Construction and Building Materials 114 (2016) 517–527

Contents lists available at ScienceDirect

Construction and Building Materials

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

Experimental study on strength and durability of lightweight aggregate

concrete containing silica fume
Kwang-Soo Youm a, Jiho Moon b, Jae-Young Cho c, Jung J. Kim d,⇑
a
Infra Structure Team, Technical Division, GS E&C, Grand Seoul, 33 Jong-ro, Jongno-gu, Seoul 110-130, Republic of Korea
b
New Transportation Research Center, Korea Railroad Research Institute (KRRI), Uiwang-si, Gyeonggi-do, Republic of Korea
c
Infra & Offshore Team, Technical Division, GS E&C, Grand Seoul, 33 Jong-ro, Jongno-gu, Seoul 110-130, Republic of Korea
d
Department of Civil Engineering, Kyungnam University, 7Kyungnamdaehak-ro, Masanhappo-gu, Changwon-si 631-701, Republic of Korea

h i g h l i g h t s

The water absorption ratio of LWAC is proportional to that of the corresponding LWA.

The frost resistance capacity of LWAC depends on the types of LWA.

The chloride ion penetration resistance of LWAC depends on silica fume contents.

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

Article history: In this study, the mechanical property and durability performance of high strength lightweight aggregate
Received 9 September 2015 concrete (LWAC) with silica fume for 91 days were presented. LWACs were designed to have the design
Received in revised form 30 January 2016 compressive strength of 60 MPa at 28 days and the oven-dry density below 1900 kg/m3. Nine mixtures
Accepted 25 March 2016
with three aggregate types and silica fume replacement ratio of 0%, 3.5% and 7.0% by cement weight were
Available online 5 April 2016
prepared. The splitting tensile strength, the compressive strength and the modulus of elasticity tests were
conducted at 7 days, 28 days, 56 days and 91 days. The chloride penetration resistance tests were done at
Keywords:
28 days, 56 days and 91 days. The chloride diffusion coefficient based on the measurement of chloride
Lightweight aggregate
High strength concrete
penetration depth was also measured at 7 days, 14 days, 28 days, 56 days and 91 days. The rapid
Durability freeze-thaw cycling tests were conducted and the relative dynamic modulus was evaluated up to 300
Chloride penetration cycles. The results indicate that the durability against chemical deterioration for LWAC incorporated to
Freeze-thaw silica fume depends on the compositions of hardened cement pastes in concretes, while the durability
against physical attack depends on the types of aggregates.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction [2]. The high strength LWAC bridge was built with the average
measured the compressive strength of 73.5 MPa at 35 days, despite
Due to the relatively low density and high thermal insulating its design strength was 45 MPa at 28 days [3]. LWAC has been
capacity of lightweight aggregate concrete (LWAC) than those of found to be well suitable for the floating offshore platform since
normalweight aggregate concrete (NWAC), the applications of it helps to avoid an excessive hull displacement [4]. The durability
LWAC have been growing and increasing in many fields recently performance of several concrete offshore structures mainly below
[1–4]. For instance, energy saving houses constructed with LWAC 40 MPa was investigated after 20 year operation [5,6].
could decrease the heating energy consumption by more than The incorporation of pozzolanic materials in concrete has many
30% during winter season because of lower thermal conductivity beneficial effects to enhance the mechanical properties of concrete.
of LWAC than that of NWAC [1]. Since the dead weight of bridge The calcium silicate hydrates in cement matrix of concrete increase
could be reduced, many LWAC bridges with the design compres- by pozzolanic reactions. The fine pozzolanic particles fill spaces
sive strength of concrete as 55 MPa–60 MPa were built in Norway between clinker grains, thereby resulting in a denser cement
matrix and interfacial transition zone (ITZ) between cement matrix
and aggregates; this lowers the permeability and increases the
⇑ Corresponding author.
compressive strength of concrete. Seleem et al. [7] investigated
E-mail addresses: [email protected] (K.-S. Youm), [email protected]
the chloride penetration resistance of NWAC and showed that
(J. Moon), [email protected] (J.-Y. Cho), [email protected] (J.J. Kim).

http://dx.doi.org/10.1016/j.conbuildmat.2016.03.165
0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
518 K.-S. Youm et al. / Construction and Building Materials 114 (2016) 517–527

