Construction and Building Materials 191 (2018) 912–922

Contents lists available at ScienceDirect

Construction and Building Materials

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

Compressive strength and durability properties of roller-compacted

concrete pavement containing electric arc furnace slag aggregate and fly
ash
My Ngoc-Tra Lam a, Duc-Hien Le b,⇑, Saravut Jaritngam c
a
Faculty of Civil Engineering, Ho Chi Minh City Open University, 97 Vo Van Tan, District 3, Ho Chi Minh City, Vietnam
b
Sustainable Developments in Civil Engineering Research Group, Faculty of Civil Engineering, Ton Duc Thang University, Ho Chi Minh City, Vietnam
c
Department of Civil Engineering, Faculty of Engineering, Prince of Songkla University, Hat Yai, Songkhla, Thailand

h i g h l i g h t s

EAF slag as coarse aggregate substitution decreases compressive strength and sulfate resistance of RCCP.

Replacing natural aggregate by EAF slag increases water absorption and abrasion resistance of RCCP.

20% fly ash as cement alternative improves compressive strength and durability properties of slag-RCCP at long-term age.

The weathering treatment process reduces the expansive potential of EAF slag.

EAF slag is non-active aggregate by alkali-silica reaction.

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

Article history: This study investigates the effects of electric arc furnace (EAF) slag aggregate and fly ash on compressive
Received 20 October 2017 strength and durability properties (i.e. water absorption, abrasion resistance, and sulfate resistance) of
Received in revised form 29 June 2018 roller-compacted concrete pavement (RCCP). The EAF slag aggregate as natural aggregate substitution
Accepted 13 October 2018
was used at three levels (i.e. 0%, 50%, and 100%) and cement was replaced by fly ash at three ratios
(i.e. 0%, 20%, and 40%). The EAF slag aggregate used in this study was exposed to outdoor condition for
several months to reduce the volume instability. The RCCP mixing proportions were determined by soil
Keywords:
compaction method. The compressive strength of RCCP was examined at 3-, 7-, 28-, and 91-day age.
Electric arc furnace slag aggregate
Fly ash
Meanwhile, the water absorption, abrasion resistance, and sulfate resistance of RCCP at 91-day age were
Roller-compacted concrete pavement used to study the durability properties. Furthermore, the length change of mortar bars made with EAF
Compressive strength slag aggregate was measured to evaluate the expansive potential of EAF slag caused by autoclave condi-
Durability properties tion and alkali-silica reaction (ASR) condition. Additionally, X-ray diffraction analysis was offered to iden-
Expansion tify the crystalline phases of mortar patterns under autoclave and ASR testing condition. The results
X-ray diffraction analysis presented that the compressive strength and sulfate resistance of RCCP containing EAF slag aggregate
declined slightly, whereas the water absorption and abrasion resistance increased in comparison to those
of traditional RCCP. Besides, the use of fly ash as cement substitution improved the compressive strength
of slag-RCCP at long-term age. A replacement of 20% fly ash provided the slag-RCCP fulfilling the strength
and durability requirements for pavement. In addition, the expansion in terms of length change of mortar
bars indicated that EAF slag aggregate after the proper treatment performed the volume stability under
the autoclave condition and ASR test.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction
Steel production in the world has been rapidly increased over
the last ten years, approximately 1560, 1650 and 1670 million tons
⇑ Corresponding author.
E-mail address: leduchien@tdtu.edu.vn (D.-H. Le).
of crude steel were produced in 2012, 2013, and 2014, respectively
[1]. Electric arc furnace (EAF) slag is a by-product generated from
steel manufacturing process, about 150–200 kg of EAF slag is pro-
duced from each ton of liquid steel. In Vietnam, approximately one
million ton of EAF slag was obtained in 2015 [1]. For the last dec-
ade, the use of EAF slag as natural aggregates in concrete has
become increasingly popular to reduce environmental pollution,

https://doi.org/10.1016/j.conbuildmat.2018.10.080
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
 M.N.-T. Lam et al. / Construction and Building Materials 191 (2018) 912–922 913

