See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/276120695

Utilization of carbonated and granulated steel

slag aggregate in concrete

Article in Construction and Building Materials June 2015

DOI: 10.1016/j.conbuildmat.2015.03.008

CITATIONS READS

7 26

3 authors, including:

Bo Pang
Southeast University (China)
4 PUBLICATIONS 12 CITATIONS

SEE PROFILE

All content following this page was uploaded by Bo Pang on 29 May 2017.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
 Construction and Building Materials 84 (2015) 454467

Contents lists available at ScienceDirect

Construction and Building Materials

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

Utilization of carbonated and granulated steel slag aggregate in concrete

Bo Pang, Zonghui Zhou , Hongxin Xu
Shandong Provincial Key Laboratory of Preparation and Measurement of Building Materials, Engineering Center of Advanced Building Materials of Ministry of Education, University
of Jinan, Jinan 250022, China

h i g h l i g h t s

 The concrete aggregate is fully replaced by carbonated steel slag aggregate (CSA).
 Volume stability of concrete with CSA is satisfactory.
 The basic properties of concrete are improved by CSA.
 The results prove the possibility of using large number of steel slag as aggregate to instead of nature aggregate.

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

Article history: An experimental program was carried out to investigate the possibility of producing carbonated granu-
Received 2 December 2014 lated steel slag aggregate (CSA) to replace common natural aggregate (NA) like limestone, quartzite, etc.
Received in revised form 24 February 2015 and steel slag aggregate (SSA). Slump and compressive strength were tested to evaluate the fresh and
Accepted 4 March 2015
hardened properties of concrete with different aggregate. Mineralogy, morphology, pore structure and
Available online 30 March 2015
volume stability properties of aggregate and concrete were analyzed. Test results showed that carbona-
tion treatments can signicantly improve the strength and volume stability of CSA and reduce water
Keywords:
absorption, porosity and free calcium oxide. The workability of concrete with CSA was not signicantly
Carbonation
Steel slag
affected by the high water absorption of CSA. Besides, bleeding, and segregation were slighter and poros-
Aggregate ity of cement matrix was greatly reduced. After carbonation, harmful pores in aggregate reduced by 24.4%
Strength while harmless pores increased by 67.9%. The products were spindle and rod-like columnar calcium car-
Porosity bonate particles from SEM. Carbonation reduced the free CaO content from about 7 wt.% to less than
Volume stability 1 wt.% in 3 h. Compressive strength of concrete with CSA could be improved by 20% at 28 days.
Strength of CSA concrete exceeded SSA concrete at 60 days. The volume stability of CSA met the required
standards of autoclave chake and immersion expansion test.
2015 Elsevier Ltd. All rights reserved.

1. Introduction
Steel slag is a kind of solid waste produced in the process of
steel production which could be categorized as stainless steel slag
and carbon steel slag etc. according to the types of steel. Some
results showed that the recycling rates of steel slag in the US,
Germany and Japan had reached up to 50%, 31% and 25%, respec-
tively and most of them were used in molten iron and other indus-
tries [1,2]. The annual production of steel slag in China is about
100 m, however, the utilization rate of steel slag is only about
22%. Deposition of steel slag lead to the occupation of farm land
and pollution to the environment.
Corresponding author.
E-mail addresses: pbo1990@hotmail.com (B. Pang), mse_Zhouzh@ujn.edu.cn
(Z. Zhou), xuhx1990@hotmail.com (H. Xu).
The main composition of steel slag are CaO, SiO2, Fe2O3, MgO,
Al2O3, small amount of phosphorus and manganese oxides and
other impurities. Up to now, the main methods of using steel slag
were used as sinter material, hot metal dephosphorizing agent,
waste water treatment materials, reclamation of waste steel, con-
crete admixtures, CO2 capture, agriculture and aggregate in con-
struction etc. Typically, the crushed steel slag was used as coarse
aggregate directly, for example, all the recycled steel slag was used
as aggregate in concrete in Germany [3].
However, shapes, composition and structure of crushed steel
slag aggregate were difcult to control.
From the perspective of shapes and structure, a large number of
harmful pores and defects existed in both internal and surfaces of
steel slag, therefore causing poor workability, durability and com-
pressive strength of concrete. Anastasiou et al. carried out a
research showed that the use of ne construction and demolition