silica fume as a partial replacement of cement is the most effective replacement ratio of silica fume. The water-to-binder ratio was determined as
0.281 for NWAC, 0.259 for LWAC with Stalite, and 0.242 for LWAC with Argex.
in preventing chloride ingress among other pozzolanic materials.
The used cement contents were varying from 539 kg/m3 to 640 kg/m3 according
Leng et al. [8] showed that the chloride ion diffusion coefficient to the coarse aggregate types. The high range water-reducing superplasticizer of
of concrete decreased by replacing cement of pulverized fuel ash. polycarboxylate was added at constant rate of 1.0% of the binder content weight.
Nonetheless, there was few published study on the durability per- The reason why changes water-to-binder ratio and cement contents is to make high
formance of high strength LWAC over 60 MPa [9]. The possibility of strength concrete above 60 MPa considering the characteristics of types of
aggregates.
high strength LWAC below 45 MPa was shown by three mixtures
of LWAC with the lightweight aggregate of scoria [10]. Liu et al.
2.3. Mixing and curing
[11] experimentally evaluated the chloride penetration resistance
for normal strength LWAC. Kockal et al. [12] examined the Lightweight aggregates were immersed in water before 24 h of casting. For each
enhancement of LWAC durability with a compressive strength up mixture, the specimens were cast and cured in water (20 ± 2 °C) until the time of
to 60 MPa using fly ash aggregate through rapid chloride ion pen- testing after demolding at 24 h. For each mixture, a total of 91–100  200 cylinders
(diameter  height) was prepared for the compressive strength, the splitting tensile
etration test and rapid freeze-thaw test. Furthermore, Youm et al.
strength, the modulus of elasticity and the chloride ion penetration tests. Six
[13] conducted the long-term experiments to find out the upper 100  100  400 mm prisms were prepared for the freeze-thaw cycling tests for
limit on LWAC strength or strength ceiling of LWAC with regard each mixture.
to the various types of lightweight aggregates.
Since the floating offshore plant has been emerging as a promis- 2.4. Test methods
ing business for the next generation in Korea, many Korean ship-
building and plant construction companies have drawn much 2.4.1. Physical and mechanical testing
The slump and air content of fresh concrete were recorded in accordance with
more attention to the merits and feasibility of LWAC as an offshore American Standard Test Method (ASTM) C143 and ASTM C231 respectively. Wet-
structural material. Therefore, it is urgently needed to evaluate the density of LWAC was measured in compliance with ASTM C567 at 28 days while
mechanical properties of high strength LWAC and investigate the oven-dry density was recorded after specimens were dried at 105 ± 5 °C according
durability performance of it in harsh marine environment. The to BS EN 206-1. The splitting tensile strength and the modulus of elasticity tests
were performed at 7 days, 28 days, 56 days and 91 days while the compressive
aim of this research is twofold. One is to evaluate the mechanical
strength was measured at 2 days, 5 days, 7 days, 28 days, 56 days and 91 days to
properties of high strength LWAC for the design compressive find out the early strength development as per ASTM C469, ASTM C39 and ASTM
strength of 60 MPa at 28 days and the oven-dry density below C469 respectively.
1900 kg/m3. Two types of lightweight coarse aggregates for high
strength LWAC were selected from the previous experimental 2.4.2. Resistance to chloride penetration testing
results [13], where LWAC with the selected lightweight aggregates Two kinds of chloride penetration tests were implemented to investigate the
chloride penetration resistance of LWAC and NWAC. One is rapid chloride ion pen-
have recorded the first and second highest compressive strength at
etration testing (RCPT) in accordance with ASTM C1202 [14]. This method is widely
28 days among other five different LWACs. The other aim of this used and adopted as a standard test in order to measure not just chloride ions but
study is to investigate the durability of high strength LWAC with all the ionic movement by determining the total charge passing the specimen,
regard to the resistance on chloride ion penetration and freeze- expressed in coulombs. After cutting the specimen with 50 mm thick and
thaw cycling. Silica fume is used as a partial cement replacement 100 mm diameter, the cut slice was installed in the cell containing with 0.3 mol
NaOH solution and 3.0% NaCl solution on both sides of the slice. A potential of
material in order to reduce the permeability and increase the resis-
60 V is applied for six hours recording of the current at every 30 min. RCPT was
tance on harsh environment. done on six specimens for each mixture at the age of 28 days, 56 days and 91 days.
The other is Nordic Standard NT Build 492 [15] to determine the chloride diffu-
sion coefficient based on the measurement of chloride penetration depth. In order
2. Experimental programs to force the chloride ions to migrate into the concrete, an adjusting external electri-
cal potential is applied to the specimens filling in the cathode of 10% NaCl solution
2.1. Materials and anode of 0.3 mol NaOH respectively. After passing the specified duration, the
specimens are cut in half and 0.1 mol AgNO3 solution is sprayed to measure the
Two types of commercial lightweight coarse aggregates imported from the USA chloride penetration depth. Finally, the chloride diffusion coefficient can be esti-
and Belgium were selected for high strength LWAC based on the previous experi- mated by Eq. (1).
mental results [13]. Table 1 summarizes scanning electron microscope (SEM)
" rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi#
images, 3-Dimensional (D) and 2-D sliced images scanned by 3-D X-ray computed 0:0239ð273 þ TÞL ð273 þ TÞLxd
tomography and the physical properties of lightweight aggregates. Crushed Dnssm ¼ xd  0:0238 ð1Þ
ðU  2Þt U2
expanded clay aggregate, named ‘Argex’, has the particle density of 1130 kg/m3
and 24 h water absorption ratio of 14%. Crushed slate aggregate, named ‘Stalite’, where Dnssm is the non-steady-state migration coefficient in  1012 m2/s, U is the
has the particle density of 1470 kg/m3 as well as the water absorption ratio of 6%. absolute value of the applied voltage in V, T is the average value of the initial and
SEM images as shown in Table 1, ‘Argex’ and ‘Stalite’ have open pores and closed final temperature in the anolyte solution in °C, L is the thickness of the specimen
pores respectively. The particle size distributions for ‘Argex’ and ‘Satalite’ provided in mm, xd is the average value of the penetration depths in mm, and t is the test dura-
by manufacturer were presented in Fig. 1. Normalweight coarse aggregate with a tion in hour. NT Build 492 test were conducted on six specimens for each mixtures at
specific density of 2.65 was made from crushed granite in Korea. Two kinds of fine 7, 14, 28, 56 and 91 days. Considering that the amount of charge passed from the
aggregate materials were used as follows: one is clearly washed sea-sand with a experiments of RCPT is not a physical property of concrete, we try to link it to the
fineness modulus of 2.3 and the other is crushed fine aggregate to secure a reason- diffusion coefficient, which is a physical property of concrete, from the experiments
able consistency and minimum segregation. Type I Ordinary Portland Cement (OPC) of NT Build 492.
with a Blaine surface area of 3315 cm2/g and a specific density of 3.15 manufac-
tured in Korea and silica fume containing about 94% SiO2 with a BET specific surface
of 20.3 m2/g were used in this study. The particle size distribution for silica fume 2.4.3. Rapid freeze-thaw cycling testing
was presented in Fig. 2. It is recommended that the rapid freeze-thaw cycling test in accordance with
ASTM C330 [16], which modifies test procedures of ASTM C666 [17], should be
more suitable and reasonable in case of LWAC since this method reflects the char-
2.2. Mixture design acteristics of lightweight aggregate and LWAC [18]. The LWAC specimens dried in
for another 14 days to a relative humidity of 50% and a temperature of 20 °C prior
Nine concrete mixtures with three different coarse aggregates and silica fume to the freezing and thawing test after 14 days of water curing according to ASTM
contents were designed to obtain the design cylindrical compressive strength of C330, while the rapid freeze-thaw experiments on NWAC were performed after
60 MPa and the oven-dry density below 1900 kg/m3 at 28 days as summarized in 14 days in water prior to testing in compliance with the procedure A of ASTM
Table 2. The specimens were labeled ‘NN’, ‘St’ and ‘Ag’ according to the types of C666 in which the specimens were periodically frozen and thawed in water at
coarse aggregates of normal weight aggregate, Stalite and Argex respectively. the specified cycle per day. Both testing were decided at six cycles per day up to
Cement was partially replaced with silica fume of 0%, 3.5% and 7.0% by weight. 300 cycles and the freeze-thaw resistance of specimens was evaluated by the
The partial replacement ratio of silica fume with cement was denoted as ‘0.0’, changes in the relative dynamic modulus at every 30 cycles until its values dropped
‘3.5’ and ‘7.0’; for instance, ‘St-3.0’ represents the Stalite-based LWAC with 3.0% below 60% of the initial value.
Table 1
Types of lightweight coarse aggregates.