waste storage areas and save cost. EAF slag can replace natural expansive potential of EAF slag. Furthermore, X-ray diffraction
aggregates in many kinds of concrete, such as conventional con- (XRD) analysis was offered to identify the crystalline phases of
crete, high performance concrete, alkali-activated concrete [2–7]. mortar patterns under autoclave condition and ASR test.
For conventional concrete, Manso et al. [2] described that the com-
pressive strength of slag-concrete was higher than that of concrete
using natural aggregate. However, the durability of slag-concrete 2. Testing program
exposed to freezing and thawing cycles was lower than that of con-
2.1. Materials used
ventional concrete. Furthermore, the compressive strength after
subjecting to 30 cycles of wetting and drying of slag-concrete 2.1.1. Cementitious materials
decreased by 40% in comparison to that of conventional concrete. In this study, Ordinary Portland Cement (OPC) type I conforming to ASTM C150
Adegologe et al. [3] showed that compressive strength of slag- was used [13]. Type F fly ash with chemical composition as listed in Table 1 col-
lected from Southern Vietnam was used as partial cement replacement.
concrete increased by 9% as compared to referenced concrete. Ten-
sile strength increased by 3% when natural aggregate was replaced
by slag aggregate. Faleschini et al. [4] used EAF slag aggregate to 2.1.2. Aggregates
produce high performance concrete. Test results revealed that Natural aggregate used in this study with a gradation as shown in Table 2 is
the concrete made with EAF slag aggregate increased the strength crushed stone. EAF slag aggregate from Southern Vietnam with size from
4.75 mm to 19 mm was used as a natural aggregate substitution. EAF slag aggregate
in range of 35%–45% and improved the durability in terms of fulfilling the requirements of ASTM C33 standard is eligible to replace natural
exposing chloride. Alkali-activated concrete containing steel slag coarse aggregate (Table 3) [14–17]. However, the presence of unstable compounds
as coarse aggregate attained sufficient strength for structure, (i.e. free CaO, free MgO) in EAF slag could create deleterious behavior in concrete.
whereas the values of durability reduced slightly [7]. Although, When free CaO reacts with water to transform into Ca(OH)2 and free MgO reacts
with water to transform into Mg(OH)2, their volume has become about twice their
the mechanical and durability properties of concrete using EAF slag
original size leading to the expansion and cracking in concrete. Therefore, EAF slag
aggregate were attracted many research, few studies have investi- aggregate used in this study was left outdoor condition for several months to
gated the mechanical and durability properties of roller- reduce the volume expansion [18]. After treatment process, chemical composition
compacted concrete made with EAF slag aggregate [8]. of the EAF slag was examined following ASTM C25 standard [19] and free CaO con-
Roller-compacted concrete is a stiff and no-slump concrete, tent was determined following a test method of ministry of transportation of
Ontario [20] through ethylene glycol and complexometric titration as listed in
which is usually used for pavement due to its fast construction Table 4. Because the free CaO content is less than 0.1%, the EAF slag used in this
and cost efficiency [9]. Materials of roller-compacted concrete study may be hardly create cracking or unstable problems. Additionally, the volume
pavement (RCCP) are similar to those of conventional concrete, stability of EAF slag aggregate from hydration reactions was confirmed by the
whereas the gradation of RCCP is similar to that of hot-mix asphalt expansion test conforming to ASTM D4792 standard in the previous research [8,21].
pavement. RCCP mixture has a high volume of fine aggregate and a
low volume of binder, coarse aggregate and water. Recent studies
2.2. Mixing proportions
have concentrated on the recycling of fly ash as cement alternative
in RCCP to reduce expense and environmental impacts. Mardani- This study is composed of two phases. In first phase, mixing proportions of RCCP
Aghabaglou et al. [10,11] have used 20%, 40%, and 60% fly ash to designed by soil compaction method were reported in the previous research [8].
According to this method, the mixture with different moisture contents (e.g. 5%,
replace cement in RCCP designed by maximum density method.
6%, 7%, 8% and 9%) was compacted in the Proctor mold in accordance with ASTM
The results of this study reported that strength of RCCP decreased D1557 [22] to establish a compaction curve. From the compaction curves, the water
with increasing fly ash content in mixture at early ages. Besides, optimum content of each mixture was determined. The binder (cement + fly ash)
the cement replacement by fly ash resulted in the durability reduc- dosage used in this research was 12% of total weight of binder and aggregates,
tion. However, RCCP using proper fly ash content attained the sat- and fly ash substituted partially for cement at three ratios (i.e. 0%, 20%, and 40%).
Three types of coarse aggregate were prepared to produce RCCP, i.e. 100% crushed
isfactory strength and durability at later ages. In a similar study,
stone aggregate (group A), 50% crushed stone plus 50% EAF slag aggregate (group B),
Yerramala and Babu [12] have studied the transport properties of and 100% EAF slag aggregate (group C). As a result, nine mixing proportions of RCCP
RCCP containing fly ash high-volume. Fly ash as cement substitu- were shown in Table 5.
tion was used in the range of 40–85% of cement weight. It was
observed that these mixtures containing 60% to 70% of fly ash
Table 1
improved the permeability, absorption, sorption and chloride dif-
Chemical composition and physical properties of OPC and fly ash.
fusivity of RCCP as compared to referenced concrete. Meanwhile,
the incorporation of EAF slag as a natural coarse aggregate substi- OPC, type I Fly ash, class F
tution and fly ash as a cement replacement in RCCP has been dis- Chemical composition (%)
cussed previously [8]. The recycling of EAF slag aggregate and fly Silica (SiO2) 20.7 52.3
ash in RCCP could reduce the material cost, the environmental pol- Alumina (Al2O3) 4.5 24.9
Ferric oxide (Fe2O3) 3.3 14.1
lution, and the consumption of natural aggregate in concrete Calcium oxide (CaO) 63.0 –
industry. Therefore, the use of RCCP containing EAF slag aggregate Magnesium oxide (MgO) 1.8 –
and fly ash is one of the solution to develop sustainable Sodium oxide (Na2O) 0.10 0.67
infrastructure. Potassium oxide (K2O) 0.74 –
Sulphuric anhydride (SO3) 2.3 0.47
The main aim of this study is to investigate the effects of EAF
Loss on ignition (LOI) 2.8 0.15
slag aggregate and fly ash on compressive strength and durability
Physical characteristics
properties of RCCP. EAF slag aggregate was exposed to outdoor
Fineness (Blaine) (m2/kg) 347 289
condition for several months as a treatment process before using Specific gravity 3.14 2.40
to produce RCCP. In this work, RCCP mixing proportions were Initial setting time (min) 110 –
determined by soil compaction method. Compressive strength of Final setting time (min) 170 –
standard cylinder specimens was used to assess the mechanical Particle composition
Retaining on 45 mm sieve (%) – 7.92
properties of RCCP. Besides, the water absorption, abrasion resis- Compressive strength (N/mm2)
tance, and sulfate resistance are parameters to study the durability 1 day 14.6 –
properties of RCCP. In addition, the length change of mortar bars 3 days 26.2 –
containing EAF slag aggregate caused by autoclave condition and 7 days 33.0 –
28 days 43.0 –
alkali-silica reaction (ASR) test was measured to evaluate the
914 M.N.-T. Lam et al. / Construction and Building Materials 191 (2018) 912–922