http://dx.doi.org/10.1016/j.conbuildmat.2015.03.008
0950-0618/ 2015 Elsevier Ltd. All rights reserved.
 B. Pang et al. / Construction and Building Materials 84 (2015) 454467
waste aggregate increased porosity and reduced strength and
durability of concrete, while the combination with steel slag aggre-
gate partly recovered strength and durability loss. Additionally, 28-
day compressive strength of the concrete with mixed construction
and demolition waste as ne aggregate and steel slag as coarse
aggregate only reached 30 MPa, adequate durability were shown
for only low grade applications [4]. Turhan Bilir carried out an
investigate that the use of non-ground slag and bottom ash as ne
aggregate generally increased permeability by increasing porosity
due to physical properties. These by-products as ne aggregate
could also reduce permeability of concrete due to their chemical
and mechanical properties in terms of permeability durability tests
[5]. In fresh concrete, the free water and air-voids were blocked
coming out from matrix due to irregular shapes of crushed steel
slag, then accumulated and deposited under the bottom of aggre-
gate which result in formation of holes and voids near the inter-
faces harden concrete, which eventually lead to the decrease in
strength and durability.
The composition aspect was most important, steel slag con-
tained about 50 wt.% of CaO (wherein more than 5 wt.% of free
CaO) and about 10 wt.% of MgO, the formation of Ca(OH)2 and
Mg(OH)2 causing volume expansion of 98% and 148%, respectively.
Besides, the uneven distribution of the RO phase (the phase with
the maximum density and hardness in steel slag which account
for about 2030 wt.% of steel slag) introduced defects in concrete
structure [6]. Thus, utilization of crashed steel slag as aggregate
directly resulted in an unstable service of concrete.
Thus, it was necessary to improve the durability and strength of
steel slag aggregate concrete. Carbonation was an effective way to
reduce the content of free calcium oxide and magnesium in steel
slay and enhance the strength and durability. The articial reefs
showed a high stability in seawater due to the consisted of
CaCO3, like shells and coral, and they acted as great breeding habi-
tats for seaweeds and coral [7,8].
In this study, the granulated steel slag aggregate made into
spherical shape of different sizes with all ne ground steel slag
was treated by accelerated carbonation process with a certain tem-
perature, humidity and CO2 pressure for 4 h [911].
2. Experimental study
2.1. Material
Shanshui 42.5 type Portland cement with the Chinese National Standard GB
175-1999, having Blaine neness of 311 m2/kg was used and the chemical composi-
tions of the cement are shown in Table 1.
For the manufacture of steel slag aggregate, steel slag which was a kind of Basic
Oxygen Furnace slag provided by Jinan Iron and Steel Works Company was used as
a raw material. The sample was collected from the open air storage site after crush-
ing and magnetic separation for iron and steel recovery. Chemical composition of
steel slag was illustrated in Table 1.
The particle size of the raw material was ranged from 2 mm to 8 mm. In order to
improve the pore size distribution and specic gravity of raw CSA, a certain amount
of y ash (FA) (5 wt.%) was used, the FA were provided from Shandong Huangtai
thermal power plant in which bituminous coal was used. FA with a specic gravity
of 2.12 and a specic surface of 276 m2/kg used in this study could be classied as F
type according to ASTM C618.
In order to compare the performance of normal aggregate concrete (NAC) to
steel slag aggregate concrete (SSC) and carbonated steel slag aggregate concrete
(CSC), natural coarse aggregate and ne aggregate were used. Crushed limestone
Table 1
Chemical composition of the cement and steel slag studied here.
CaO SiO2 Al2O3 Fe2O3 MgO SO3 K2O
(%) (%) (%) (%) (%) (%) (%)
Cement 55.85 22.91 7.12 3.36 3.28 2.30 0.69
Steel 46.73 14.77 5.52 18.42 6.27
slag
455
aggregate of 2.3618 mm in size was used as coarse aggregate. Natural river sands
with neness of 2.95 was used as ne aggregate. Polycarboxylic superplasticizer
(SP) with a water reducing rate of 40% was used to adjust the uidity of concrete.
2.2. Production of CSA
In the rst stage of the experimental program, raw slag was ground by a lab ball
mill for 30 min and sieved through a sieve of 600 lm. Then 5% y ash and 95% steel
slay powder were mixed by V-type mixing machine for 10 min. The amount of
sprayed water used during pelletization process has been determined as the coagu-
lant to form spherical pellets with the motion of rolling disc, and in this work, 20%
wt. water was used. The formation of pellets occurred between 10 and 12 min in
trial productions. The total pelletization time was determined as 20 min for the
compaction of fresh pellets [2527]. The untreated CSA was produced at 30 min
and then immediately dried in an oven at 70 C for 3 h. Additionally, parts of steel
slag powder were made into bricks (1  4  9 cm3) by compression molding in a
pressure of 8 MPa.
In the second stage, the dried CSA and the bricks were carbonated in a carbona-
tion reactor. The reactor had been sealed to 70 C and vacuumed to 0.3 MPa before
carbonation. Then CO2 was introduced into the reactor until the pressure reached
0.3 MPa. The produced CSA was sieved into fractions from 2.36 to 18 mm sizes as
coarse aggregates, carbonized bricks were crushed and sieved into fractions from
0.3 mm to 4.75 mm sizes as ne aggregates after carbonation. All the technological
process diagram was shown in Fig. 1.
2.3. Design and amendment of mix proportions
The packing of aggregate particles is closely related to the mesostructures of
concrete and mortar [12]. The depositional processes of concrete aggregate could
signicantly affect the performance of concrete on workability and service
[13,14]. Aggregate shapes and particle distribution were important factors to affect
the ITZ [15]. To get the best bulk density, bulk density for the best evaluation
method were used. Fuller curve proposed by William B. Fuller and Sanford E.
Thomson [16] in 1907 was used to calculate the mortar and concrete aggregate bulk
density required and to achieve maximum packing density regardless of undesir-
able workability or uidity, then the optimum distribution would be achieved.
Fuller curve is calculated as:
 n
d
Ud 100
D
U(d) cumulative volume of particles under d mm (%)
n correction factor
d particle size
D the maximum diameter of particles (mm)
The correction factor was modied from 0.5 to 0.33 by Andreasen [17,18] in
1930 when the irregular shape and polygon of natural rock aggregate were consid-
