardless
The shapeof of the
whether steam curing
wasfor
applied,
the quantity
isotherm
HPC isbut
influenced
by many p
of expansion is smaller
compared
with
the caseextent
of and
especially those that influence
DEF expansion. chemical
No expansion
due
to
DEF
is
obreactions and, in turn, determ
served even if steam
curing
performed
when a (waterstructure
and ispore
size distribution
quartzite aggregate
is used.
From
this fact,composition,
it is unratio,
cement
chemical
SF
derstood that withcuring
humidtime
air and
curing
at
40°C,
the al- mix
method, temperature,
kali-silica reaction
takesInplace
but DEF does
not formulatio
ocetc.).
the literature
various
cur.
found to describe the sorption isotherm
concrete (Xi et al. 1994). However, in th
paper the semi-empirical expression pro
Norling Mjornell (1997) is adopted b
3.0
Expansion(%)
-0.50
Expansion (%)
0.00
The water content w can be expressed a
of the evaporable water we (capillary wa
vapor, and adsorbed water) and the non-e
(chemically bound) water90℃w (Mil
ASR n
Pantazopoulo & Mills 1995).
It
20℃- is reas
assume that the evaporable ASR
water is a fu
90℃relative humidity, h, degree of hydration
degree of silica fume reaction,quartz
αs, i.e. we=w
20℃= age-dependent sorption/desorption
quartz
(Norling Mjonell 1997). Under this assum
by substituting Equation 1 into Equati
50
100
150
obtains
Age (days)
Proceedings of FraMCoS-7, May 23-28, 2010
J
= − D ( h , T ) ∇h
(1)
The proportionality coefficient D(h,T) is called
moisture permeability and it is a nonlinear function
of the relative humidity h and temperature T (Bažant
& Najjar 1972). The moisture mass balance requires
that the variation in time of the water mass per unit
volume of concrete (water content w) be equal to the
divergence of the moisture flux J
− ∂ = ∇•J
∂
(2)
w
t
The water content w can be expressed as the sum
of the evaporable water we (capillary water, water
vapor, and adsorbed water) and the non-evaporable
(chemically bound) water wn (Mills 1966,
Pantazopoulo & Mills 1995). It is reasonable to
assume that the evaporable water is a function of
relative humidity, h, degree of hydration, αc, and
degree of silica fume reaction, αs, i.e. we=we(h,αc,αs)
= age-dependent sorption/desorption isotherm
(Norling Mjonell 1997). Under this assumption and
by substituting Equation 1 into Equation 2 one
obtains
Figure 10. BEI of ASR damaged hardened mortar (ASR gel
∂w
∂w in
formed
∂hthe crack of ASR aggregate).
e α ∂we α w
+ ∇ • ( D ∇h ) =
− e
(3)
c
s
n
h
∂α
∂α
∂h ∂t
c
&+
s
&+ &
where ∂we/∂h is the slope of the sorption/desorption
isotherm (also called moisture capacity). The
governing equation (Equation 3) must be completed
by appropriate boundary and initial conditions.
The relation between the amount of evaporable
water and relative humidity is called ‘‘adsorption
isotherm” if measured with increasing relativity
humidity and ‘‘desorption isotherm” in the opposite
case. Neglecting their difference (Xi et al. 1994), in
the following, ‘‘sorption isotherm” will be used with
reference to both sorption and desorption conditions.
By the way, if the hysteresis of the moisture
isotherm would be taken into account, two different
relation, evaporable water vs relative humidity, must
be used according to the sign of the variation of the
relativity humidity. The shape of the sorption
isotherm for HPC is influenced by many parameters,
especially those that influence extent and rate of the
chemical reactions and, in turn, determine pore
structure and pore size distribution (water-to-cement
ratio, cement chemical composition, SF content,
curing time and method, temperature, mix additives,
etc.). In the literature various formulations can be
found to describe the sorption isotherm of normal
concrete
(Xi ofet DEF
al. 1994).
in the(AFm
present
Figure
11. BEI
damagedHowever,
hardened mortar
parpaper the
expression proposed by
ticipated
in thesemi-empirical
crack around the aggregate).