Table 2 2.3. Testing approaches and sample preparation

The suggested limits and the designing gradation of RCCP used in this study.
2.3.1. Compressive strength
Sieve size Passing mass (%) Compressive strength of RCCP was determined at 3-, 7-, 28-, and 91-day age.
The lower The upper The designing Specimens with diameter of 150 mm and height of 300 mm were produced by a
limit limit gradation vibrating hammer conforming to ASTM C1435 [23]. Mixture was compacted in four
layers in molds and the compaction time was 20 s per layer. At 24 ± 8 h after cast-
19 mm 95 100 100 ing, all samples were removed from the molds and cured in water bath at the tem-
12.5 mm 70 95 83 perature of 23 ± 2 °C until testing day. At least 16 h before testing, the cylinder
9.5 mm 60 85 73 specimens underwent a capping procedure in accordance with ASTM C617 [24],
4.75 mm (No. 4) 40 60 54 in which sulfur mortar was used to give them a plane surface.
2.36 mm (No. 8) 30 50 41
1.18 mm (No. 16) 20 40 30
2.3.2. Water absorption
600 mm (No. 30) 15 30 22
Total water absorption of RCCP in this study was measured at 91-day age in
300 mm (No. 50) 10 25 15
accordance with BS EN 1338 standard [25]. The testing specimens (100 mm diam-
150 mm (No. 100) 2 16 10
eter and 200 mm height) were made by a vibrating hammer with two layers of
<75 mm (No. 200) 0 8 7
compaction and cured in water bath until 91 days. Before testing, all testing spec-
imens were removed dust and placed in a room at a temperature of 20 ± 5 °C for
2 days. After that, the specimens were immersed in water bath for at least 3 days.
The initial mass of specimens was recorded when the difference between first
and second weight performed at an interval of 24 h is less than 1%. Then, the spec-
Table 3 imens were placed in oven at a temperature of 105 ± 5 °C until they reached con-
Physical properties of crushed stone aggregate and EAF slag aggregate. stant mass. The total water absorption of specimen is the ratio of the difference
between initial and constant mass of sample and the constant mass of sample.
Properties Crushed stone EAF slag
Fine Coarse Coarse 2.3.3. Abrasion resistance
aggregate aggregate aggregate Abrasion resistance of RCCP was determined at 91-day age by wide wheel abra-
(0 – 4.75 mm) (4.75–19 mm) (4.75–19 mm) sion machine (Fig. 1) conforming to BS EN 1338 standard [25]. Eighteen samples
with diameter of 100 mm and height of 50 mm were prepared. The samples shall
Apparent specific gravity 2.71 2.72 3.40 be flat within a tolerance of ±1 mm, clean and dry. Each sample was put on the
Density (OD) (kg/m3) 2609 2680 3085 machine contact with wide abrasion wheel to measure abrasion resistance. After
Density (SSD) (kg/m3) 2644 2691 3176 the abrasion wheel of machine rotates 75 revolutions, the length of the groove pro-
Bulk density (kg/m3) 1493 1466 1686 duced on the sample was measured. The result of each RCCP mixture is the average
Water absorption (%) 1.36 0.44 2.93 of two readings on two different samples.
Los Angeles abrasion – 13.98 19.37
value (%)
2.3.4. Sulfate resistance
For assessing sulfate resistance, 27 cubic (100  100  100 mm) samples of
each RCCP mixture were prepared. After 91 days of curing, the samples were
divided in two groups; one group (12 cubic samples) was immersed in solution
of 5% sodium sulfate (Na2SO4), whereas the second group (15 cubic samples) was
cured continuously in water to study the influence of sulfate attack on the mass
Table 4 and compressive strength. The sodium sulfate solution was renewed every 2 weeks.
Chemical composition of the EAF slag aggregate. Every 2 weeks, mass and compressive strength change were determined. The mass
change is the difference between the average mass of samples placed in sodium sul-
Oxide composition wt. % fate solution and the average mass of samples placed in water at the same testing
CaO 25.94 time. Consequently, the mass change was calculated by Eq. (1)
P
Iron oxides 34.73 mS  mW
Mass change ð%Þ ¼  100 ð1Þ
SiO2 16.32 mW
MgO 6.86
Al2O3 8.31 where mW and ms are the average mass of samples placed in water and in sodium
MnO 5.18 sulfate solution, respectively (g).
TiO2 1.98 In addition, compressive strength change was measured by the change in com-
Na2O 0.30 pressive strength of samples immersed in sodium sulfate solution as compared to
K2O 0.10 that of samples cured in water at the same testing age. The compressive strength
P2O5 0.25 change was calculated by Eq. (2)
Free CaO <0.1% RS  RW
Compressiv e strength change ð%Þ ¼  100 ð2Þ
RW

Table 5
Mixing proportions of RCCP with optimum water content.

Group Mixture Binder Coarse aggregate Fine aggregate Optimum water

content
Fly ash Fly ash OPC EAF slag EAF slag Crushed stone Crushed stone
(%) (kg/m3) (kg/m3) (%) (kg/m3) (kg/m3) (kg/m3) (kg/m3)
A A00 0 0 274 0 0 842 1163 158
A20 20 54 217 0 0 835 1154 149
A40 40 108 161 0 0 830 1146 144
B B00 0 0 275 50 424 424 1172 182
B20 20 55 220 50 423 423 1171 175
B40 40 109 163 50 419 419 1158 168
C C00 0 0 282 100 867 0 1197 196
C20 20 56 225 100 865 0 1195 188
C40 40 112 168 100 863 0 1191 184