ered to reach the maximum of packing density.
In order to reduce the impacts of aggregate shape on the properties of concrete,
aggregate gradation was corrected by Fuller curve. The correction factors of the NA
and SSA were set as 0.33 while the factor of CSA was set as 0.5 (spherical coarse
aggregate) [19]. The particles with the size less than 0.6 mm were difcult to be
removed in experiment, so the original grading curve was amended (Fig. 2).
Before amendment, there was a great deviation between natural gradation curve
and Fuller curve. Besides, the offset in region of high particle diameter was getting
larger after Gauss tting. After amendment by selecting different correction factors,
aggregate gradation was signicant improved.
All the samples were forming in 100 mm3 molds and vibrating for 30 s. In order
to minimize the inuence of water absorption of aggregate on water and SP, all the
aggregate was added after uniformly mixing of water with other ingredients for
3 min (the properties of all the aggregate were shown in Table 2).
3. Results and discussion
3.1. The basic properties of the aggregate
From the mechanical aspect, the compressive strength of CSA
increased by four times after carbonation treat, but it was lower
Na2O MnO TiO2 P2O5 V2 O5 CaO free Ignition loss
(%) (%) (%) (%) (%) (%) (%)
0.22 0.19 1.25 1.44
2.76 1.18 1.67 0.90 6.52 3.04
456 B. Pang et al. / Construction and Building Materials 84 (2015) 454467
Fig. 1. The technological process diagram of aggregate.
compared with NA and SSA (shown in Fig. 3a). The compressive
strength of aggregate was obtained by separate measurement of
each aggregate compressive strength under press load while the
index of crushing was the results based on the weight percentage
of crushed aggregate (3 kg aggregate under a compression speed of
1 kN/s and pressurized to 200 kN, then the crushed aggregate was
sieved by 2.36 mm sieve). The strength of SSA was higher than NA
actually, but the index of crushing relatively was lower because of
its irregular edges and defects.
According to Berndt et al. [20], utilization of aggregate like RCA
(recycled concrete aggregate) with high water absorption (about
46%) would result in a very stiff admixture condition and almost
zero slump. And these concrete mixes had rapid workability loss
due to absorption of mix water by the aggregate and were difcult
to compact. However, from Table 3, there was no signicant
decline in slump, though the water absorption of CSA was much
higher than NA and SSA. Thus, the water absorption of different
aggregate with time was tested and the results showed that water
absorption of CSA was a slow process which lasted nearly 10 h
(shown in Fig. 3b).
After splitting the CSA specimens, the freshly split surface was
cleaned and sprayed with a phenolphthalein pH indicator. The
indicator used was a phenolphthalein 1% ethanol solution with
1 g phenolphthalein and 90 ml 95.0 V/V% ethanol diluted in water
to 100 ml [21]. From the early carbonation level, the carbonation
rate was relatively fast in rst 2 h and the CSA was almost com-
pletely carbonated in 3 h. The nal depth of carbonation was
between 2 and 3 cm in 4 h (shown in Fig. 3c).
Free CaO content was tested by glycol extraction and titration.
From the results in Fig. 3d, free CaO content was greatly reduced
by carbonation from about 7% wt. to less than 1% wt. in steel slag
at 3 h.
3.2. Mercury intrusion porosimetry
In order to study the impacts of carbonation reaction on the
internal structure (porosity) of the CSA, the mercury analysis was
used. Due to the unstable and complex composition distribution
of NA and SSA, a small amount of material structural analysis is
not representative and convincing, therefore they were not tested
by mercury intrusion porosimetry.
Mercury intrusion porosimetry tests (MIP) were conducted in
Quantachrome Pore Master 60GT with an applied intrusion pres-
sure range of 0.892 PSIA to 59,653.277 PSIA and pore diameter
range of 239.0670930.003576 lm while the extrusion pressure
range of 59,204.273 PSIA to 20.241 PSIA and pore diameter range
of 0.00360310.539027 lm at 20.0 C. For the MIP test, samples
weighing approximately 1.5 g was cut with a thin-saw cut from
the aggregate samples. The drying temperature was adopted in
order to avoid structural damage [22,23].
Xudong Chen et al. investigated the effect of water-to-cement
ratio (w/c) and age on the pore structure of cement mortar through
MIP and found that the degree of hydration increased with increas-
ing curing time and water-to-cement ratio of the cement mortar
for the ages of cement mortar varying between 14 and 180 days.
An increase in the water-to-cement ratio increased the total poros-
ity [24,25].
Shaopeng Wu et al. studied that the porosity of steel slag was as
large as about 24 times of basalt (5.76% and 0.24%, respectively),
which indicated that steel slag had a porous structure. Half of
pores size in steel slag were at a range of 0.010.1 mm, and most
pores of basalt were larger than 0.1 mm [26].
Wu et al. researched the inuence of pore size on the perfor-
mance and service of the concrete. In terms of the different effect
of pore size on concrete performance, the pore in concrete was
classied as harmless pore (<20 nm), less-harmful pore (20
50 nm), harmful pore (50200 nm) and more-harmful pore
(>200 nm) [27]. In order to analyze and compare conveniently,
the pore system of concrete was divided into three ranges (harm-
ful, less harmful and harmless) according to this sort method in
this study.
The results, as shown in Fig. 4a, indicated two levels of pore vol-
ume. The volume of harmful pores especially with the diameter
larger than 1 lm were decreased to some extent, whats more,
the volume of pore with the diameter from 10 nm to 200 nm were
also decreased, and they were all transformed into harmless pores
with diameter less than 10 nm. Thus, the performance of CSA was
further improved. After calculation, the harmful pores in aggregate
reduced by 24.4% while the harmless pores increased 67.9%.
The cumulative intrusion curves in function of the pore diame-
ter were shown in Fig. 4b. According to the volume of the mercury
intruded per unit mass of the sample, the internal air voids less
than 1 lm in CSA were great impacted by carbonation treatment
that the aggregate was twofold reduction in total volume of pores
less than 0.1 lm.
The pore size distribution curve (Fig. 4c) indicated a threshold
radius below which there was relatively little intrusion and imme-
diately above which rapid intrusion commences [28]. Relatively,
 B. Pang et al. / Construction and Building Materials 84 (2015) 454467 457