Aggregates X-ray CT images SEM images Raw Maximum size Particle density Water
material (mm) (kg/m3) absorption (%)
3D configuration 2D-sliced image
Argex Clay 8 1130 14.0

K.-S. Youm et al. / Construction and Building Materials 114 (2016) 517–527
Stalite Slate 13 1470 6.0

519
520 K.-S. Youm et al. / Construction and Building Materials 114 (2016) 517–527

Table 2
Design mixture for high strength lightweight concrete.

Specimen Aggregate types W/B (%) Mixtures [unit: kg/m3]

Cement Silica fume Water Coarse aggregate Fine aggregate
NN-0.0 Normal-weight 28.10 580.0 0.0 163 950.1 702.6
NN-3.5 28.10 559.7 20.3 163 946.5 700.0
NN-7.0 28.10 539.4 40.6 163 943.0 697.3
St-0.0 Lightweight (Stalite) 25.90 630.0 0.0 163 520.0 685.0
St-3.5 25.90 608.0 22.0 163 518.0 682.0
St-7.0 25.90 586.0 44.0 163 515.0 679.0
Ag-0.0 Lightweight (Argex) 24.22 640.0 0.0 155 530.7 696.8
Ag-3.5 24.22 617.6 22.4 155 528.5 693.9
Ag-7.0 24.22 595.2 44.8 155 526.2 690.9

Table 3
100
Test results of physical properties.

Specimen Slump (mm) Ari content (%) Density at 28 days

80
(kg/m3)
Passing %

Argex Wet Oven-dry

60
Stalite
NN-0.0 220 1.2 2448 2357
NN-3.5 220 1.0 2429 2346
40 NN-7.0 160 1.0 2397 2329
St-0.0 210 1.6 2014 1885
20 St-3.5 210 1.6 2009 1893
St-7.0 200 1.6 2001 1895

0 Ag-0.0 230 1.5 2053 1852

0 5 10 15 20 Ag-3.5 220 1.9 2038 1856
Ag-7.0 160 1.8 2006 1843
Sieve size (mm)

Fig. 1. The maximum and minimum grading curves for Satalite and Argex.
the fine particle size of silica fume with a particle mean diameter
of 0.61 lm and agreed with many other researches [21,22]. While
15
the wet-cured density of concrete gradually decreased according to
the increase of silica fume, the oven-dry density virtually con-
12 verged to constant values. The mean weight differences between
wet-cured and oven-dry densities in NWAC, LWAC with Stalite
Fractions %

9 and Argex were 80.7 kg/m3, 117.0 kg/m3 and 182 kg/m3, respec-
tively, as depicted in Fig. 3. The mean weight difference to the
wet-cured density ratios for NWAC, LWAC with Stalite and Argex
6
are 3.3%, 6.0% and 11.2% respectively. These values nearly matched
the water absorption ratio of each coarse lightweight aggregate as
3 presented in Table 1. The pore images for both lightweight aggre-
gates as shown in SEM images of Table 1, the independently closed
0 pores in Stalite is preventing water emigration during pre-soaking
0.01 0.1 1 10 stage, whereas the raw material and open pores in Argex have a
Particle diameter (µm) great influence on absorbing the water.
Fig. 2. Particle size distribution for silica fume.

