Construction and Building Materials 48 (2013) 764–771

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier.com/locate/conbuildmat

New perspective of service life prediction of fly ash concrete

Zhuqing Yu a,⇑, Guang Ye a,b
a
Microlab, Faculty of Civil Engineering and Geosciences, Delft University of Technology, Delft 2600 GA, The Netherlands
b
Magnel Lab for Concrete Research, Department of Structural Engineering, Ghent University, Ghent, Belgium

h i g h l i g h t s

 The chloride migration coefficients of fly ash concrete are determined with maximum curing age of 2 years.
 w/b Ratio largely influences the chloride migration coefficients of fly ash concrete.
 A reasonable reference time, t0 is determined for fly ash concrete.
 The service life of fly ash concrete is predicted based on DuraCrete model.
 The long-term chloride migration coefficient of fly ash concrete is predicted under laboratory condition.

a r t i c l e i n f o a b s t r a c t

Article history: Rapid Chloride Migration (RCM) test developed in Sweden in the 90s is widely used to determine the
Received 22 April 2013 chloride migration coefficient (DRCM) of concrete because of its simplicity and rapidity. In Europe, the
Received in revised form 16 July 2013 time-dependent DRCM has been introduced in the DuraCrete project as input parameter to predict the ser-
Accepted 21 July 2013
vice life of concrete. In the DuraCrete project, when determining the DRCM of fly ash concrete, reference
Available online 24 August 2013
time (the maturity age of concrete) is normally set as 28 days like what used in Portland cement concrete
and blast furnace slag (BFS) concrete. However, fly ash does not fully play its role in concrete at 28 days
Keywords:
due to slow Pozzolanic reaction of fly ash. Therefore, 28 days is not a realistic reference time for predict-
Fly ash concrete
Chloride migration coefficient
ing service life of fly ash concrete. This study is aimed at determining a reasonable reference time (t0) for
RCM test fly ash concrete. The DRCM of fly ash concrete is determined by the RCM test for a long-term period of cur-
The DuraCrete project ing up to 2 years. The influences of fly ash dosage and w/b ratio on the DRCM of fly ash concrete were con-
Service life prediction sidered. The results show that the DRCM of fly ash concrete significantly decreases after 28 days of curing.
The reference time, t0 for fly ash concrete is proposed according to the concrete maturity indicated by the
content of calcium hydroxide (CH) in cement paste blended with fly ash. This proposed reference time is
applied to predict the service life of fly ash concrete and predict the long-term chloride migration coef-
ficient of fly ash concrete under laboratory condition.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction of Portland cement is able to improve the resistance of concrete

Chloride ingression is one of the most important factors affect-
ing the durability of reinforced concrete structure. Corrosion
caused by chloride penetration leads to the deterioration of bridge,
marine structure, manufacturing plant, and construction building.
In Europe, around 5 billion Euros were spent on maintenance of
the infrastructure due to reinforcement corrosion annually [1].
Nowadays, much attention has been devoted to extend the service
life of concrete structure suffering from chloride ingress. To pre-
vent such kind of corrosion, the concrete structure with good
impermeability is desired. Until now, a number of studies [2–7]
have been performed and shown that fly ash as partial replacement
⇑ Corresponding author. Tel.: +31 (0)15 27 89342; fax: +31 (0)15 27 86383.
E-mail address: yu.zhuqing@hotmail.com (Z. Yu).
against chloride ingress. At present, different test methods for
measuring and evaluating chloride ingress have been proposed.
These methods can be categorized as diffusion test, migration test,
and indirect tests based on resistivity or conductivity [8]. Rapid
Chloride Migration (RCM) test is a non-steady state migration test
applying an external electrical field for accelerating chloride pene-
tration. The parameter, non-steady state migration coefficient
(DRCM) obtained from RCM test describes the property of chloride
transport under a condition of reduced chloride binding [9]. Now-
adays, it is widely used in Europe.
The DRCM obtained from the RCM test is not only used to esti-
mate the resistance of concrete to chloride ingress, but used for
predicting the service life of concrete by means of the model pro-
posed in the DuraCrete-project [10]. The DuraCrete project is one
of the guidelines for durability design and assessment of concrete