Fig. 2. Original grading curve and amended grading curve.

Table 2
Properties of all the aggregate.
diameter less than 0.01 lm. This suggested that the pore size in
this region could affect the permeability of the CSA at some level
Property Coarse Fine Coarse Fine Coarse Fine
[28].
NA NA CSA CSA SSA SSA
Carbonation reaction reduced the volume of internal air voids in
Rodded bulk density 1497 1929 2163 the aggregate, thus the shape, tortuosity and other properties of
(kg/m3)
Water absorption (%) 1.69 3.99 8.91 8.12 3.31 4.23
inside pores changed. It could be seen from Fig. 4d, pore shape fac-
(24 h) tor of the aggregate less than 1 lm was signicantly increased (the
Specic gravity (SSD) 2.49 2.56 2.86 3.31 3.42 3.58 smaller pore shape factor is, the shape of the pore is closer to ball
Water content (%) 2.31 3.44 2.38 2.72 1.56 2.80 shape) after carbonation, so the water absorption rate decreased
Powder content (%) 0.67 2.97 <0.1 <0.1 1.12 3.21
while its impermeability properties increased.
The test resulted of MIP in this study included the pore struc-
ture parameters such as total specic pore volume, most probable
pore diameter, bulk density, porosity, surface area, pore volume,
the pore volume of the diameter in range of 0.33 lm were permeability (neglecting the tortuosity effects) and the volume,
reduced, and these larger holes transformed into smaller pores surface area and pore number fraction at a certain pore diameter
which were reected by the increasing in pore volume of the (mean and median) which were shown in Table 4.
458 B. Pang et al. / Construction and Building Materials 84 (2015) 454467

Fig. 3. (a) The mechanism properties of aggregate, (b) time vs. water absorption of NA, CSA and SSA, (c) carbonate time vs. free-CaO content of CSA and (d) carbonate time vs.
carbonation depth of CSA.