Note: Binder (cement + fly ash) is fixed at 12% of total weight of binder and aggregates.
 M.N.-T. Lam et al. / Construction and Building Materials 191 (2018) 912–922
Fig. 1. Sample on wide wheel abrasion machine and after testing.
where RW and RS are the average compressive strength of three samples cured in
water and immersed in sodium sulfate solution, respectively (MPa).
2.3.5. Expansion of mortar bars
2.3.5.1. Expansion by autoclave testing. EAF slag can be used as a good aggregate in
concrete after proper treatment which makes slag to become stabilization. After
subjecting to environmental condition for at least three months, the EAF slag used
in this study was indicative of innocuous expansion behavior in accordance with
ASTM D4792 standard [8,21]. Nevertheless, the hydration of free CaO is quicker
and easier than the hydration of free MgO. Lun et al. [26] have demonstrated that
the hydration of free CaO and free MgO almost accomplished under autoclave con-
dition. Therefore, the autoclave method was proposed in this part to observe the
expansive potential of EAF slag aggregate after treatment. The autoclave testing
was performed according to the procedure described in ASTM C151 standard
[27]. Accordingly, mortar bars with dimension of 25  25  250 mm after molding
24 h ± 30 min were removed moist room and measured immediately an initial
length for each bar. Next, the mortar bars underwent autoclave testing. And then,
their length changes were recorded. In this work, the EAF slag was crushed in a
jaw crusher. Fine EAF slag and natural aggregate passed the No.4 (4.75 mm) sieve
were collected for making mortar bars. The gradation of fine aggregates conformed
to ASTM C1260 standard [28] (Table 6). Three mortar mixtures (i.e. MA, MB, and
MC) (see Table 9) with three fine aggregate types were prepared (i.e. 100% natural
aggregate, 50% natural aggregate plus 50% EAF slag aggregate, and 100% EAF slag
aggregate). For each mortar mixture, the cement dosage was 262 g to produce
two specimens. The cement, aggregate and water mixing ratio was 1:2.75:0.485.
Furthermore, three patterns extracted from MA, MB, and MC mortar bars were ana-
lyzed XRD to examine the crystalline phases.
2.3.5.2. Expansion by alkali-silica reaction (ASR). Aggregates containing certain con-
stituents can react with alkali hydroxides. Products of ASR may be expansion due
to the water absorption. If products of ASR are high-swelling, they will result in
cracking of concrete. In present work, the potential ASR of EAF slag aggregate
was evaluated through the expansion of mortar bars made with EAF slag aggregate
following ASTM C1260 standard [28]. In addition, the effects of fly ash as a cement
substitution on expansive ASR mitigation were also investigated. The EAF slag was
crushed to meet the gradation for making samples (Table 6). Three fine aggregate
groups (i.e. 100% natural aggregate, 50% natural aggregate plus 50% EAF slag aggre-
gate, and 100% EAF slag aggregate) and three fly ash ratios replacing cement (i.e. 0%,
20%, and 40%) were prepared to produce mortar mixtures. As a result, a total of nine
mortar mixtures were provided, as shown in Table 7. The aggregate - binder
(cement + fly ash) ratio was 2.25 and the water -binder ratio was 0.47 by mass in
Table 6
Grading requirement conforming ASTM C1260.
Sieve size Mass, %
Passing Retained on
4.75 mm (No. 4) 2.36 mm (No. 8) 10
2.36 mm (No. 8) 1.18 mm (No. 16) 25
1.18 mm (No. 16) 600 mm (No. 30) 25
600 mm (No. 30) 300 mm (No. 50) 25
300 mm (No. 50) 150 mm (No. 100) 15
915
the mixture. In each mortar mixture, three testing specimens
(25  25  250 mm) were produced. After casting, the specimens were remained
in the molds for 24 ± 2 h in the moist cabinet. Next, the specimens were removed
from the molds and recorded the initial comparator reading. Each sample group
was placed in a storage container with sufficient water to totally immerse them.
Then, the containers were put in an oven at 80.0 ± 2.0 °C for a period of 24 h. After
24 h, the bars were removed the oven and taken the zero reading of each bar imme-
diately. After that, all specimens made with each mortar mixture were placed in a
container with 1 N NaOH solution and returned the oven at 80.0 ± 2.0 °C. Make sub-
sequent comparator readings of the specimens periodically. Additionally, XRD anal-
ysis of six patterns extracted from six mortar samples (i.e. MA00, MB00, MC00,
MA40, MB40, and MC40) were conducted to detect the crystalline phases.
Table 8 shows the testing approaches and the quantity of samples in this study.
3. Results and discussion
3.1. Compressive strength
Fig. 2 presents the compressive strength of RCCP mixtures at 3,
7, 28, and 91-day ages. At 3-day age, compressive strength of A00,
B00, and C00 was 25.87 MPa, 22.08 MPa, and 21.58 MPa, respec-
tively. There was a slight strength reduction in RCCP containing
EAF slag as coarse aggregate replacement. A strength decrease of
B00 was 14.65% and that of C00 was 16.58% in comparison to
A00. Similar to strength at 3-day age, the compressive strength
of RCCP using EAF slag aggregate also declined slightly at 7- and
28-day age. The strength reduction of RCCP containing EAF slag
caused by particles of limestone (see Fig. 3). Limestone (CaCO3) is
added to the steelmaking process to remove impurities such as
nitrogen, silicon, phosphorus, sulfur, etc. These impurities are
removed by combining with calcium oxide (CaO) produced from
heating process of limestone in the furnace to form slag. X-ray
diffraction analysis of EAF slag pattern shows complex crystalline
phases of slag such as iron oxide (FeO), calcium aluminum oxide
(CaO)12(Al2O3)7), calcium silicate Ca2(SiO4), and calcium iron alu-
minum oxide (Ca2((FeO922Al1O78) O5) (Fig. 4). On the other hand,
calcium oxide is unstable in the presence of moisture and converts
to calcium hydroxide (Ca(OH)2). Then, calcium hydroxide absorbs
carbon dioxide (CO2) from the air to revert back to limestone
(Fig. 5). The crystalline phases of reverted limestone were calcium
silicate (Ca2(SiO4)), calcium carbonate (Ca(CO3)), and iron calcium
oxide (Fe2CaO4) (Fig. 4). Although the compressive strength of
RCCP using EAF slag as coarse aggregate decreased in comparison
to that of RCCP using natural aggregate, the compressive strength
of B00 and C00 mixture at 28-day age exceeded the compressive
strength requirement for pavement. For instance, the B00 attained
the compressive strength of 41.46 MPa at 28-day age, an increase
916 M.N.-T. Lam et al. / Construction and Building Materials 191 (2018) 912–922

Table 7
Mixture proportions of mortars were used in ASR test.

Mortar ID. Aggregates (g) Binder (g)

Crushed stone EAF slag Cement Fly ash
Mortar made with 100% natural aggregate MA00 990 0 440 0
MA20 990 0 352 88
MA40 990 0 264 176
Mortar made with 50% natural aggregate plus 50% EAF slag aggregate MB00 495 495 440 0
MB20 495 495 352 88
MB40 495 495 264 176
Mortar made with 100% EAF slag aggregate MC00 0 990 440 0
MC20 0 990 352 88
MC40 0 990 264 176

Table 8
The testing approaches and the quantity of samples used in this study.

No. Testing items Sample dimension (mm) Quantity of sample Standards

1 Compressive strength 150  300, cylinder 108 ASTM C39
2 Water absorption 100  200, cylinder 18 BS EN 1338
3 Abrasion resistance 100  50, cylinder 18 BS EN 1338
4 Sulfate resistance 100  100  100 243 –
5 Expansion of mortar bars by autoclave testing 25  25  250 6 ASTM C151
6 Expansion of mortar bars by alkali-silica reaction 25  25  250 27 ASTM C1260
7 X-ray diffraction analysis – 11 –

RCCP using EAF slag aggregate and 40% fly ash as cement substitu-
tion can be considered to apply in subbase.