2550
3. Results and discussion wet oven-dry

3.1. Physical properties 2350

Density (kg/m3)

Especially in LWAC, oven-dry or equilibrium density is an

important factor which can directly affect the strength in case of
2150
LWAC. This has already confirmed that the mechanical and physi-
cal properties of lightweight aggregate itself directly affect the
strength and permeability of LWAC [19,20]. The physical proper-
1950
ties of fresh concrete and density were measured and listed in
Table 3. The average slump was in the range of 200–230 mm
except 160 mm for the specimen containing 7% of silica fume. This
indicated that the higher dosage of superplasticizer should be 1750
St-0.0 St-3.5 St-7.0 Ag-0.0 Ag-3.5 Ag-7.0 NN-0.0 NN-3.5 NN-7.0
added to maintain constant slump for the concrete having the rel-
Specimens
atively high silica fume contents. The concrete mixture was also
sticky from the higher usage of silica fume. This is attributed to Fig. 3. Comparison of wet- and oven dry-density values.
 K.-S. Youm et al. / Construction and Building Materials 114 (2016) 517–527 521

3.2. Mechanical properties tures is estimated as raw materials of coarse aggregates; clay for
Argex, slate for Stalite and granite for normalweight aggregate.
Fig. 4 showed that the compressive strength rapidly increased Fig. 6 showed the measured modulus of elasticity. The modulus
at early stage up to 14 days. However, the rate of strength develop- of elasticity of all mixtures increased before 14 days and was grad-
ment was steady or reached to an ultimate strength as the age ually flattened until the testing age of 91 days. Although the dis-
increased. Table 4 summarized the compressive strength, the split- crepancy of the modulus of elasticity among aggregate types
ting tensile strength and the modulus of elasticity at 28 days and obviously confirmed, the increase in the silica fume replacement
91 days for reference. NWAC and LWAC with Stalite exhibited ratio had entirely no effect on the change of secant modulus. It is
the similar behavior as silica fume is added to the mixtures. The evident that modulus of elasticity is strongly linked to the aggre-
effect of silica fume on the compressive strength was clearly con- gate skeleton rather than the cement paste [23–26]. The increase
firmed when silica fume contents increased from 0% to 3.5%. It of normalweight coarse aggregate volume seemed to raise the elas-
was observed that the compressive strength of NWAC and LWAC tic modulus of NWAC since the test values in this study were close
with Stalite incorporated to silica fume were 8%–11% stronger than to the results measured at 365 days summarized in Table 4. On the
those without silica fume at the age of 28 days. In contrast, the other hand, the increase of lightweight coarse aggregate volume
usage of silica fume made no difference to LWAC with Argex. Con- resulted in reducing the elastic modulus of LWAC.
sidering the previous experimental results for LWAC with Argex
having w/c of 0.29 and cement contents of 530 kg/m3 show the
compressive strength of 56 MPa [13], the LWAC with Argex in this 3.3. Evaluation of code equations for predicting mechanical properties
study was designed to have the compressive strength over 60 MPa
as w/c as 0.25 and cement contents of 600 kg/m3. However, the Many equations for the modulus of elasticity and the splitting
LWAC with Argex in this study could reach the strength ceiling tensile strength derived from the design compressive strength
of about 50 MPa. have been proposed and used in many international codes [23–
The splitting tensile strength development of all mixtures was 25]. In order to calculate these two values for LWAC, the reduction
measured and plotted in Fig. 5. While there was no clear relation- or modification factors are carefully chosen considering the low
ship between the splitting tensile strength and silica fume con- density of LWAC and the volume fraction of lightweight fine and
tents, there was a slightly distinction of the splitting tensile coarse aggregates. For instance, CEB-FIP [27] uses the oven-dry
strength according to the types of aggregate. Many researchers density of LWAC, whereas American concrete institute (ACI) [28]
reached to the similar conclusions [22]. It is evident that the dom- and Korean concrete institute (KCI) [29] choose the equilibrium
inant cause of splitting tensile strength difference among all mix- density of LWAC.

(a) Stalite
80 (c) Normal
80
Compressive strength (MPa)

70
Compressive strength (MPa)

70

60
60

50
50

40
40
St-0.0 St-3.5 St-7.0
NN-0.0 NN-3.5 NN-7.0
30
0 20 40 60 80 100 30
Age (days) 0 20 40 60 80 100
Age (days)
(b) Argex
80
Compressive strength (MPa)

70

60

50

40
Ag-0.0 Ag-3.5 Ag-7.0

30
0 20 40 60 80 100
Age (days)

Fig. 4. Evolution of measured compressive strength values.

522 K.-S. Youm et al. / Construction and Building Materials 114 (2016) 517–527

Table 4
Average test results of mechanical properties [unit: MPa].

Specimen 28 days 91 days

Comp. strength Splitting tensile strength Modulus of elasticity Comp. strength Splitting tensile strength Modulus of elasticity
NN-0.0 68.6 5.09 36,245 71.6 5.31 36,404
NN-3.5 73.0 4.88 35,459 78.7 5.12 37,904
NN-7.0 74.2 5.32 36,286 76.2 5.60 37,604
St-0.0 65.4 4.27 27,398 69.8 4.31 28,620
St-3.5 71.0 4.73 26,474 71.1 4.51 27,460
St-7.0 72.3 4.47 26,350 71.6 4.52 28,294
Ag-0.0 46.1 3.58 19,019 46.0 3.89 20,896
Ag-3.5 47.0 3.60 19,925 51.0 3.22 21,022
Ag-7.0 47.9 3.66 19,689 49.5 3.45 20,971
N-50 [13] 65.1 – 33,431 86.9⁄ – 35,963⁄
S-50 [13] 59.6 – 26,650 71.4⁄ – 29,292⁄
G-50 [13] 55.4 – 24,630 60.4⁄ – 27,108⁄

1) N stands for NWC, S for LWAC with Stalite, G for LWAC with Argex.
2) 50 mean the design compressive strength of concrete mixture.
3) * means the measured values at 365 days.