Table 3
Mixture proportions of the concretes.

Concrete w/c Cement Water Fine aggregate (kg/m3) Coarse aggregate (kg/m3) SP Slump Bleeding and
(kg/m3) (kg/m3) (%) (mm) segregation
0.30.6 0.61.18 1.182.36 2.364.75 4.759.5 9.516 1619 19
26.5
NAC 0.42 375 158 149.36 112.02 224.04 242.71 242.71 298.72 317.39 280.05 3 205 Serious
CSC 0.42 375 158 380.07 152.03 481.42 430.75 329.40 126.69 329.40 304.06 3 200 Non
SSC 0.42 375 158 224.04 168.03 336.06 364.07 364.07 448.09 476.09 420.08 3 185 Slight

The results showed that the most probable pore size decreased
from 8.38 nm to 4.03 nm while the pore volume decreased from
0.2363 cc/g to 0.1488 cc/g. Performance and compactness of CSA
improved dramatically, especially the permeability (neglecting
tortuosity effects) decreased almost an order of magnitude (from
0.0655 nm to 0.008 nm) by carbonation. The surface area of the
particles inside the aggregate and the pore number fraction etc.
were also increased signicantly.
3.3. Mineralogy XRD analysis
The mineralogical composition was determined by powder XRD
analysis with Cu Ka radiation using a BRUKER D8 ADVANCE diffrac-
tometer equipped with a copper tube operated at 40 kV and 40 mA,
applying an angular step of 0.02 for 2 s with 2 h spanning from 10
to 60. Mineralogical phases were identied comparing the
obtained diffraction patterns with reference ones reported in the
ICDD PDF database, selected also on the basis of the ndings of
previous studies regarding the mineralogy of BOF and carbonate
masses.
The XRD patterns of the untreated steel slag, the residual parts
of the steel slag after ball milling and slag carbonated for 4 h are
presented in Fig. 5. The mineralogy of the steel slag was somewhat
similar to that of the cement clinker, exhibiting a predominance of
Ca phases (Ca hydroxide, Ca-containing oxides and Ca-silicate) and
other metal oxides, such as titanium oxide, vanadium oxide, mag-
netite and MnMg oxide. The calcium which exists as various
forms of compounds like Ca2SiO4 (C2S), Ca3SiO5 (C3S) and
Ca3Al2O6 (C3A) provided the potentialities for carbonation and
hydration of steel slag. After the carbonation treatment, the inten-
sity of portlandite peaks decreased signicantly while the ratio
between calcite and portlandite peaks increased, indicating the
reaction of Ca(OH)2 with CO2. Relatively, the carbonation did not
have much impact on other minerals [29]. For the residual parts
of the steel slag after milling, wherein the active minerals were
relatively less than other parts, and the contents of RO phases
 B. Pang et al. / Construction and Building Materials 84 (2015) 454467 459

Fig. 4. (a) Pore diameter vs. pore volume of untreated CSA and CSA. (b) Pore diameter vs. volume intruded of untreated CSA and CSA, (c) pore diameter vs. dV/d log D of
untreated CSA and CSA and (d) pore diameter vs. pore shape factor of untreated CSA and CSA.

Table 4
Total specic pore volume, bulk density, porosity, surface area, pore volume, permeability etc. of the untreated CSA and CSA.