3.2. Water absorption

The total water absorption of RCCP was determined at 91 day-

Fig. 2. Compressive strength of RCCP mixtures at various ages.
of 48% as compared to 28 MPa that is the required value for pave-
ments in areas without freeze–thaw conditions, and an increase of
34% as compared to 31 MPa that is the required value for pave-
ments in areas exposed to freeze–thaw conditions [29]. And, the
C00 mixture attained the compressive strength of 35.56 MPa at
28-day age. This result has demonstrated that the feasibility of
using EAF slag to replace 100% natural coarse aggregate in RCCP.
Besides, a cement substitution by fly ash resulted in a compres-
sive strength decrease of RCCP at early age. At 3-day age, A20’s
strength was 22.52 MPa; B20’s strength was 17.45 MPa; C20’s
strength was 15.73 MPa. At 28-day age, compressive strength of
A20, B20, C20 was 42.44 MPa, 37.82 MPa, 30.07 MPa, respectively.
This observation has proved that the incorporation EAF slag aggre-
gate and 20% fly ash produced RCCP which was sufficient strength
for pavement application [29]. After 28 days, compressive strength
of mixtures containing fly ash improved significantly due to the
pozzolanic properties of fly ash. As shown in the graph, A20, B20
and C20 mixture reached the compressive strength of 58.40 MPa,
42.84 MPa, and 43.35 MPa at 91 days, respectively. Nevertheless,
when cement content was replaced by 40% fly ash, compressive
strength fell dramatically because fly ash is not as good as cement
in terms of strength contribution. At 28-day age, compressive
strength of B40 and C40 was 26.99 MPa and 23.20 MPa. Therefore,
age as presented in Fig. 6. It was observed that the total water
absorption of A00 mixture was 4.31%, whereas the total water
absorption of B00 and C00 mixture was 4.88% and 5.36%, respec-
tively, which grew by 13% and 24% when compared to A00 mix-
ture. The water absorption of EAF slag aggregate was higher than
that of natural aggregate leading to the water absorption increase
in B00 and C00 mixture. According to CEB-FIP [30], A00 and B00
mixture were classified as ‘‘average concrete”, whereas C00 mix-
ture was classified as ‘‘poor concrete”. In addition, the using of
20% fly ash as a cement substitution resulted in a water absorption
reduction. As shown in the Fig. 6, the water absorption values of
A20, B20, and C20 were 4.15%, 4.67%, and 5.14%, respectively. In
this case, fly ash as filler has contributed to reduce volume of large
pore structure in concrete [31]. Nevertheless, the replacement of
40% cement by fly ash led to increase the water absorption. The
similar observation was found in Yerramala et al. research [12].
3.3. Abrasion resistance
Abrasion resistance of RCCP at 91-day age was determined by
the dimension of groove created by wide wheel abrasion machine.
The average dimension of groove obtained from two different spec-
imens as shown in Fig. 7. The bigger the dimension of groove get
the lower abrasion resistance get. All grooves showed the dimen-
sion in range of 21–23 mm leading to all mixtures can be classified
as Class 3 [25]. Because of the rough texture of slag, the abrasion
resistance of group C was higher than that of group A. Besides,
abrasion resistance in each group had a strong relationship with
compressive strength [32]. Higher compressive strength resulted
in higher resistance. It was observed that A20 was the highest
resistance in group A and C20 was the highest resistance in group
C. Meanwhile, mixtures containing 40% fly ash (i.e. A40, B40, and
C40) performed the lowest resistance in its group.
 M.N.-T. Lam et al. / Construction and Building Materials 191 (2018) 912–922 917

Fig. 3. Particles of limestone (CaCO3) could be seen in a sample of RCCP containing EAF slag aggregate.

Fig. 4. X-ray diffraction analysis of limestone and slag patterns.

3.4. Sulfate resistance
3.4.1. Mass change
Mass change of nine RCCP mixtures exposed to sulfate environ-
ment is shown in Fig. 8. It can be seen that mass of samples with-
out fly ash (i.e. A00, B00, and C00) increased gradually in testing
period (8 weeks). A00 increased from 0.33% in 2 weeks to 0.38%
in 8 weeks, B00 increased from 0.37% in 2 weeks to 0.39% in
8 weeks, and C00 increased from 0.39% in 2 weeks to 0.46% in
8 weeks. This phenomenon may be due to swelling and filling up
the pores in samples by gypsum and ettringite produced by sulfate
attack reactions [33]. Mass increase in C00 with a faster rate was a
result of higher water absorption of EAF slag leading to produce a
high-volume of gypsum and ettringite. Similarly, there was a dra-
matical mass increase in mixtures containing 20% fly ash (i.e. A20,
B20, and C20) in first two weeks; a mass increase of A20, B20, and
C20 was 0.85%, 1.04%, and 1.03%, respectively. Fly ash hardly
affected on sulfate resistance in early period of exposure because
the pozzolanic reaction occurred slowly. After that, the mass
increase rate of samples reduced steadily. In this case, fly ash reacts
with portlandite to produce calcium silicate hydrate, thereby
reducing the gypsum and ettringite formation [34]. In contrast,
918 M.N.-T. Lam et al. / Construction and Building Materials 191 (2018) 912–922

Fig. 8. Mass change of RCCP exposed to sulfate environment.

Fig. 5. EAF slag aggregate and reverted limestone.

Fig. 6. Total water absorption values of RCCP at 91-day age.

Fig. 9. Cracking image of C40’s sample after exposing sulfate environment.