Norling Mjornell (1997) is adopted because it
Proceedings of FraMCoS-7, May 23-28, 2010
explicitly
accounts
evolution
of hydration
With water
curingfor
at the
20°C,
the expansion
in the
reaction
and steam
SF content.
Thissmall
sorption
isotherm
case
without
curing was
regardless
of
readstype of aggregate, and expansion due to DEF
the
was produced when steam curing was applied. The
quantity of expansion was
as large as 2.0%
with
⎡
⎤
1 aggregate.
both reactive aggregate and
quartzite
⎢
⎥
w (h, α c , αstoring
α c , α s )⎢1 −
+ Un⎥to DEF
s ) = G1 (conditions,
dere such
expansion
due
∞
10(g α
− α )h ⎥
⎢
can be confirmed, but no
e 1 c by calkali-silica
⎣ expansion
⎦
(4)
reaction is observed. Although the alkali-silica reac⎤
)h
α ∞the− αtemperature
10(g at
tion appears at an early⎡⎢ age
of
c
1 c
K1 (αslows
e
− 1⎥a longer
40°C, the reaction
at 20°C and
c , α s )⎢down
⎥
⎣
curing age is required to cause
expansion. ⎦
Figures 10 and 11 show the BEI of the alkaliwherereaction
the firstdamaged
term (gel
isotherm)
the
silica
mortar
and therepresents
BEI of DEF
physicallymortar,
bound respectively.
(adsorbed) water
and the
second
damaged
Through
naked
eye
term (capillary
the hardened
capillary
observation,
bothisotherm)
ASR and represents
DEF damaged
water.
This
expression
is
valid
only
for
low
content
specimens could be seen to have similar textures
the amount
of
of SF.cracks
The covering
coefficienttheGwhole
with
specimen,
producing
1 represents
per unit
volume held
in theitgel
porestoatdistin100%
awater
map-like
appearance.
However
is easy
relativeDEF
humidity,
and mortar
it can befrom
expressed
(Norling
guish
damaged
ASR damaged
Mjornell
1997)
as
mortar by the observation of BEI images of polished
sections. The phase in cracks around the aggregate is
determined
by EMPA. The AFt
G1 (α c , α s ) = as
k c ettringite
α c + k s α(AFt)
(5)
c vg by
s s XRD. In ASR damaged
phase
is also vg
determined
mortar, aggregate reacted at the surface and the center
of the
alkali-silica
gel formed
in
and ksvg areand
material
parameters.
From the
where
kcvg aggregate
the
cracks
and
in
the
space
of
air
voids,
and
cracks
maximum amount of water per unit volume that can
formed
across
the capillary
reacted aggregate.
The
distribufill all pores
(both
pores and gel
pores),
one
tion
of cracksKwas
limited
as revealed by BEI imobtains
can calculate
1 as one
ages. In DEF damaged mortar, cracks were distributed uniformly as seen in BEI
images,
and
⎡
⎛
⎞ ⎤ the
∞
⎜ g α − α ⎟h ⎥
⎢
c
c
thickness of
the
cracks
caused
by
DEF
is
the
⎝
⎠ same
w −
α s + α s −G ⎢ −e
⎥
c
s
⎢
thickness at 10 microns. Ettringite
formed in⎥ (6)
the
⎦
⎣
)=
K (α c α sCrack
cracks.
appeared ⎛in the paste
parts
and
around
⎞
α ∞ − α across
⎜ g form
⎟h
the aggregate and did not
the aggregate.
c
c
⎝
⎠ −
e
These observations coincide with the findings of a
previous paper (Thomas 2008).