0950-0618/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.conbuildmat.2013.07.035
 Z. Yu, G. Ye / Construction and Building Materials 48 (2013) 764–771
structure, which is also widely used in Europe. The chloride con-
tent at certain depth and time is calculated by the following equa-
tion [11,12].
" #
x tures. Graded Dutch river sands with a maximum grain size of 4 mm and gravel
Cðx; tÞ ¼ C s  ðC s  C i Þerf pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
f4kDðtÞtg
where C(x,t) is the chloride content at depth x and time t, Cs is the
surface chloride content, Ci is the initial chloride content in the con-
crete, k is a correction factor, which depends on the binder type, the
environment and the curing conditions. D(t) is the apparent migra-
tion coefficient at time t. It is a time-dependent function.
 n
t0
DðtÞ ¼ D0
t The w/b ratio of 0.4 was used to prepare the cement paste blended with 50% fly
where D0 is the achieved chloride migration coefficient at the matu-
rity age t0 of the concrete [13].
The parameter n is the ageing coefficient. It is theoretically
range from 0 to 1 [13]. It depends on the type of binder, the rate
of cement hydration and environmental factors. D0 and n are the
parameters determined by regression analysis of the test results
[13]. The ageing coefficient, n, was determined by various
researchers for different concrete binders under different exposure
conditions. According to Fick’s second law of diffusion, some diffu-
sion coefficients are determined for the same concrete binder at
different ages. The ageing coefficient can be obtained from the
slope of the diffusion coefficient against time [14]. This way is
adopted in the DuraCrete and European standard [15]. Another
possibility to obtain ageing coefficient is to use the actual RCM val-
ues from specimens of different age. But this RCM test results
should from field exposure describing the actual use conditions
[14]. In DuraCrete, the DRCM,28days measured at the curing age of
28 days (reference time, t0) has been suggested to use for predict-
ing the service life of concrete structures. At 28 days, a substantial
percentage of the hydration has taken place for Portland cement
concrete with a typical water/cement (w/c) ratio and over 90% of
final strength is reached at 28 days. When Portland cement is
partially replaced by fly ash, the addition of fly ash improves the
long-term resistance of concrete to chloride ingress. Cox [4]
demonstrated that the DRCM value of fly ash concrete decreases sig-
nificantly with the increase of curing age, although fly ash concrete
has a much higher DRCM value at 28 days than Portland cement
concrete. Most of the studies about RCM test for Portland cement
concrete and fly ash concrete focused on the short-term curing
age (3 months). The Pozzolanic reaction of fly ash in concrete con-
suming calcium hydroxide (CH) is very slow during the early age
hydration [16,17]. Fly ash concrete needs longer curing time to
reach equal maturity like Portland cement concrete at the same
temperature. The predicted service life of fly ash concrete could
be underestimated when 28 days is chosen as reference time (t0).
The aim of this study is to determine a reasonable reference
time, t0 which can be used to predict the service life of fly ash con-
crete based on the DuraCrete model. In this paper, the chloride
migration coefficient of fly ash concrete blended with different
dosages of fly ash and different w/b ratios were determined by
the RCM test and compared with Portland cement concrete. The
curing periods are up to 2 years. Meanwhile, the content of CH
was determined in blended cement paste. The Pozzolanic reaction
of fly ash is characterized and the reference time is redefined. The
long-term chloride migration coefficient of fly ash concrete is pre-
dicted under laboratory condition which is as an application on the
basis of the redefined reference time.
765
2. Materials and methods
2.1. Materials and properties
Portland cement (CEM I 42,5 N), fly ash (low calcium) with mean particle
size of 21.46 lm, aggregate, and tap water were used in preparing concrete mix-
ð1Þ
with a 16 mm maximum grain size were used as fine and coarse aggregates,
respectively. Demineralized water was applied to cast blended cement paste.
The chemical compositions of fly ash and Portland cement used are shown in
Table 1.
2.2. Mix proportions and specimen preparation
The dosages of fly ash were 30% and 50% as the replacement of Portland cement
by weight for fly ash concrete mixtures and blended cement paste mixtures. The
water/binder (w/b) ratios used for fly ash concrete were 0.4, 0.5, and 0.6. Two w/
ð2Þ b ratios, 0.4 and 0.5 were used to cast the cement paste blended with 30% fly ash.
ash. The fly ash concrete with 50% of fly ash and with a w/b ratio of 0.6, the cement
paste blended with 30% fly ash and a w/c ratio of 0.6, and the cement paste blended
with 50% fly ash and two w/c ratios of 0.5 and 0.6 were excluded because of segre-
gation and bleeding. The mixture proportions of concrete and blended cement paste
are given in Tables 2 and 3.
After mixing for 2 min, the concrete mixtures used in the RCM test were cast in
cylindrical molds with a diameter of 100 mm and height of 300 mm. After 24 h, the
concrete specimens were demolded and cured in the fog room (95% ± 5% relative
humidity and 20 °C ± 1 °C temperature). After preconditioning, the RCM test was
performed at 28, 91, 180, 365 and 730 days. Before testing, the hardened concrete
specimens were cut into three cylindrical samples (height = 50 mm) as shown in
Fig. 1.
The blended cement pastes to be used for thermal gravimetric analysis (TGA)
measurement was cast at room temperature around 20 ± 2 °C. After mixing for
3 min, the specimens were sealed in a plastic bottle to prevent moisture loss and
rotated for 24 h to avoid possible segregation. Then, the specimens were stored at
room temperature. At the cuing age of 28, 90, 180, 365, and 730 days, the specimens
were demolded from the plastic bottle and crushed into small pieces around 1 cm
diameter. Afterwards, the specimens were frozen by immersion into liquid nitrogen
for 5 min in order to stop hydration [18], and then dried in a freeze-dryer until a
constant weight loss (0.05%) was reached. The dried specimens were ground for
TGA test.
2.3. Test methods
2.3.1. Rapid Chloride Migration (RCM) test
The Rapid Chloride Migration (RCM) test was conducted to determine the chlo-
ride migration coefficient of the concrete according to the NT Build 492 standard
[19]. Fig. 2 shows the RCM test set-up used in this study.
Three slices with the thickness of 50 mm were cut from each concrete speci-
men. The surface-dry specimens were placed vertically in a desiccator connected
to a vacuum-pump for 3 h. While the vacuum-pump was still running, the saturated
limewater was slowly filled into the desiccator to immerse all the specimens com-
pletely. After that, the vacuum was maintained for an additional hour before allow-
ing air to re-enter the desiccator. The specimens were kept in the limewater for
18 ± 2 h.
After above preconditioning, the specimens were put in a rubber sleeve and fas-
tened with two stainless steel clamps. The used anolyte solution was 0.3 M NaOH
while the used catholyte solution was 10% NaCl. With the preset voltage 30 V,
the initial current through each specimen was recorded. Following the standard
[19], a corrected voltage and appropriate test duration were chosen. After chloride
migration test, all specimens were split and sprayed with 0.1 M silver nitrate
(AgNO3) solution to determine the penetration depth of chloride. Fig. 3 shows a typ-
ical split specimen after RCM test. The white silver chloride precipitation on the
split surface is clearly visible (after 15 min). The average penetration depth, Xd of
chloride was recorded.
The non-steady-state migration coefficient was calculated by the following
equation [18]:
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi!
0:0239  ð273 þ TÞ  L ð273 þ TÞ  L  xd
DRCM ¼
ðU  2Þ  td
 xd  0:0238
U2
ð3Þ
where DRCM is the non-steady-state migration coefficient, 1012 m2/s; U is absolute
value of the applied voltage, V; T is average value of the initial and final temperatures
in the anolyte solution, °C; L is thickness of the specimen, mm; Xd is average value of
the penetration depths, mm; and td is the test duration, h.
766 Z. Yu, G. Ye / Construction and Building Materials 48 (2013) 764–771