(a) Stalite
6.0 (c) Normal
6.0
Splitting tensile strength (MPa)

Splitting tensile strength (MPa)

5.5
5.5
5.0
5.0
4.5
4.5
4.0
4.0
3.5
St-0.0 St-3.5 St-7.0 3.5
N-0.0 N-3.5 N-7.0
3.0
0 20 40 60 80 100 3.0
Age (days) 0 20 40 60 80 100
Age (days)
(b) Argex
6.0
Splitting tensile strength (MPa)

Ag-0.0 Ag-3.5 Ag-7.0

5.5

5.0

4.5

4.0

3.5

3.0
0 20 40 60 80 100
Age (days)

Fig. 5. Evolution of measured splitting tensile strength.

The modulus of elasticity of LWAC can be calculated by Eqs. (4) where q is the oven-dry density of LWAC in kg/m3, fcm is the mean
multiplying the reduction factor gE in accordance with CEB-FIP. value of compressive strength, fck is the characteristic strength
 1=3 defined for the various concrete grades and Df is 8 MPa.
f cm
Elc ¼ gE  20523  ð4-1Þ However, the modulus of elasticity of LWAC from ACI in Eq. (5)
10 and KCI in Eq. (6) taking account of the density between 1440 kg/
m3 and 2560 kg/m3 can be directly taken as
 q 2
gE ¼ ð4-2Þ
2200 qffiffiffiffi
0
Ec ¼ q1:5
e  0:043 fc ð5Þ
f cm ¼ f ck þ Df ð4-3Þ
 K.-S. Youm et al. / Construction and Building Materials 114 (2016) 517–527 523

(a) Stalite
40 (c) Normal
40
Modulus of elasticity (GPa)

35

Modulus of elasticity (GPa)

35
30
30
25
25
20
20
15
St-0.0 St-3.5 St-7.0 15
NN-0.0 NN-3.5 NN-7.0
10
0 20 40 60 80 100 10
Age (days) 0 20 40 60 80 100
Age (days)
(b) Argex
40
Modulus of elasticity (GPa)

35

30

25

20

15
Ag-0.0 Ag-3.5 Ag-7.0

10
0 20 40 60 80 100
Age (days)

Fig. 6. Evolution of measured modulus of elasticity.

qffiffiffiffi
0 of volumetric fractions, when a portion of the lightweight fine
E ¼ q1:5  0:077
3
fc ð6Þ
aggregate is replaced with normal weight fine aggregate.
0 Table 5 lists the modulus of elasticity and splitting tensile
where qe is the equilibrium density of LWAC,f c is the specified com-
strength predicted from code equations using the design compres-
pressive strength of concrete.
sive strength of 60 MPa compared with the experimental results at
CEB-FIP suggests another reduction factor to determine the
28 days. The calculated elastic modulus from CEB-FIP and ACI for
splitting tensile strength, fctm, of LWAC using following equations.
NWAC and LWAC with Stalite slightly overestimated the test
f ctm ¼ 0:3ðf ck Þ
2=3
ð7-1Þ results while Eq. (5) proposed by KCI tended to closely approach
the values. In case of LWAC with Argex, Table 5 showed that the
difference was evident for all equations. This indicated that the
f ctm ¼ 2:12  lnð1 þ 0:1  ðf ck þ Df ÞÞ ð7-2Þ
modulus of elasticity might be affected by more the volume change
 q  of coarse aggregate than the increase of total binder content and
gl ¼ 0:4 þ 0:6 ð7-3Þ reduction of dosage of plasticizer comparing the values listed in
2200
Table 4. Except for LWAC with Argex, it appears that the prediction
where Eq. (7-1) can be estimated for concrete grades 650 MPa, Eq. accuracy of proposed equations from these codes provides the
(7-2) for concrete grades >50 MPa, gl is a reduction factor for the applicable approximation in engineering practice, however,
tensile strength. slightly un-conservative. Meanwhile, all proposed equations for
ACI and KCI present the same following equation to calculate
the splitting tensile strength by introducing the modification factor
reflecting the reduced mechanical properties of LWAC to reduce
shear strength, friction properties, splitting resistance and increase Table 5
the development length of reinforcement compared with NWAC of Average modulus of elasticity and splitting tensile strength predicted from code
equations compared with the corresponding observations at 28 days [unit: MPa].
the same compressive strength.
qffiffiffiffi Types Measured CEB-FIP ACI 318-11 KCI (Eqs. 6
0 values model (Eqs. 4 (Eqs. 5 and 8) and 8)
f ct ¼ k  0:56 f c ð8Þ
and 7)

where k is 0.85 for sand-lightweight concrete containing only nor- Ec fctm Ec fctm Ec fct Ec fct
malweight fine aggregate and 0.75 for all-lightweight concrete con- Normal 35,997 5.09 38,881 4.35 38,771 4.67 35,095 4.67
taining only lightweight coarse and fine aggregates. Linear Stalite 26,740 4.49 28,726 3.99 28,670 3.97 25,947 3.97
Argex 19,544 3.61 23,314 3.58 23,261 3.01 22,524 3.01
interpolation between 0.75 and 0.85 shall be permitted, on the basis
524 K.-S. Youm et al. / Construction and Building Materials 114 (2016) 517–527

all mixtures conservatively underestimate the splitting tensile Table 6

strength below 17%. Chloride penetration testing results.