3.4.2. Compressive strength change

Fig. 10 shows compressive strength change of RCCP exposed to
sulfate solution. For mixtures without fly ash, the strength change
trend of mixtures using EAF slag aggregate (i.e. B00 and C00) was
similar to that of mixture using natural aggregate (A00). A strength
development occurred within first four weeks. This is a result of
gypsum and ettringite formation leading to fill up the pores of
samples in early phase of exposure [35]. After that, the compres-
sive strength reduction was obtained owing to the loss of adhesion
and stiffness caused by expansion force of ettringite [36]. Because
EAF slag aggregate absorbs water higher than natural one, the

Fig. 7. Abrasion resistance of RCCP at 91-day age.

the B40 and C40 manifested a mass decrease in early period of

immersion. It is well known that concrete with a high strength
resists sulfate attack better than concrete with a low strength.
Because of the low compressive strength of B40 and C40 samples,
the formation of gypsum and ettringite during the sulfate attack
caused a loss of bond between the cement paste and aggregate
leading to cracking of samples at the cube edges. The cracking of
C40’s sample at the cube edges was observed after the first 2 week
of immersing period (Fig. 9). This is a main cause of a mass loss of
B40 and C40 for all immersing periods. Fig. 10. Compressive strength change of RCCP exposed to sulfate environment.
 M.N.-T. Lam et al. / Construction and Building Materials 191 (2018) 912–922 919

sulfate solution absorption of slag-RCCP (i.e. B00 and C00) was
higher than conventional RCCP (A00). Therefore, sulfate environ-
ment affected significantly in mixtures containing slag aggregate.
Consequently, C00 mixture displayed the least sulfate resistance.
Similar to mixtures containing fly ash, there was a continuous
strength development in early exposing period. The partial cement
replacement by 20% fly ash improved sulfate resistance of RCCP,
regardless of amount of EAF slag aggregate in mixtures. The poz-
zolanic reactions consumed portlandite leading to decrease the
gypsum and ettringite formation would a cause of this observation.
As shown in the Fig. 10, A20 mixture exhibited the best resistance
to sulfate attack. The effect of fly ash on B20 and C20 mixture dis-
played similar behavior of A20 mixture. Nevertheless, when 40% fly
ash was used to replace cement, compressive strength of RCCP
decreased dramatically. The low compressive strength of mixtures
containing 40% fly ash is a main cause of the less resistance. Indeed,
the C40 mixtures were observed the cracking on the samples after
2 weeks of immersion (Fig. 9). The manifestation of cracking on the
sample owing to the loss of bond between the cement paste and
aggregate produced by gypsum and ettringite.
3.5. Expansion of mortar bars
sion of MA mixture occurred due to the hydration of CaO which is
often present in cement only. Meanwhile, the expansion of MC
mixture related to the expansion of EAF slag aggregate [8] and
the hydration of CaO in cement. As a result, the intensity of port-
landite (Ca(OH)2) observed at 18° in XRD analysis of MC’s pattern
was the highest value (Fig. 11). Therefore, the natural aggregate
substitution by EAF slag aggregate resulted in an increase of expan-
sion; an increase of MB was 13%, and an increase of MC was 50% in
comparison to control mixture (MA). The XRD analysis revealed
that the crystalline phases of mortar made with natural aggregate
after autoclave condition were portlandite, calcite, quartz, albite,
microcline and clinochlore. Meanwhile, portlandite, calcite, larnite
and wustite were the crystalline phases of mortar made with EAF
slag aggregate.
In addition, the picture of mortar bars after testing (Fig. 12)
shows many pop-outs in MB and MC sample. Kuo et al. [37]
explained that chemical composition of pop-out was SiO2, MgO,
Ca(OH)2 and Mg(OH)2. Additionally, all specimens were not broken
when the entire hydration reactions took place under autoclave
condition [26], so it can be said that the EAF slag is volume stability
and can be replaced natural aggregate in RCCP after leaving several
months in environmental condition.

3.5.1. Expansion by autoclave testing 3.5.2. Expansion by alkali-silica reaction (ASR)

Expansion in terms of length change of mortar bars between
before and after autoclave testing is listed in Table 9. The expan-
sion values of MA, MB, and MC mixture were 0.038%, 0.043% and
0.057%, respectively. MC mixture was the highest expansion value,
whereas MA mixture was the lowest expansion value. The expan-
Fig. 13 shows the expansion values in terms of length change of
mortar bars caused by alkali-silica reaction (ASR). It can be seen
that 14-day expansion value of MC00 is 0.036% below the limit
value (0.1%) according to ASTM C1260, so EAF slag aggregate is
non-reactive aggregate. The reaction between the reactive silica

Table 9
Autoclave expansion results.

Sample ID. Test no. Expansion value (%)

Each mortar bar Average
Mortar bar made with 100% natural aggregate. MA MA-1 0.038 0.038
MA-2 0.038
Mortar bar made with 50% natural aggregate plus 50% EAF slag aggregate. MB MB-1 0.042 0.043
MB-2 0.044
Mortar bar made with 100% EAF slag aggregate. MC MC-1 0.057 0.057
MC-2 0.056