Intrusion statistics Extrusion statistics Total intruded Total surface Bulk (particle) Permeability
volume (cc/g) area (m2/g) density (g/cc) (neglecting
Mean Mode Median Mean Mode Median
tortuosity effects)
(df/dD) (df/dD)
(nm)
CSA Diameter (lm) 0.11 0.004 0.83 0.39 0.0036 0.54 0.15 5.34 1.69 0.0080
Volume (cc/g) 0.14 1.43 0.07 0.15 0.55 0.10
Surface area (m2/g) 0.80 356.50 2.67 5.31 152.40 4.86
Pore number fraction 0.88 136.40 0.50 0.07 670.30 0.51
Control Diameter (lm) 0.29 0.01 1.14 1.42 0.0036 0.43 0.24 3.27 1.06 0.0655
Volume (cc/g) 0.21 0.50 0.12 0.24 0.28 0.20
Surface area (m2/g) 0.83 59.23 1.64 3.22 77.26 3.12
Pore number fraction 0.00 99.61 0.50 0.11 1063.00 0.51

(between peaks of 35 and 40) such as Calcium Manganese Oxide,
Calcium Manganese Oxide and Srebrodolskite etc. were relatively
larger. These parts of the solid contain more metal elements,
dopants and inert substances which had lower activity and higher
strength properties.
There were three main existence form of MgO in steel slag
which were combined state (calcium magnesium olivine,
manganolite etc.), solid solution (MgO, FeO, MnO and other innite
solid solution, i.e. the RO phase) and free (periclase) [30]. The origi-
nal steel slag combined other divalent metal into solid solution,
after that the MgO in steel slag was reconstructed by high tem-
perature melting and re-crystallization, so the diffraction peak
was more sharp but the width of the peak got more narrow in
XRD. Compared with original steel slag, the free MgO got higher
degree of crystallization in steel slag after the reconstruction, and
there was a little bit offset at the diffraction peak position and
the diffraction angle [3133]. It could be seen from the diffraction
patterns that the peak intensity of magnesium oxide in slag weak-
ened after carbonation reaction.
Considering the limitations of the sample with XRD due to its
small amount, the atomic uorescence spectrometry (AFS) was
used to determine the chemical composition of the CSA before
and after carbonation treat. The results showed that the content
of magnesium oxide in steel slag was about 5.25.3 wt.%, while
the difference between loss on ignition of the sample proved about
6.5 wt.% calcium oxide or magnesium oxide were carbonated dur-
ing the carbonate treatment (shown in Table 5).
3.4. Thermogravimetric analysis
Fig. 6 showed the weight loss of the samples obtained by TGA
after 4 h carbonation. This weight loss occured following two
460 B. Pang et al. / Construction and Building Materials 84 (2015) 454467

Fig. 5. XRD pattern of the untreated slag, the aggregate carbonated for 4 h and of the residual parts of the steel slag after milling.

phases: the rst one corresponds to the Ca(OH)2 or Mg(OH)2 con-
tent, which followed by the second one that carbon dioxide lost
and resulted from the decomposition of CaCO3. It was clearly that
Ca(OH)2 in steel slag was almost carbonated to CaCO3, addition-
ally, about 3% of the clinker minerals was carbonated by
calculation.
During the carbonation process of steel slag, an oversaturated
calcium carbonate solution formed rapidly at rst, which
prompted homogeneous nucleation of calcium carbonate to form
numerous amorphous calcium carbonate particles, meanwhile, cal-
cium hydroxide particles around were adsorbed because of their
high activity. This state lead to two opposite effects: The chemical
reaction rate of calcium hydroxide and carbon dioxide were
reduced while a kind of linear intermediate were forming by cal-
cium hydroxide particles. However, the amorphous calcium car-
bonate soon transformed to crystalline one due to its instability
 B. Pang et al. / Construction and Building Materials 84 (2015) 454467 461

Table 5
Chemical composition of the untreated and carbonated CSA.

CaO % SiO2 % Al2O3 % Fe2O3 % MgO % SO3 % Other % Sum% Ignition loss%
Untreated 46.2 14.94 6.05 16.15 5.22 0.74 0.09 95.9 4.1
Carbonated 46.36 12.24 5.98 18.35 5.30 0.58 0.06 89.4 10.6

[34]. As reaction proceeded, linear intermediate dissolved and dis-
appeared while combination of calcium carbonate was going on,
crystal orientation and morphology (in particular, the rod-like
and spindle) of particles were formed.
3.5. Microstructures of aggregate
Fig. 7 presented microstructures of SSA, CSA and untreated CSA.
The macroscopic morphology photos of SSA and CSA were also pre-
sented. Slag was produced at temperature between 1500 and
1700 C, therefore, a large number of dense sintered product were
observed. But there were still large amount of cracks and pores
inside the slag, from the macroscopic photo of SSA, not only
numerous pores were observed on surfaces of SSA, the composi-
tions were also quite unstable [35,36]. Although the un-carbonated
granulated steel slay aggregate (untreated) were not attached by
carbonation products, the surfaces were lled by hydration prod-
ucts such as CSH (occs) and Ca(OH)2 (hexagonal plate-like),
which due to the fact that the granulated aggregate was formed
by admixture of water and SS powder whose hydration mineral
was physically activated by grinding. Even so, its internal structure
was still relatively loose.
For the microstructural conguration of CSA which was shown
in the middle of Fig. 4, the degree of compaction of aggregate on

Fig. 6. Thermogravimetric analysis of the untreated and carbonated CSA.