Specimen Total charge passed Chloride diffusion

3.4. Resistance to chloride penetration (Coulombs) coefficient ( 1012 m2/s)
28 days 91 days 28 days 91 days
Fig. 7 depicts the development of total charge passed for each NN-0.0 4612 2951 8.4 6.8
mixture conducted by RCPT in accordance with ASTM C1202 to NN-3.5 1402 899 2.9 2.3
evaluate the difference among types of coarse aggregate and the NN-7.0 565 326 1.4 0.8
silica fume replacement level. The resistance to chloride penetra- St-0.0 6450 2399 8.2 6.7
tion for all mixtures was increased with the silica fume content. St-3.5 2456 652 4.4 2.2
St-7.0 1166 242 2.5 1.1
It was well known that calcium hydroxide in hydrated cements
was converted to calcium silicate hydrates by reacting with sili- Ag-0.0 9751 3007 11.9 6.6
Ag-3.5 4146 910 7.1 4.2
cates in silica fume. This reaction was called as pozzolanic reaction
Ag-7.0 2075 372 4.9 2.2
[30]. Therefore, it could be considered as the resistance to chloride
penetration of concrete was affected by not only the coarse aggre-
gate types but also the microstructures of hydrated cements. As
presented in Fig. 7, the total charge passed amount decreased with
Fig. 8 shows the chloride diffusion coefficient of all mixtures
the increase of silica fume content and the curing age. Regardless
measured by NT Build 492 with coarse aggregate types and silica
of aggregate types, the charge passed amount of LWAC and NWAC
fume replacement ratio. It was calculated from the penetration
were converged at 91 days. This indicated that the main effect on
depth of chloride ion in all specimens using Eq. (1). As shown in
increasing the resistance to chloride ion penetration was the silica
Fig. 8, it is evident that the decreasing trends of all mixture were
fume content [31]. Meanwhile, the types of coarse aggregate and
very similar with that of RCPT due to the time dependence and
their raw materials affected the chloride permeability only at early
the increase of silica fume content [11]. These results indicated
ages. This chloride penetration testing results are listed in Table 6.
that there was significant effect of silica fume in case of the
Moreover, the results by RCPT were compared to the chloride per-
replacement ratio up to 3.5% whereas the gap between all mixtures
meability level referred in ASTM C1202. The charge passed for all
with 3.5%–7.0% silica fume contents was within a relatively small
specimens without silica fume had moderate chloride permeability
range. The diffusion coefficient of LWAC with Argex showed about
level at 91 days whereas specimens with silica fume content of
2–3 times higher than that of NWAC and LWAC with Stalite. Mean-
more than 3.5% corresponded to very low chloride permeability
while, when silica fume was added, all mixtures showed almost
level below 1000 C.

(a) Stalite
10000
(c) Normal
10000
Charge Passed (Coulombs)

Charge Passed (Coulombs)

1000
1000

St-0.0 St-3.5 St-7.0

100 NN-0.0 NN-3.5 NN-7.0
10 100 100
Ages (Days) 10 100
Ages (Days)
(b) Argex
10000
Charge Passed (Coulombs)

1000

Ag-0.0 Ag-3.5 Ag-7.0

100
10 100
Ages (Days)

Fig. 7. Evolution of chloride permeability measured by ASTM C1202.

 K.-S. Youm et al. / Construction and Building Materials 114 (2016) 517–527 525

(a) Stalite
20
(c) Normal
Diffusion coeffcient(x 10-12 m2/s)

St-0.0 St-3.5 St-7.0 20

Diffusion coeffcient(x 10-12 m2/s)

15 NN-0.0 NN-3.5 NN-7.0

15

10

10

0
0 20 40 60 80 100
Ages (days) 0
0 20 40 60 80 100

(b) Argex Ages (days)

20
Diffusion coeffcient(x 10-12 m2/s)

Ag-0.0 Ag-3.5 Ag-7.0

15

10

0
0 20 40 60 80 100
Ages (days)

Fig. 8. Evolution of chloride diffusion coefficient measured by NT492.

the same values of chloride diffusion coefficient. This behavior is ability. While LWAC with Argex showed lower mechanical proper-
due to the enhancement of pore distribution and microstructure ties compared with NWAC and LWAC with Stalite as shown in
by using silica fume over aging period and the discrepancy of Fig. 9(a), it showed similar permeability compared with others as
coarse aggregate types. The above results indicated that while depicted in Fig. 9(b). It is well-known that time-dependent
the effect of silica fume directly influenced the improvement of improvement in the reduction of permeability LWAC, called ‘inter-
durability against chemical deterioration by changing hardened nal curing effect’, is due to hydration of cementitious materials
cement microstructure in concrete, coarse aggregate types more provided by internal water available from the pores of lightweight
clearly affected the mechanical properties than concrete perme- aggregate [18,30,32–36]. This effect tends to be increased when

(a) Compressive strength (b) Chloride diffusion coefficient

80 20
Diffusion coeffcient(x 10-12 m2/s)

NN-7.0 Ag-7.0 St-7.0

Compressive strength (MPa)

70
15

60

10
50

5
40

NN-7.0 Ag-7.0 St-7.0

30 0
0 20 40 60 80 100 0 20 40 60 80 100
Age (days) Ages (days)

Fig. 9. Comparison of compressive strength and chloride diffusion coefficient at silica fume replacement ratio of 7.0%.
526 K.-S. Youm et al. / Construction and Building Materials 114 (2016) 517–527

(a) Stalite (c) Normal

10000
St-0.0 St-3.5 St-7.0 10000
Charge passed (Coulombs)

NN-0 NN-3.5 NN-7.0

8000

Charge passed (Coulombs)