ks and g1 can
The material parameters kcvg and
The alkali-silica reaction and
DEFvgare phenombe calibrated by fitting experimental data relevant to
ena of expansion due to the swelling and expansion
free (evaporable) water content in concrete at
of alkali silica gel and secondarily generated ettringvarious ages (Di Luzio & Cusatis 2009b).
ite, produced in the presence of alkali or sulfate as
the causal substance. The alkali-silica reaction and
DEF
have so far been
discussed as phenomena simi2.2 Temperature
evolution
lar to each other. Table 2 summarizes the relating
Note that, at early age, since the chemical reactions
factors of alkali-silica reaction and DEF. Although
associated with cement hydration and SF reaction
there is no factor common to both, they are produced
are exothermic, the temperature field is not uniform
equally when an alkali is contained in the cement in
for non-adiabatic systems even if the environmental
the form of alkaline sulfate. As conditions for the
temperature is constant. Heat conduction can be
simultaneous occurrence of ASR and DEF, there is
described in concrete, at least for temperature not
some deviation in the conditions common to both
exceeding 100°C (Bažant & Kaplan 1996), by
phenomena. In DEF, ettringite decomposes at the
Fourier’s law, which reads
time of steam curing and the sulfate ion content in
the
pore water of hardened specimens remain high
q = − λ ∇T
even thereafter, causing secondary production of (7)
ettringite. During the time when the sulfate ion content
where
q is
the is heat
T is theion absolute
in
the pore
water
high, flux,
the hydroxide
content
temperature,
and
λ
is
the
heat
conductivity;
this
in the pore water is kept low, and the pH of theinpore
water does not go up while DEF is taking place. For
10
0
0.188
0.22
1
1
,
1
10
1
1
1
that reason, if there is any case where the alkalisilica reaction and DEF are produced simultaneously, it should be considered that an alkali-silica reaction took place after occurrence of DEF.
Table 2. Impact factors and characteristics for alkali-silica reaction (ASR) and DEF.
ASR
DEF
ASR aggregate
○
×
Quantity of alkali
○
△
Sulfate
×
○
Steam curing
×
○
Reaction temperature
40°C
20°C
Water supply
△
○
Expansion
0.1%
1.0%
○：More effective, △: Effective,
×：Less effective
5 CONCLUSIONS
The results obtained by this research are as follows.
- DEF expansion was affected by the heating rate of
mortar to maximum temperature. The heating rate
influences the hydration of cement and the formation
and deformation of ettringite at an early age when
calcium sulfate is added. The addition of anhydrite
CaSO4 to a concrete subjected to quick heating
through high-temperature steam curing increases the
risk of occurrence of expansion due to DEF.
- Expansion due to DEF is affected by the type of
aggregate. Both DEF and ASR damaged hardened
specimens had the same texture with cracks covered
the whole specimen, producing a map-like appearance as seen with the naked eye. However, it is easy
to distinguish DEF damaged mortar from ASR damaged mortar by the observation by BEI images of
polished sections. In ASR damaged mortar, cracks
formed across the reacted aggregate. In DEF damaged mortar, cracks were distributed uniformly as
seen in BEI images, and the thickness of the cracks
caused by DEF is the same thickness at 10 microns.
D (hcracks.
, T ) ∇h
Ettringite formedJin= −the
Crack appeared in
the paste parts and around the aggregate and did not
form across the aggregate.
The proportionality coefficient D(h,T)
moisture permeability and it is a nonlinea
- With regard to the
conditions
for the simultaneous
of the
relative humidity
h and temperature
occurrence of the&alkali
aggregate
reaction
and DEF,
Najjar 1972). The moisture
mass balanc
there is some deviation
in
common
points.