Table 1
Chemical composition of fly ash and cement used in this study.

Chemical composition (% by mass) SiO2 Al2O3 CaO Free-CaO Fe2O3 P2O5 K2O MgO SO3 Na2O
CEM I 42,5 N 20.36 4.96 64.4 0.6 3.17 0.18 0.64 2.09 2.57 0.14
Fly ash 48.36 31.36 7.14 – 4.44 1.90 1.64 1.35 1.18 0.718

Table 2
Mixture proportions of concrete (kg/m3).

Number CEMI 42.5 N Fly ash Tap water Masonry sand 0/2 Sand 0/4 Gravel 4/16
R04 0.390 – 0.156 0.178 0.498 1.103
FA3004 0.273 0.117 0.156 0.178 0.498 1.103
FA3005 0.273 0.117 0.195 0.178 0.498 1.103
FA3006 0.273 0.117 0.234 0.178 0.498 1.103
FA5004 0.195 0.195 0.156 0.178 0.498 1.103
FA5005 0.195 0.195 0.195 0.178 0.498 1.103

Table 3
Mixture proportions of blended cement paste (g).

Number CEMI 42.5 N Fly ash Distilled water

OPC04 500 – 200
OFA3004 350 150 200
OFA3005 350 150 250
OFA5004 250 20 200

Φ100

50mm
#1
Fig. 3. Typical concrete sample after RCM test.
#2

#3

3. Results and discussions

Fig. 1. The sawing plan of test specimens.

3.1. The influence of fly ash dosage on the DRCM of concrete

The development of the DRCM of concrete with different dosages

Sodium chloride
Sodium hydroxide
Fig. 2. The chloride migration set-up in laboratory.
2.3.2. Thermal gravimetric analysis (TGA) test
Thermal gravimetric analysis (TGA) test was carried out in order to determine
the content of CH in cement paste blended with or without fly ash. About 50 mg
of powder sample was taken in an aluminum oxide (Al2O3) crucible and heated
at a temperature range between 40 °C and 1100 °C at the heating rate of
10 °C min1 using argon as a protect gas.
of fly ash varying between 0% and 50% over 2 years is shown in
Fig. 4. The DRCM values of Portland cement concrete keep almost
plateau with increasing curing age from 28 days to 730 days in
the double logarithmic coordinate system. It indicates that curing
age has no obvious influence on the resistance of Portland cement
concrete to chloride ingress. This may be because Portland cement
concrete with a w/c ratio of 0.4 is nearly fully hydrated and
matured after 28 days of moist curing [20].
From Fig. 4, the DRCM values of fly ash concrete are much lower
than that of Portland cement concrete after 28 days, indicating a
higher resistance of fly ash concrete to chloride ingress. It is
reported that the incorporation of fly ash results in a decrease in
the capillary pore size [21] and an improvement in the interfacial
zone of concrete [22,23]. Therefore, a more durable and impenetra-
ble concrete is obtained.
The results presented in Fig. 4 are strongly coincided with the
mathematical model used to describe the time-dependent chloride
migration coefficient as mentioned in Eq. (2). After taking loga-
rithm on both sides of Eqs. (2) and (5) can be derived.
lgðDðtÞÞ ¼ lgðD0 Þ þ n lgðt 0 Þ  n lgðtÞ ð4Þ
) lgðDðtÞÞ ¼ n lgðtÞ þ C; C ¼ lgðD0 Þ þ n lgðt 0 Þ ð5Þ
 Z. Yu, G. Ye / Construction and Building Materials 48 (2013) 764–771 767

100 100

Chloride Migration Coefficients

Chloride Migration Coefficients

(×10 -12 m2/s)

10
(×10 -12 m2 /s)