8000

6000
6000

4000
4000

2000
2000

0
0.0 2.0 4.0 6.0 8.0 10.0 12.0 0
0.0 2.0 4.0 6.0 8.0 10.0 12.0
Diffusion coefficient (x 10-12 m2/s)
Diffusion coefficient (x 10-12 m2/s)
(b) Argex
10000
Ag-0.0 Ag-3.5 Ag-7.0
Charge passed (Coulombs)

8000

6000

4000

2000

0
0.0 2.0 4.0 6.0 8.0 10.0 12.0

Diffusion coefficient (x 10-12 m2/s)

Fig. 10. Linear correlation between measured chloride permeability by ASTM C1202 and chloride diffusion coefficient by NT 492 for tested specimens.

supplementary cementitious materials (in this study, silica fume) 110

are included in the mixture by accelerating the pozzolanic reaction
as this needs calcium hydroxide and available moisture. 100
Relave dynamic modulus

The total charge passed values are plotted against the chloride 90
diffusion coefficient for all mixtures from 28 days to 91 days
of elascity (%)

shown in Fig. 10. This showed excellent correlation between these 80

values with the R-squared value exceeding more than 0.9388 for 70
all mixtures summarized in Table 7. It is estimated that both ASTM Limit
charge passed and diffusion coefficient were a good indicator of 60
the chloride resistance of concrete. In case of NWC and LWAC with
50 St-0.0 St-3.5 St-7.0
Stalite using silica fume content of 3.5% and 7.0%, charge passed Ag-0.0 Ag-3.5 Ag-7.0
below 3000 C showed significant consistency with the chloride dif- 40 NN-0.0 NN-3.5 NN-7.0
fusion coefficient below 4.0  1012 m2/s and had the highly chlo-
30
ride resistance of them. 0 50 100 150 200 250 300
Freeze-thaw cycles
3.5. Resistance to freeze-thaw cycling
Fig. 11. Change in relative dynamic modulus of elasticity for tested specimens.

In this study, rapid freeze-thaw cycling test according to ASTM

C330 for LWAC and C666 for NWC was carried out to determine
the frost resistance of all mixtures regarding the effect of silica fume and aggregate types at 28 days, 56 days and 91 days.
Fig. 11 presented the relative dynamic modulus of elasticity of all
tested specimens. As the number of freeze-thaw cycles increased,
Table 7 LWAC with Stalite showed insignificant changes in the relative
Correlation between measured charges passed and chloride diffusion coefficient
dynamic modulus of elasticity. On the other hand, the value of
[Y = aX + b].
LWAC with Argex dropped rapidly below 60% of the initial value
Types of aggregate a b R2 before 150 freeze-thaw cycles without regard to the replacement
Stalite 922.44 958.51 0.9388 ratio of silica fume. It is noted that the difference of raw material
Argex 964.43 254.4 0.9576 of LWAC was estimated to the main cause of this results since
Normal 508.63 220.91 0.9493
Argex had higher water absorption ratio and open pores compared
 K.-S. Youm et al. / Construction and Building Materials 114 (2016) 517–527 527