Considerthat the variation in time of the water mas
ing the behavior volume
of sulfate
ions and (water
hydroxide
ionsw) be eq
of concrete
content
in pore water, it divergence
seems reasonable
to consider
of the moisture
flux Jthat
the alkali aggregate reaction takes place after the occurrence of DEF. ∂w
−
∂
t
= ∇•J
REFERENCES
The water content w can be expressed a
wa
of
the evaporable water we (capillary
Hanehara S. 2003, Sulfate deterioration and DEF. Cement
and
vapor,JCA,
andNo.671,
adsorbed
water) and the non-e
Concrete (in Japanese),
pp.61-62,
(chemically
wn (Mil
Hanehara S., Fukuda
S., Oyamada bound)
T., Fujiwarawater
T., 2006a,
Pantazopoulo
& ofMills
1995).
It is reas
Study on the occurrence
condi-tion
expan-tion
of concrete by delayedassume
ettringitethat
formation,
Proceedings water
of the is a fu
the evaporable
Japan Concrete Insti-tute,
Vol.28,
No.1,
pp.748-748
relative humidity, h, degree of hydration
Hanehara S., Fukuda S., Oyamada T., Fujiwara T., 2006b, Indegree
of silica fume reaction, αs, i.e. we=w
fluence of material, condition of steam-curing and stored
=
age-dependent
sorption/desorption
condition on expansion of concrete by delayed
ettringite
(Norling
Mjonell
1997).
Under this assum
formation, Cement Science and Concrete Technology,
No.60, pp.335-341
by substituting Equation 1 into Equati
Hanehara S., Fukudaobtains
S., 2007, Delayed Ettringite Formation in
Concrete Precast, Concrete technology, Vol.26, No.3, pp.9-16
∂w 2008a, Dew Fukuda
Hanehara S., Oyamada ∂T.,
S., Fujiwara T.,
∂h
e α ∂we α w
( D ∇ag-gregate
) =
+ ∇ •alkali
− e
h
layed ettringite formation
and
reaction,
c
s
h
∂α
∂α
∂h ∂t
The 8th International Conference on Creep, Shrinkage
ands
c
durability of concrete and concrete structure, pp.1051-1056,
(Iseshima-Japan) where
ISBN:978-0-415-48508-1
is the
slope of the sorption/
∂we/∂h T.,
Hanehara S., Oyamada T., Fujiwara
2008b, Re-production
isotherm
(also
called
of delayed ettringite formation in con-crete andmoisture
its mecha- capac
governing
equation
(Equation
3) must be
nism, the 1st International confe-rence on microstructure
and(Nanjing,
initial conditi
related durabilitybyofappropriate
cementitiousboundary
composites,
China) ISBN:978-2-35158-065-3
The relation between the amount of e
Shimada Y., 2005, Dissertation
chemicalhumidity
path of ettringite
water
andofrelative
is called ‘‘
formation in heat-cured mortar and its rela-tionship to exisotherm”
if
measured
with
increasing
pansion, Northwestern University
humidity
and
‘‘desorption
isotherm”
Taylor H. F. W., Famy C., Scrivener K. L., 2001, Delayed Et- in th
case.
Neglecting
theirResearch,
difference
(Xi et al.
tringite Formation,
Cement
and Concrete
Vol.31,
pp.683-693
the following, ‘‘sorption isotherm” will be
Thomas M., Folliardreference
K., Drimalas
Ramlochan
2008,
DitoT.,both
sorptionT.,and
desorption
c
agnosing delayedBy
ettringite
formation
in
concrete
structures,
the
way,
if
the
hysteresis
of
the
Cement and Concrete Research, Volume 38, 2008, Pages
isotherm would be taken into account, two
841-847
&+
&+
relation, evaporable water vs relative humi
be used according to the sign of the varia
relativity humidity. The shape of the
isotherm for HPC is influenced by many p
especially those that influence extent and
chemical reactions and, in turn, determ
structure and pore size distribution (waterratio, cement chemical composition, SF
curing time and method, temperature, mix
etc.). In the literature various formulatio
found to describe the sorption isotherm
concrete (Xi et al. 1994). However, in th
paper the semi-empirical expression pro
Norling Mjornell (1997) is adopted b
Proceedings of FraMCoS-7, May 23-28, 2010