10

1
1
0% (fly ash)_0.4; n=0.0147 50% (fly ash)_0.4; n=0.796
30% (fly ash)_0.4; n=0.895 50% (fly ash)_0.5; n=0.808
50% (fly ash)_0.4; n=0.796 0,1
10 100 1000
0.1 Curing age (days)
10 100 1000
Curing age (days) Fig. 6. The influence of w/b ratio and curing age on the DRCM of fly ash concrete
(FA50).
Fig. 4. The influence of fly ash dosages and curing age on the DRCM of concrete (w/b
ratio = 0.4).

w/b ratio mainly influences the initial chloride diffusivity, while

The negative slope of this curve, namely the ageing coefficient, n
is calculated and shown in Fig. 4. It seems that the addition of fly
ash in concrete mainly influences the slope of the curve, namely
the development of chloride migration. Similar results about the
influence of other Pozzolanic material (condensed silica fume) on
the DRCM of concrete are found by Maage [13].
3.2. The influence of w/b ratio on the DRCM of fly ash concrete
The development of the DRCM with time for fly ash concrete
with different w/b ratios under curing age up to 2 years is illus-
trated in Figs. 5 and 6. The DRCM of fly ash concrete decreases dra-
matically with the increase of curing age. The resistance of fly ash
concrete to chloride ingress is improved by lowering w/b ratio.
Higher w/b ratio causes an increase in the porosity of concrete
and results in lower resistance to chloride ion ingress [24,25].
Compared with the influence of fly ash dosage on the DRCM of con-
crete (Fig. 4), lowering w/b ratio is a more effective to reduce the
chloride penetration ratio of fly ash concrete.
Straight lines are found in the double logarithmic coordinate
system for fly ash concrete. The age coefficient, n is also calcu-
lated and shown in Figs. 5 and 6. There is no significant change
of the value of n for fly ash concrete with two different dosages
of fly ash and three different w/b ratios under the same curing
conditions. The value of n is around 0.8. It indicates that the
the ageing coefficient has no obvious change for all fly ash con-
crete mixtures.
3.3. The relationship between the DRCM and w/b ratio
The relations between the DRCM and w/b ratio for fly ash con-
crete at each curing age from 28 days to 730 days are plotted in
Fig. 7. The DRCM of fly ash concrete increases with the increase of
w/b ratio. The w/b ratio is a very important parameter in concrete
since it directly affects the rate of hydration, the pore structure
development and so on [26].
There is a linear relationship between the DRCM and w/b ratios
for fly ash concrete (FA30) at each curing age from 28 days to
730 days. It should not be denied that the linear relationship about
the DRCM and w/b ratio also exists in higher volume fly ash con-
crete, i.e., FA50, although only two different w/b ratios are pre-
sented for FA50 fly ash concrete.
The liner relationships between the DRCM and w/b ratio for fly
ash concrete at each curing age are strongly in accordance with
Eq. (6) proposed by [12].
DRCM ð28 daysÞ ¼ Aðw=bÞ þ B ð6Þ
where A and B are constants related to cement types and amount of
fly ash used.

100
Chloride Migration Coefficients

Chloride Migration Coefficients

35

30 FA30 28 days
FA50
(×10 -12 m2 /s)

10 25
(×10 -12 m2 /s)

20

15
1 91 days
10
30% (fly ash)_0.4; n=0.895
180 days
30% (fly ash)_0.5; n=0.862 5
30% (fly ash)_0.6; n=0.766 365, 730 days
0.1 0
10 100 1000 0.3 0.4 0.5 0.6 0.7
Curing age (days) water to binder ratio
Fig. 5. The influence of w/b ratio and curing age on the DRCM of fly ash concrete Fig. 7. The relations between the DRCM and w/b ratio for fly ash concrete at each
(FA30). curing age up to 730 days.
768 Z. Yu, G. Ye / Construction and Building Materials 48 (2013) 764–771

3.4. Discussion of the DRCM of concrete 18 14

Chloride Migration Coefficients

16 12

Ca(OH)2 (g/g ceemnt ) (%)

As discussed above, the incorporation of fly ash reduces the rate
14 Drcm_50% fly ash; w/b=0.4
of chloride ingress in concrete which was confirmed by other stud- 10
ies [26–28]. The Pozzolanic reaction of fly ash in concrete is one of 12

(×10 -12 m2/s)

CH_50% fly ash; w/b=0.4"
the main reasons which improve the resistance to chloride ingress 10 8
[29]. The Pozzolanic reaction of fly ash takes place consuming CH
8 6
and forming additional (C–S–H) phases which results in a more
dense concrete structure. CH phase is the weak phase in the hard- 6
4
ened concrete structure [30]. The content of CH in concrete struc- 4
ture influences the resistance to chloride ingress. 2
2
The content of CH measured by TGA in blended cement paste is
0 0
illustrated in Fig. 8. For Portland cement paste, the content of CH 0 200 400 600 800
has no significant change after 28 days. This trend is remarkable Curing age (days)
consistent with the development of the DRCM of Portland cement
concrete as shown in Fig. 4. Under the same curing condition, the Fig. 9. Comparision of the DRCM and the amount of CH.
resistance of Portland cement concrete to chloride ingress is di- 2 0 13
rectly influenced by the rate of cement in concrete. The content d
6 B x C7
g ¼ cdcr  cd ðx; tÞ ¼ cdcr  cds;cl 41  erf @ qffiffiffiffiffiffiffiffiA5 ð7Þ
of CH in blended cement paste decreases dramatically after 2 RdtðtÞ
28 days due to the Pozzolanic reaction of fly ash. Meanwhile, more cl