Fig. 11. X-ray diffraction analysis of MA’s, MB’s, and MC’s pattern after autoclave testing condition.
920 M.N.-T. Lam et al. / Construction and Building Materials 191 (2018) 912–922
Fig. 12. The mortar bars after autoclave testing condition.
Fig. 13. Expansion values of mortar bars immersing in 1 N NaOH solution.
Fig. 14. X-ray diffraction analysis of patterns after immersing 28 days in 1 N NaOH solution.
contained in the aggregates and alkali within the cement paste to
form alkali-silicate gel is called ASR. Therefore, the absence of SiO2
content in EAF slag aggregate leads to the ASR reduction of MC00
mixture (Fig. 4). The 14-day expansion result which was received
from accelerated mortar bar testing method (ASTM C1260) corre-
lated well with 52-week expansion result received from testing
method with concrete prism (ASTM C1293) [38]. Mortar bars con-
taining 100% EAF slag aggregate (MC00) with 14-day expansion
value less than 0.10% leads to EAF slag aggregate could exhibit
innocuous behavior under field condition. Consequently, the use
of EAF slag as aggregate in RCCP can be accepted. Moreover, the
14-day expansion values of MA00 and MB00 were 0.172% and
0.113%, it is essential to extend experiment until 28 days because
the expansion values vary in range of 0.1–0.2%. When the immers-
ing time was extended until 28 days, the expansion of MA00,
MB00, and MC00 was 0.32%, 0.21%, and 0.05%, respectively. Roger
[39] suggested that aggregate was classified as a reactive aggregate
when expansion value exceeded 0.15% at 14 days, 0.33% at 28 days,
and 0.48% at 56 days. In this case, a combination of 50% EAF slag
and 50% natural aggregate is considered a non-reactive aggregate.
Additionally, the results of XRD analysis (Fig. 14) of dried powder
samples extracted from mortar bars after 28-day testing revealed
that the crystalline phases of MA00 were portlandite, calcite, albite,
microcline, quartz, and clinochlore, whereas MC00 pattern dis-
played the presence of portlandite, larnite, and wustite.
In addition, the cement replacement by 20% fly ash decreased
the expansion due to alkali-silica reaction, especially in MA20
and MB20 mixture. The 14-day expansion values of MA20, MB20,
and MC20 were 0.058%, 0.044%, and 0.034%, respectively. A
decrease of 66% in MA20 in comparison to MA00; a decrease of
61% in MB20 in comparison to MB00. The pozzolanic properties
of fly ash contribute to ASR reduction [40]. The ASR mitigation of
fly ash maintained during the testing time (28 days). Furthermore,
the higher dosages of fly ash resulted in the higher reduction of
expansion. At 40% fly ash level of replacement, pozzolanic reaction
might consume almost of portlandite presented by the absence of
portlandite in the XRD analysis results of MA40, MB40, and MC40
 M.N.-T. Lam et al. / Construction and Building Materials 191 (2018) 912–922
(Fig. 14). It is known that portlandite is one of the composition
causing expansion [41]. The consumption of portlandite led to
decline the deleterious expansion.
4. Conclusion
1) Compressive strength of group A with natural aggregate is
higher than that of group B (with 50% EAF slag coarse aggre-
gate) and group C (with 100% EAF slag coarse aggregate).
This result related to the limestone which was added to
the steelmaking process to remove impurities. However,
RCCP containing 100% EAF slag as coarse aggregate replace-
ment reached the satisfactory strength for pavement appli-
cation. Moreover, a cement substitution by 20% fly ash
improved compressive strength of slag-RCCP at long-term
age. Increasing fly ash content more than 20% resulted in a
dramatical compressive strength reduction. Consequently,
the slag-RCCP containing 40% fly ash can be considered to
apply for sub-base construction.
2) A natural aggregate alternative by EAF slag leads to increase
the water absorption and abrasion resistance of RCCP. This
may be due to high water absorption and rough-surface tex-
ture of slag. With proper fly ash dosage (20%) as cement sub-
stitution, the slag-RCCP was improved significantly the
water absorption and abrasion resistance properties.
3) The sulfate resistance of RCCP made with EAF slag aggregate
is similar to that of conventional RCCP. There was a contin-
uous increase of compressive strength and mass in early per-
iod of exposure. This phenomenon may be resulted from the
gypsum and ettringite formation leading to fill up pores of
samples. After that, the loss of adhesion and stiffness caused
by expansion force resulted in the strength reduction. More-
over, similar to conventional RCCP, fly ash improved sulfate
resistance in slag-RCCP. However, the slag-RCCP containing
40% fly ash has a less resistance. It was observed that crack-
ing of B40’s and C40’s samples surface after exposing sulfate
environment. This occurrence could be as a result of a loss of
bond between aggregate and the cement paste caused by
gypsum and ettringite.
4) Under autoclave condition, the expansion in terms of length
change of mortar bars made with EAF slag aggregate is
higher than that of made with natural aggregate. The hydra-
tion of CaO both in cement and slag aggregate could be a
main cause for this behavior. After autoclave testing, all
mortar bars were not broken. This result confirms that EAF
slag aggregate was left outdoor condition for several months
is volume stability and can be used in RCCP.
5) 14-day expansion value of mortar bar made with EAF slag
aggregate (MC00) caused by alkali-silica reaction was
0.036%. This result indicated that EAF slag aggregate is
non-active aggregate. XRD analysis result of MC00 shows
the presence of portlandite is a main cause of this expansion.
Additionally, fly ash mitigated the expansion of slag-mortar
bar by pozzolanic reactions.
Conflict of interests
The authors declare that there are no competing interests
regarding the publication of this paper.
Acknowledgments
This work was supported by the Higher Education Research
Promotion and the Thailand’s Education Hub for Southern Region
of ASEAN Countries Project Office of the Higher Education Commis-
921
sion (Contract No. TEH-AC 032/2015) and PSU.GS. Financial Sup-
port for Thesis scholarship. Besides, we gratefully acknowledge