462 B. Pang et al. / Construction and Building Materials 84 (2015) 454467
the basis of the untreated one had been great improved. A great
number of spindle calcium carbonate crystals which were con-
rmed by Energy Dispersive Spectrometer (EDS) were grown on
the substrate of CSA. The carbonated products were interspersed
with each other and constantly lling the pores from the surface
to the interior. Moreover, the carbonation reaction was carried
out in a high pressure, high temperature and high humidity
environment, and therefore the activity of SS powder was further
activated and the hydration rate was accelerated which further
improve the performance of CSA. Most of small CaCO3 particles
gathered into rod-like and spindle aggregation which was very
similar to the morphology formed in liquid phase like palmitic acid
and stearic acid [3739]. Moreover, the size of calcium carbonate
crystals was relatively large (even reach up to 1.5 lm), and it could
be concluded by mechanism of crystallite nucleation and the XRD
pattern of the CSA that the absorption rate of Ca(OH)2 by CaCO3
was extremely fast, the formation of crystal nucleus, especially,
was the most important factor of reaction [40].
3.6. The volume stability of aggregate
Volume stability of steel slag has drawn great concerns that
cementitious material could not undergo any appreciable volume
change after hardening [41]. However, the high alkalinity and high
temperature environment would accelerate reaction of CaO and
MgO in steel slag, then result in volume expansion. The high-tem-
perature rapid catalytic test used in this study was autoclave
expansion and immersion expansion test which conformed to
GB/T 24175-2009 in China. In order to test the stability of ne
CSA, the carbonated bricks were tested before crushing. Besides,
Fig. 7. SEM images of SSA, CSA and untreated granulated aggregate.
the immersion expansion was tested in 1 mol/L NaOH solution to
facilitate the reaction of calcium and magnesium [42]. The results
shown in Fig. 8 reected that carbonation treatments signicantly
reduced the expansion ratio of the steel slag, approximately 50% of
the SSA and 3% of CSA were broken or cracked during the experi-
ment and the rupture points were almost in the center of the
SSA, while most of spalling occured around the surface of CSA.
Moreover, the volume stability of carbonated bricks was satisfac-
tory, the chaked ratio and expansion ratio were 0% (the duplicate
test came to the same conclusions). Expansion of aggregate in solu-
tion was different with that in cement-based, therefore, the vol-
ume stability of concrete was needed to be investigated.
3.7. Workability
Hisham et al. [6] presented a study in using SSAs to enhance the
mechanical properties of concrete. They found that the reduction
in workability was attributed to the fact that slag aggregate parti-
cles were angular resulting in reduction of the mobility of concrete.
Similar results were also obtained in this experiment. Besides, it
was noteworthy that different aggregates have signicant impacts
on bleeding and segregation of fresh concrete. NAC and SSC had a
serious and slight bleeding and segregation after mixing,
respectively.
From Fig. 9, the conditions of casting surface of the concrete
samples were quite different due to bleeding and segregation, a
large number of pores and occulent hydration products presented
on the surfaces of NAC samples [22,43]. The SSA samples have a
relatively better condition, but there were still a lot of pores.
 B. Pang et al. / Construction and Building Materials 84 (2015) 454467
Fig. 8. The picture the aggregate after autoclave chake and immersion expansion test and the chaked ratio and expansion ratio of CSA and SSA.
Relatively speaking, the surfaces of CSC samples were more
smooth and almost had no pore.
From the side faces of the hardened concrete, although NAC had
a high uidity, the bubbles were not easy to exclude during the
mixing and molding, this kind of situation was caused by the angu-
lar morphology of aggregate particles. Even after vibrating for 30 s,
the water and bubbles were blocked by the recesses, edges and
corners of the irregularly shaped aggregate. Besides, there were
more pores on the side faces of SSC samples, and a large number
of striped pores were appeared [6]. The side faces of CSC were
dense and smooth relatively.
3.8. Compressive strength
Compressive strength mostly reected the overall performance
of the hardened concrete through service life of the structure, and
its trends and changes could reect changes inside the concrete
indirectly. The results shown in Fig. 10 were the mean made on
three specimens. Using CSA instead of NA could effectively
improve the compressive strength 1550% at curing time of 3, 7
and 28 days. Although the strength growth of SSC had increased
signicantly, the later strength at curing time of 60 days decreased
463
slightly. It was mainly due to the high content of f-CaO and f-MgO,
steel slag contained free (unhydrated) lime (CaO) that resulting in
volumetric instability (expansion). During the prolonged hydration
ages, the strength of SSC was more unstable than the other two
[44,45]. Hisham Qasrawi carried out an investigate that the use
of SSA as partial replacement of natural aggregate, resulted in
increase in compressive strength depending on the used replace-
ment ratio and on the w/c ratio. Increase of strength reached a
value of about 20% from the original value when the whole natural
aggregate was replaced by SSA [6,46]. Higher strength of concrete
containing steel slag aggregate could be attributed to the higher
angularity of steel slag aggregate which enhanced the interaction
between the slag aggregate and the cementitious matrix.
Moreover, strength of steel slag aggregate was higher when com-
pared to that of normal aggregate, an increase up to 17% in the
modulus of elasticity was observed when steel slag was used.
Higher replacements produced higher increase in the modulus of
elasticity, which was veried in this experiment [6,47].
From the crushing strength of the aggregate, CSA has lower
crushing strength than NA, however, the compressive strength of
concrete was not affected but improved. Besides, from the cross
sections of the samples (Fig. 9), the cracks did not go through the
464 B. Pang et al. / Construction and Building Materials 84 (2015) 454467