hydration products like C–S–H are produced and more dense

d
microstructure is formed which benefit to resistance to chloride in- where cdcr is design value of the critical chloride concentration; cs ; cl
gress. Around 180 days, the content of CH does not change too is design value of the chloride surface concentration; xd is design va-
much which indicates a relative mature structure is formed. Like lue of the cover thickness; Rdcl ðtÞ is design value of the chloride resis-
Portland cement concrete, the variation of the DRCM of fly ash con- tance; and t is the time.
crete is surprising similar with the change of the content of CH in The design value of the critical chloride concentration, surface
blended cement paste. Fig. 9 shows an example of the comparison chloride concentration, cover thickness, and the time dependent
of the DRCM of concrete with 50% of fly ash and the content of CH in resistance can be determined from following equation:
cement paste blended with 50% of fly ash (w/b = 0.4). After
180 days, not only the content of CH in blended cement paste, 1
cdcr ¼ cccr  ; cds;cl ¼ Acs;cl  ðw=bÞ  ccs;cl ; xd ¼ xc  Dx; Rdcl ðtÞ
but the chloride migration coefficient of fly ash concrete changes cccr
little. Therefore, it can be concluded that fly ash concrete reaches
Rccl;0
a relative maturity is at around 180 days. ¼ ð8Þ
c c nc
ke;cl  kc;cl  ðtt0 Þ cl
 cRcl
3.5. Service life prediction of fly ash concrete

As mentioned in the introduction, 28 days is normally used as
reference time t0 to predict the service life of Portland cement con-
crete in DuraCrete project because Portland cement concrete is
matured at this age. However, for fly ash concrete, 28 days is not
a reasonable reference time since it takes about 180 days to reach
optimum maturity as discussed in Section 3.4. Therefore, 180 days
is defined as the new reference time, t0 to predict the service life of
fly ash concrete.
3.5.1. DuraCrete model
According to the DuraCrete [10], corrosion of reinforcement is
initiated when the chloride concentration around the reinforce-
ment exceeds the critical chloride concentration:
16
Ca(OH)2 (g/gceemnt) (%)
where cccr is the characteristic value of the critical chloride concen-
tration. It is in the form of total chloride by weight of binder [31]. In
order to investigate the critical chloride concentration, several
experimental setups are included such as: a steel electrode of inter-
est embedded in a cement based material; chloride ions present at
the steel surface; detection of the depassivation of the steel elec-
trode; quantification of the total chloride content at the steel sur-
face using X-ray fluorescence spectrometry [32]. The critical
chloride concentration mainly depends on the quantity of the
steel–concrete interface, the pH of the pore solution, the electro-
chemical potential of the steel, the type of cement, w/c ratio and
so on [32]. cccr is the partial factor of the critical chloride concentra-
tion; Acs;cl is a regression parameter describing the relation between
the chloride surface concentration and the water-binder ratio; w/b
is water to binder ratio; ccs;cl is the partial factor for the surface con-
centration; xc is the cover thickness; Dx is the margin for the cover

14
thickness; Rccl;0 is resistance with respect to chloride ingress deter-
12 mined on the basis of compliance tests. For example,
10 Rccl;0 ðyear=mm2 Þ ¼ D 1 c
; ke;cl is environment factor.
2 13
RCM;28days ðm =sÞ3:153610
8 c
It can be split up into two factors: ke;0 describing the environment
6 c c c c c
and ke;c describing the type of cement. ke;cl ¼ ke;0  ke;c ; kc;cl is curing
4 Portland cement paste; w/c=0.4 factor; t0 is the age of concrete when the compliance test is per-
Cement paste blended with 30% fly ash; w/b=0.4 formed. For example, t0 = 0.767 years (corresponding to 28 days);
2 Cement paste blended with 50% fly ash; w/b=0.4
ncl is age factor which is same as ageing coefficient (n) mentioned
Cement paste blended with 30% fly ash; w/b=0.5
0 above; and cRcl is the partial factor for the resistance with respect
0 200 400 600 800
to chloride ingress.
Curing age (days)
By rewriting Eqs. (7) and (8), the time to initiation of corrosion,
Fig. 8. The content of CH in cement paste blended with or without fly ash. namely service life, is determined from the following equation:
 Z. Yu, G. Ye / Construction and Building Materials 48 (2013) 764–771 769

Table 4 Table 6
The partial factors for structures in a marine environment. Predicted service life of fly ash concrete with different reference time, t0.

Cost of mitigation of risk High Normal Low t0 (days) t0 (years) DRCM,28days/1012 m2/s Rccl;0 /year/mm2 tdi /years
relative to the cost of repair
28 0.0767 15.18 0.0025 33
Dx 20 14 8 180 0.4932 2.12 0.0149 620
cccr 1.20 1.06 1.03
ccs;cl 1.70 1.40 1.20
cRcl 3.25 2.35 1.50
Table 7
The chloride migration coefficients of fly ash concrete.