to Stalite. It is estimated that the silica fume replacement level had [7] H.E.H. Seleem, A.M. Rashad, B.A. El-Sabbagh, Durability and strength
evaluation of high-performance concrete in marine structures, Constr. Build.
little effect on the resistance to freeze-thaw cycles. The value of
Mater. 24 (2010) 878–884.
NWC slightly decreased and pop-out or spalling was not observed [8] F. Leng, N. Feng, X. Liu, An international study on the properties resistance to
up to 300 cycles. diffusion of chloride ions of fly ash and blast furnace slag concrete, Cem. Concr.
Res. 30 (2) (2000) 989–992.
[9] Thomas A. Holm, T.W. Bremner, High-Durability Structural Low-Density
4. Conclusions Concrete for Application in Severe Marine Environments, Engineer Research
and Development Center, 2000, pp. 1–95.
[10] A. Kılıç, C.D. Atisß, E. Yasßar, F. Özcan, High-strength lightweight concrete made
This study presents the mechanical properties and durability with scoria aggregate containing mineral admixtures, Cem. Concr. Res. 33 (10)
evaluation of high strength LWAC with two types of lightweight (2003) 1595–1599.
coarse aggregates increasing the silica fume replacement ratio up [11] X. Liu, H. Du, M.-H. Zhang, A model to estimate the durability performance of
both normal and light-weight concrete, Constr. Build. Mater. 80 (2015) 255–
to 7.0% of cement weight.
261.
[12] N.U. Kockal, T. Ozturan, Durability of lightweight concretes with lightweight
1) Depending on the lightweight aggregate types, the differ- fly ash aggregates, Constr. Build. Mater. 25 (3) (2011) 1430–1438.
[13] K.S. Youm, Y.J. Jeong, E.S.H. Han, T.S. Yun, Experimental investigation on
ence between wet-cured and oven-dry density matched
annual changes in mechanical properties of structural concretes with various
the water absorption ratio of each lightweight aggregate. types of lightweight aggregates, Constr. Build. Mater. 73 (2014) 442–451.
2) The effects of silica fume increased the compressive strength [14] ASTM C1202, Standard Test Method for Electrical Indication of Concrete’s
of NWC and Stalite LWAC except that of Argex LWAC. Mean- Ability to Resist Chloride Ion Penetration, American Society for Testing and
Materials, 2003.
while, the splitting tensile strength and modulus of elasticity [15] NORDTEST, NT Build 492 Chloride Migration Coefficient from Non-Steady-
had nothing to do with the silica fume content regardless of State Migration Experiments, NORDTEST, Espoo, Finland, 1995.
aggregate types. [16] ASTM C330, Standard Specification for Lightweight Aggregates for Structural
Concrete, American Society for Testing and Materials, 2005.
3) The resistance to chloride penetration for all mixtures was [17] ASTM C666, Standard Test Method for Resistance of Concrete to Rapid Freezing
affected by the increase of silica fume. This was mainly and Thawing, American Society for Testing and Materials, 2003.
due to the addition of silica fume to intensify the microstruc- [18] T.A. Holm, Lightweight Concrete and Aggregates. STP 169D Significance of
Tests and Properties of Concrete and Concrete-Making Materials, ASTM
tural composition of hardened cement pastes of LWAC. International, 2006, pp. 522–532.
4) The rapid freeze-thaw cycling tests proved that the water [19] H.G. Russell, Lightweight Concrete – Material Properties for Structural Design.
absorption ratio and pore characteristics of lightweight Lightweight Aggregate Concrete Bridges, 2008, pp. 1–21.
[20] J. Mao, K. Ayuta, Freeze-thaw resistance of lightweight concrete and aggregate
aggregate mainly affected the frost resistance of LWAC irre-
at different freezing rates, J. Mater. Civ. Eng. 20 (1) (2006) 78–84.
spective of silica fume content. [21] M. Mazloom, A.A. Ramezanianpour, J.J. Brooks, Effect of silica fume on
5) Although the discrepancy of coarse aggregate types in con- mechanical properties of high-strength concrete, Cement Concr. Compos. 26
(4) (2004) 347–357.
crete resulted in the difference of mechanical properties of
[22] R. Siddique, Utilization of silica fume in concrete: review of hardened
concrete, it had little influence on the resistance to chloride properties, Resour. Conserv. Recycl. 55 (11) (2011) 923–932.
ion penetration. The difference of chloride ion penetration [23] P.-C. Aïtcin, High-Performance Concrete, E & FN Spon, London, 1998.
resistance resulted from the silica fume content level in con- [24] M. Suzuki, M.S. Meddah, R. Sato, Use of porous ceramic waste aggregates for
internal curing of high-performance concrete, Cem. Concr. Res. 39 (5) (2009)
crete. These results indicates that while the effect of silica 373–381.
fume directly influence the improvement of durability [25] W. Baalbaki, P.-C. Aïtcin, G. Ballivy, On prediction modulus of elasticity in high
against chemical deterioration by changing the microstruc- strength concrete, ACI Mater. J. 89 (5) (1992) 517–520.
[26] M.S. Meddah, S. Zitouni, S. Belaâbes, Effect of content and particle size
ture of hardened cement, coarse aggregate types more distribution of coarse aggregate on the compressive strength of concrete,
clearly affected the mechanical properties and concrete Constr. Build. Mater. 24 (4) (2010) 505–512.
durability against physical attack. [27] CEB-FIP, CEB-FIP Model Code for Concrete Structures 2010, Comité Euro-
International du Béton (2010).
[28] American Concrete Institute, ACI Committee 318: Building Code Requirements
for Structural Concrete and Commentary, ACI Standard, Farmington Hills, MI,
Acknowledgements 2009.
[29] Korean Standard Specification of Concrete, Ministry of Land, Transport and
Maritime Affairs of Korean Government, 2009.
This work was supported by the Techology Innovation Program [30] A.M. Nevile, Properties of Concrete, fifth ed., Pearson Education Limited,
funded by the Ministry of Trade, Industry & Energy (MI, Korea) (No. Edinburgh, England, 2011.
[31] T. Gupta, S. Chaudhary, R.K. Sharma, Assessment of mechanical and durability
10042556).
properties of concrete containing waste rubber tire as fine aggregate, Constr.
Build. Mater. 73 (2014) 562–574.
References [32] H.K. Kim, J.G. Jang, Y.C. Choi, H.K. Lee, Improved chloride resistance of high-
strength concrete amended with coal bottom ash for internal curing, Constr.
[1] T.S. Yun, Y.J. Jeong, T.S. Han, K.S. Youm, Evaluation of thermal conductivity for Build. Mater. 71 (2014) 334–343.
thermally insulated concretes, Energy Build. 61 (2013) 125–132. [33] K. Kohno, T. Okamoto, Y. Isikawa, T. Sibata, H. Mori, Effects of artificial
[2] Technical and Economic Mixture Optimisation of High Strength Lightweight lightweight aggregate on autogenous shrinkage of concrete, Cem. Concr. Res.
Aggregate Concrete, EuroLightCon, 2000. 29 (2) (1999) 611–614.
[3] Lightweight Concrete and the New Benicia-Martinez Bridge, HPC Bridge Views, [34] A. Bentur, S. Igarashi, K. Kovler, Prevention of autogenous shrinkage in high
2008. strength concrete by internal curing using wet lightweight aggregates, Cem.
[4] A.K. Haug, S. Fjeld, A floating concrete platform hull made of lightweight Concr. Res. 31 (11) (2001) 1587–1591.
aggregate concrete, Eng. Struct. 18 (11) (1996) 831–836. [35] M. Suzuki, M.S. Meddah, R. Sato, Use of porous ceramic waste aggregates for
[5] C. Ben, J. Gerwick, Construction of Marine and Offshore Structures, third ed., internal curing of high-performance concrete, Cem. Concr. Res. 39 (5) (2009)
CRC Press, 2007. 373–381.
[6] I. Holand, O.T. Gudmestad, E. Jersin, Design of Offshore Concrete Structures, [36] O.M. Jensen, P.F. Hansen, Water-entrained cement-based materials: I. Principle
Spon Press, NY, US, 2000. and theoretical background, Cem. Concr. Res. 31 (4) (2001) 647–654.