the testing equipment support of Quality Assurance and Testing
Center 3 (QUATEST 3), Bien Hoa Industrial Zone 1, Dong Nai pro-
vince, Vietnam.
References
[1] World Steel Association, in: World Steel in Figures 2016, World Steel Assoc.,
2016, pp. 3–30.
[2] J.M. Manso, J.A. Polanco, M. Losañez, J.J. González, Durability of concrete made
with EAF slag as aggregate, Cem. Concr. Compos. 28 (6) (2006) 528–534.
[3] G. Adegoloye, A.L. Beaucour, S. Ortola, A. Noumowé, Concretes made of EAF
slag and AOD slag aggregates from stainless steel process: mechanical
properties and durability, Constr. Build. Mater. 76 (2015) 313–321.
[4] F. Faleschini, M. Alejandro Fernandez-Ruiz, M.A. Zanini, K. Brunelli, C.
Pellegrino, E. Hernandez-Montes, ‘‘High, performance concrete with electric
arc furnace slag as aggregate: Mechanical and durability properties”, Constr.
Build. Mater. 101 (2015) 113–121.
[5] S.I. Abu-Eishah, A.S. El-Dieb, M.S. Bedir, Performance of concrete mixtures
made with electric arc furnace (EAF) steel slag aggregate produced in the
Arabian Gulf region, Constr. Build. Mater. 34 (2012) 249–256.
[6] S. Monosi, M.L. Ruello, D. Sani, Electric arc furnace slag as natural aggregate
replacement in concrete production, Cem. Concr. Compos. 66 (2016) 66–72.
[7] N. Palankar, A.U. Ravi Shankar, B.M. Mithun, Durability studies on eco-friendly
concrete mixes incorporating steel slag as coarse aggregates, J. Clean. Prod. 129
(2016) 437–448.
[8] M.N.T. Lam, S. Jaritngam, D.H. Le, Roller-compacted concrete pavement made
of Electric Arc Furnace slag aggregate: Mix design and mechanical properties,
Constr. Build. Mater. 154 (2017) 482–495.
[9] ACI Committe 325, ‘‘ACI 325.10R-95 Report on Roller-Compacted Concrete
Pavements,” vol. 95, no. Reapproved, pp. 1–32, 2001.
[10] A. Mardani-Aghabaglou, K. Ramyar, Mechanical properties of high-volume fly
ash roller compacted concrete designed by maximum density method, Constr.
Build. Mater. 38 (1) (2013) 356–364.
[11] A. Mardani-Aghabaglou, Ö. Andiç-Çakir, K. Ramyar, Freeze-thaw resistance
and transport properties of high-volume fly ash roller compacted concrete
designed by maximum density method, Cem. Concr. Compos. 37 (1) (2013)
259–266.
[12] A. Yerramala, K. Ganesh Babu, Transport properties of high volume fly ash
roller compacted concrete, Cem. Concr. Compos. 33 (10) (2011) 1057–1062.
[13] ASTM C150-07, ‘‘Standard Specification for Portland Cement.”
[14] ASTM -07, ‘‘Standard Test Method for Density, Relative Density (Specific
Gravity) and Absorption of Coarse Aggregate”.
[15] ASTM C128-07a, ‘‘Standard Test Method for Density, Relative Density (Specific
Gravity), and Absorption of Fine Aggregate”.
[16] ASTM -06, ‘‘Standard Test Method for Resistance to Degradation of Small-Size
Coarse Aggregate by Abrasion and Impact in the Los Angeles Machine.”
[17] ASTM /C33M-08, ‘‘Standard Specification for Concrete aggregates”.
[18] D. Le, Y. Sheen, Q. Bui, An assessment on volume stabilization of mortar with
stainless steel slag sand, Constr. Build. Mater. 155 (2017) 200–208.
[19] ASTM , ‘‘Standard Test Methods for Chemical Analysis of Limestone, Quicklime
, and Hydrated”.
[20] ‘‘Method of Test for Determination of Free Lime Content,” Ministry of
transportation, Ontario, Test method LS-622, Rev.no.16, Date 96 05 01, pp.
1-3.
[21] ASTM -06, ‘‘Standard Test Method for Potential Expansion of Aggregates from
Hydration Reactions”.
[22] ASTM -12, ‘‘Standard Test Methods for Laboratory Compaction Characteristics
of Soil Using Modified Effort (56 , 000 ft-lbf / ft 3 (2 , 700 kN-m / m 3)) 1”.
[23] ASTM /C1435M-08, ‘‘Standard Practice for Molding Roller-Compacted
Concrete in Cylinder Molds Using a Vibrating Hammer.”
[24] ASTM /C617M  12, ‘‘Standard Practice for Capping Cylindrical Concrete
Specimens”.
[25] BS EN 1338:2003, ‘‘Concrete paving blocks-Requirements and test methods”.
[26] Y. Lun, M. Zhou, X. Cai, F. Xu, Methods for improving volume stability of steel
slag as fine aggregate, J. Wuhan Univ. Technol. Mater. Sci. Ed. 23 (5) (2008)
737–742.
[27] ASTM C151-05, ‘‘Standard Test Method for Autoclave Expansion of Hydraulic
Cement”.
[28] ASTM C1260-07, ‘‘Standard Test Method for Potential Alkali Reactivity of
Aggregates (Mortar-Bar Method)”.
[29] American Concrete Pavement Association, ‘‘Roller-Compacted Concrete
Pavements as Exposed Wearing Surface,” pp. 1–29, 2014.
[30] CEB-FIP, Diagnosis, and assessment of concrete structures-state of art report,
CEB Bull 192 (1989) 83–85.
[31] M.D.A. Thomas, The effect of curing on the hydration and pore structure of
hardened cement paste containing pulverized fuel ash, Adv. Cem. Res. (1989)
181–188.
[32] S.K. Rao, P. Sravana, T.C. Rao, Investigating the effect of M-sand on abrasion
resistance of Fly Ash Roller Compacted Concrete (FRCC), Constr. Build. Mater.
118 (2016) 352–363.
922 M.N.-T. Lam et al. / Construction and Building Materials 191 (2018) 912–922
[33] M. Tennich, M. Ben, A. Kallel, Behavior of self-compacting concrete made with
marble and tile wastes exposed to external sulfate attack, Constr. Build. Mater.
135 (2017) 335–342.
[34] M. Sumer, Compressive strength and sulfate resistance properties of concretes
containing Class F and Class C fly ashes, Constr. Build. Mater. 34 (2012) 531–
536.
[35] S.T. Lee, Sulfate attack and role of silica fume in resisting strength loss, Cem.
Concr. Compos. 27 (2005) 65–76.
[36] W. Ouyang, J. Chen, M. Jiang, Evolution of surface hardness of concrete under
sulfate attack, Constr. Build. Mater. 53 (2014) 419–424.
[37] W. Ten Kuo, C.Y. Shu, Y.W. Han, Electric arc furnace oxidizing slag mortar with
volume stability for rapid detection, Constr. Build. Mater. 53 (2014) 635–641.
[38] W. Touma, D. Fowler, and R. Carrasquillo, ‘‘Alkali-silica reaction in Portland
cement concrete: testing methods and mitigation alternatives.”
[39] C.A. Rogers, Alkali-aggregate reactivity in Canada, Cem. Concr. Compos. 15
(1993) 13–19.
[40] M.H. Shehata, M.D.A. Thomas, The effect of fly ash composition on the
expansion of concrete due to alkali-silica reaction, Cem. Concr. Res. 30 (7)
(2000) 1063–1072.
[41] R.F. Bleszynski, M.D.A. Thomas, Microstructural studies of alkali-silica reaction
in fly ash concrete immersed in alkaline solutions, Adv. Cem. Mater. 7355 (97)
(1998).