2 !!2 3 1 DRCM (1012 m2/s) 28 days 91 days 180 days 365 days 730 days
1nc
2 1 cc 1 Rccl;0 cl

t di ¼4 c  erf 1  cr   nccl
x  Dx cccr Acs;cl  ðw=bÞ  ccs;cl
ke;cl  kc;cl  ðt0 Þ  cRcl
3.5.2. Service life prediction of fly ash concrete
As an example, the fly ash concrete blended with 50% fly ash
with a w/b ratio of 0.4 is selected to predict the service life of fly
ash concrete under the condition of marine atmospheric. The input
parameters for the DuraCrete model are also discussed in the fol-
lowing part.
The presence of fly ash increases the chloride binding capacity
of fly ash concrete and reduces the level of chlorides required to
initiate corrosion [33,34]. The critical chloride concentration is
chosen as 0.2 for fly ash concrete with 50% of fly ash according
to the work of Thomas [35]. The concrete cover is an important fac-
tor to protect the steel reinforcement bars from corrosion. Euro-
code 2 recommends the minimum cover requirements for
normal weight concrete under chloride-induced corrosion is
45 mm [36]. According to the work of van Breugel [12], the cover
depths of 45 mm and 35 mm are chosen to predict the service life
of fly ash concrete using two different reference times (t0), 28 days
and 180 days. That is because the DRCM,28days is 15.18  1012 m2/s
and DRCM,180days is 2.12  1012 m2/s as shown in Fig. 4. According
to NEN-EN 206 [15], the ageing coefficient, n for fly ash concrete
(CEM I with 21–30% fly ash) under the condition of marine atmo-
spheric is 0.8. In this research, the ageing coefficient is around
0.8 for fly ash concrete (CEM I with 30–50% fly ash, w/b ratio:
0.4, 0.5, and 0.6) under the laboratory condition as discussed in
Figs. 5 and 6. So the ageing coefficient, n, is chosen as 0.8 to predict
the service life of fly ash concrete under the condition of marine
c
atmospheric. Based on DuraCrete project, ke;0 is 0.68 under the
marine atmospheric environment. Because of the limitation of test
c
data in DuraCrete, the ke;c of blast furnace slag is used to present
c
the ke;c of fly ash, which is 2.9. The regression parameter Acs;cl is
c
4.42 for fly ash concrete. The curing factor kc;cl is 0.79 for the con-
crete after 28 days curing.
Furthermore, the partial factors relevant for structures in a mar-
ine environment are given in Table 4. Low cost of mitigation of risk
is employed in this research.
Based on above discussion, the input parameters are listed in
Table 5 to predict the service life of fly ash concrete.
The service life (tdi ) of fly ash concrete is calculated and listed in
Table 6, where two different reference time, t0, 28 days and
180 days were used, respectively.
With the reference time t0 = 28 days, the predicted service life
of fly ash concrete is only 33 years. When the reference time t0 -
= 180 days, the predicted service life of fly ash concrete is around
5 Experimental value 15.18 4.64 2.12 1.04 0.97
c c
(30% fly ash; w/
b = 0.4)
ð9Þ
Predicted value 9.39 3.66 2.12 1.20 0.69
(t0 = 180 days;
n = 0.8)
620 years. The predicted service life with 180 days as reference
time is about 18 times longer than that with 28 days as reference
time. It implies that the service life of fly ash concrete is underes-
timated when reference time (t0 = 28 days) is used as same as Port-
land cement concrete.
As mentioned above, many factors are related to the service life
of fly ash concrete, such as the cover thickness and exposure envi-
ronment. A thicker cover will always ensure a longer service life of
the concrete. However, for a certain designed service life of con-
crete structures in a given environment, the cover thickness can
be reduced when fly ash concrete is used.
3.6. Application-predicting the DRCM of fly ash concrete in laboratory
condition
In laboratory condition, all fly ash concrete specimens were
cured under 100% humidity lime bath environment until certain
curing age is reached for RCM test. The fly ash concrete has denser
structure than that exposed in field condition. Therefore, not only
the reference time has to be changed for fly ash concrete, but also
the age coefficient, n is needed to be modified in order to fit the
laboratory condition.
Based on the discussion in Section 3.5, the reference time is de-
fined as 180 days for fly ash concrete. The age coefficient, n, se-
lected for fly ash concrete is 0.8 [12]. The predicted chloride
migration coefficients of fly ash concrete are listed in Table 7.
Most of predicted chloride migration coefficients of fly ash con-
crete are smaller than experimental values. In the calculation, the
ageing coefficient, n, was chosen as 0.8. This value is advisable
for concrete blended with fly ash under the condition of above
ground or marine atmospheric [12]. However, in the laboratory
condition, no environmental loading is imposed. The ageing coeffi-
cient must be adjusted according to the environmental condition.
Several n values, 0.5, 0.8, and 1.0 are used to re-calculate the DRCM
by using Eq. (2) and to compare with experiment results as plotted
in Fig. 10.
The mean percentage relative error (MPRE) is calculated by Eq.
(10) and listed in Table 8. It is find that the predicted line t0 -
= 180 days; n = 1 has the minimum calculation error compared
with experimental data.

Table 5
Parameters from the DuraCrete note: (%): (%) relative to binder.

Parameters DuraCrete
c c c
cccr cccr Acs;cl ccs;cl xc Dx cRcl ke;0 ke;c ncl (n) kc;cl w=b

Value 1.03 0.2 4.42 1.2 45/35 8 1.50 0.68 2.9 0.8 0.79 0.4
Unit – (%) (%) – mm mm – – – – –
770 Z. Yu, G. Ye / Construction and Building Materials 48 (2013) 764–771

to predict the service life of fly ash concrete by DuraCrete model.

Chloride Migration Coefficients

1000
Based on the results obtained from this study, the following con-
clusions can be drawn:
100
(×10 -12 m2/s)

1. The addition of fly ash influences the chloride diffusivity. Unlike

Portland cement concrete, fly ash concrete shows an improve-
10 ment on the resistance to chloride ingress after 28 days.
2. The DRCM of fly ash concrete is mainly affected by w/b ratio. In
1 Experimental value FA3004 the double logarithmic coordinate system, the chloride migra-
t0=180 days; n=0.5
tion coefficient of fly ash concrete has a linear relationship with
t0=180 days; n=0.8
t0=180 days; n=1.0 the curing age. The w/b ratio has no obvious influence on the
0.1 slope, but more distinct influence on the initial chloride diffu-
10 100 1000 10000
sivity of fly ash concrete at curing age up to 2 years.
Curing age (days) 3. Like Portland cement concrete, there is a liner relationship
between DRCM and w/b ratio for fly ash concrete at curing age
Fig. 10. Experimental and calculated the DRCM value of fly ash concrete (FA3004).
up to 2 years.
4. To extend the application of DuraCrete model for fly ash con-
crete, the reference time, t0 = 180 days is determined. The pre-
Table 8
The mean percentage relative error. dicted results show that the fly ash concrete has much longer
service life when a reference time t0 = 180 days is applied com-
n 0.5 0.8 1.0
pared to that when t0 = 28 days is used like in Portland cement
MPRE 0.3040 0.2074 0.1328 concrete.

Acknowledgement
1000 No. MPRE
Chloride Migration Coefficients

FA3005 0.142 I hereby express gratitude to our technicians and colleagues in

FA3006 0.164 Microlab, Delft University of Technology, the Netherlands.
100 FA5004 0.196
FA5005 0.222
(×10 -12 m2 /s)

References
10 [1] Oskar Klinghoffer, Thomas Frølund, Ervin Poulsen. Rebar corrosion rate
Predicted Line-FA 3005 measurements for service life estimates. Paper presented at the ACI Fall
Experimental value FA3005 Convention, Toronto, Canada, Committee 365 ‘‘Practical Application of Service
Predicted Line-FA 3006 Life Models’’; 2000.
1 Experimental value FA3006 [2] Carola Edvardsen and Marianne Tange Jepsen. Chloride migration coefficients
Predicted Line-FA 5004 from non-steady-state migration experiments at environment-friendly
Experimental value FA5004
‘‘Green’’ concrete. In: Andrade C, Kropp J, editors, Proceedings of the 2nd
Predicted Line-FA 5005
Experimental value FA5005 international RILEM workshop ‘‘Testing and modelling the chloride ingress
0.1 into concrete’’; 2002. p. 203–09.
1 10 100 1000 10000 [3] Amrutha, Gopinatha Nayak, Narasimhan Mattur C, Rajeeva SV. Chloride-ion
impermeability of self-compacting high-volume fly ash concrete mixes. Int J
Curing age (days) Civ Environ Eng IJCEE–IJENS 2011;11(04):29–35.
[4] Cox K, De Belie N. Durability behavior of high-volume fly ash concrete. In:
Fig. 11. Experimental and predicted DRCM value of fly ash concrete. International conference on sustainable construction, materials and
technologies; 2007. p. 45–53.
[5] Zhang Wu-man, Ba Heng-jing, Chen Shang-jiang. Effect of fly ash and repeated
loading on diffusion coefficient in chloride migration test. Constr Build Mater
  2011;25(5):2269–74.
Measured value  Predicted value
Error ¼    100%; MPRE

[6] Collepardi S, Corinaldesi V, Moriconi G, Bonora G, Collepardi M. Durability of
Measured value high-performance concretes with pozzolanic and composite cements.
Durability Concr 2000;192:159–72.
X n
[7] Shafiq N. Effect of fly ash on chloride migration in concrete and calculation of
Error
cover depth required against the corrosion of embedded steel reinforcement.
1
¼ ð10Þ Struct Concr 2004;5(1):5–9.
n [8] Rilem Report 38: Durability of Self-Compacting Concrete. In: De Schutter G,
Audenaert, editors; April 2007.
[9] Tang L. Chloride transport in concrete-measurement and prediction, Doctoral
Chloride migration coefficient of other fly ash concrete mixtures thesis, publication P-96:6. Gothenburg, Sweden: Dept. of Building Materials,
with different w/b ratios presented in Fig. 4 are used to validate the Chalmers Universities of Technology; 1996.
[10] DuraCrete. Probabilistic Performance based Durability Design of Concrete
proposed reference time t0 = 180 days and ageing coefficient
Structures. Contract BRPR-CT95-0132, Project BE95-1347. Document BE95-
n = 1.0 in the lab condition. The predicted values, experimental 1347/R17; May 2000.
data and the MPRE between prediction and experiments are plot- [11] Polder RB, Rooij MR, de Breugel K van. Validation of models for service life
prediction – experiments from the practice. In: Proc. Int. RILEM-JCI seminar
ted in Fig. 11. It seems that the predicted values of chloride migra-
Concrete Life’06, Ein Bokek; 2006. p. 292–301.
tion coefficients are in good agreement with experiments. [12] Breugel K van, Polder R. Probability-based service life design of structural
concrete. In: Proc. 2nd Int RILEM Workshop on Concrete Durability and Service
Life Planning-ConcreteLife’09, Haifa, Israel; 2009. p. 383–91.
[13] Maage M, Helland S, Poulsen E, Vennesland O, Carlsen JE. Service life
4. Conclusions prediction of existing concrete structures exposed to marine environment.
ACI Mater J 1996;93(6):602–8.
In this paper, the resistance of fly ash concrete to chloride in- [14] de Rooij MR, Polder RB, van Oosten HH. Validation of service life performance
of in situ concrete by TEM and RCM measurements. HERON
gress was studied experimentally at curing age up to 2 years. The 2007;52(4):225–38.
reference time, t0 for fly ash concrete was determined and used [15] NEN-EN 206–1/NEN 8005: Concrete composition.
 Z. Yu, G. Ye / Construction and Building Materials 48 (2013) 764–771
[16] Sakai Etsuo, Miyahara Shigeyoshi, Ohsawa Shigenari, Lee Seung-Heun,
Daimon Masaki. Hydration of fly ash cement. Cem Concr Res 2005;35:
1135–40.
[17] Sybertz F, Wiens U. Effect of fly ash fineness on hydration characteristics and
strength development. Blended cements in construction. Papers presented at
the international conference, university of Sheffield, UK; 9–12 September
1991.
[18] Gallé C. Effect of drying on cement-based materials pore structure as identified
by mercury intrusion porosimetry: a comparative study between oven-,
vacuum-, and freeze-drying. Cem Concr Res 2001;31(10):1467–77.
[19] NT Build 492. Concrete, mortar and cement-based repair materials: chloride
migration coefficient from non-steady-state migration experiments. UDC
691.32/691.53/691.54; Approved 1999–11.
[20] Ghafoori N, Tays M. Abrasion resistance of early-opening-to-traffic portland
cement concrete pavements. J Mater Civ Eng 2007;19(11):925–35.
[21] Zhuqing Yu, Guang Ye. Chloride penetration and microstructure development
of fly ash concrete. In: The 2nd international conference ‘‘Microstructure
related durability of cementitious composites’’. Amsterdam; 11–13 April 2012.
[22] Mohammed TU, Yamaji T, Hamada H. Microstructures and interfaces in
concrete after 15 years of exposure in tidal environment. ACI Mater J
2002;99(4):260–352.
[23] Kuroda M, Watanabe T, Terashi N. Increase of bond strength at interfacial
transition zone by the use of fly ash. Cem Concr Res 2000;30(2):253–8.
[24] Zhang Yun-sheng, Sun Wei, Liu Zhi-yong, Chen Shu-dong. One and two
dimensional chloride ion diffusion of fly ash concrete under flexural stress. J
Zhejiang Univ – Sci A (Appl Phys Eng) 2011;12(9):692–701.
[25] Report. Chloride Resistance of Concrete. Cement Concrete and Aggregates
Australia; June 2009. p. 1–37.
771
[26] Chalee W, Jaturapitakkul C. Effects of W/B ratios and fly ash finenesses on
chloride diffusion coefficient of concrete in marine environment. Mater Struct
2009;42:505–14.
[27] Sengul O, Tasdemir C, Tasdemir AT. Mechanical properties and rapid chloride
permeability of concrete with ground fly ash. ACI Mater J 2005;102(6):414–21.
[28] Sujjavanich S, Sida V, Suwanvitaya P. Chloride permeability and corrosion risk
of high-volume fly ash concrete with mid-range water reducer. ACI Mater J
2005;102(3):177–82.
[29] Chindaprasirt P, Chotithanorm C, Cao HT, Sirivivatnanon V. Influence of fly ash
fineness on the chloride penetration of concrete. Constr Build Mater
2007;21(2):356–61.
[30] Concise Encyclopedia of Composite Materials. 2nd ed. In: Andreas Mortensen,
editor, ISBN: 139780080451268; 2007. p. 234.
[31] Glass GK, Buenfeld NR. The presentation of the chloride threshold level for
corrosion of steel in concrete. Corros Sci 1997;39:1001–13.
[32] Angst U, Vennesland Ø. Critical chloride content in reinforced concrete – State
of the art. In: Alexander et al., editors, Concrete repair, rehabilitation and
retrofitting II. Taylor & Francis Group, London, ISBN 978-0-415-46850-3; 2009.
P. 311–17.
[33] Arya C, Buenfeld NR, Newman JB. Factors influencing chloride-binding in
concrete. Cem Concr Res 1990;20(2):291–300.
[34] Dhir RK, Jones MR. Corrosion of reinforcement in concrete construction. In:
Page et al., editors, London: Elsevier Applied Science; 1990. p. 227–36.
[35] Thomas Michael. Chloride thresholds in marine concrete. Cem Concr Res
1996;26(4):513–9.
[36] European Standards for Reinforced Concrete: Eurocode 2 for the design for
strength, service and durability. Dr.-Ing. Hans-Ulrich Litzner, Berlin, Germany.