Fig. 9. The casting and side surface and cross section (after failure) of the concrete.

Fig. 10. Compressive strength of the concrete for curing ages of 3, 7, 28, and 60 days and strength of NA, SSA and CSA.

middle of the aggregate but traveled around. Thus, the interfacial
transition zone (ITZ) of the CSC was stronger and more rugged than
NAC. Additionally, although the water absorption of the CAS was
really high, the absorption speed was slow and therefore it did
not bring great impact on the performance of concrete, conversely,
moisture, water and gases around CSAs would be inhaled into the
aggregate like pumps in the process of the curing of the cement
slurry. It was not only the further reaction of active hydration min-
eral such as C2S, C3S and C3A etc. [48,49] inside the CSAs would
enhance the strength of the ITZ and the density of aggregate,
decrease the w/c ratio of the cement matrix surrounding and vol-
ume of air voids. From the stability of the strength of three con-
crete and aggregate, strength of SSA and SSC was high but
unstable, it was attributed to the complexity of the structure
 B. Pang et al. / Construction and Building Materials 84 (2015) 454467 465

Fig. 11. (a) The picture of the specimens for high-temperature rapid catalytic test (left:24 h, right: 120 h), (b) the expansion of each specimens, (c) the destruction of the
cement paste cubes produced by aggregate expansion and (d) the destruction conditions of the concrete produced by aggregate expansion.

Fig. 12. SEM images of the ITZ around SSA, CSA and NA.

(pores and aws) and the ingredients (solid solution and various
metal compounds). Strength of NA was the highest but that of
NAC was lowest, reasons were as follows: high clay content of
NA was the rst. Secondly, a certain amount of the at and elon-
gated aggregate which was not conducive to expel the water and
bubbles from mixture. Finally, and most importantly, though cal-
cite was the main ingredient of NA, it was not conducive to binding
to the matrix, and therefore the ITZ formed was too weak to with-
stand the stress [50].
From the morphological view of concrete after crushed (Fig. 9),
the stress sections (defective and awed) of the NAC and SSC sam-
ples were almost completely destroyed, a certain amount of aggre-
gate were broken. The shape of CSC was relatively intact. Only the
interface cracks around the aggregate was found, the strain and
stress were more evenly and the service would be also more stable.
3.9. Volume stability of concrete
Steel slag contained hydratable oxides (CaO and/or MgO) that
would result in volumetric instability (expansion) that must be
dealt with through appropriate steel slag aging and quality control,
especially appropriate testing to ensure its appropriate use in con-
struction [51]. Thus, direct use of steel slag as concrete aggregate
was prohibited in China. Lun et al. came up with suitable methods
for enhancing the volume stability of steel slag utilized as ne
aggregate [41]. Thyssen Krupp Stahl AG et al. have developed a
466 B. Pang et al. / Construction and Building Materials 84 (2015) 454467
method working on the basis of the injection of oxygen and dry
sand into the liquid slag and sufciently decreased the content of
free lime and free MgO [52]. The high-temperature rapid catalytic
test used in this study was autoclave expansion, which conformed
to ASTM C151 and heating catalysis. The specimen was placed in a
cabinet with constant temperature and humidity, and heating
catalysis was conducted at 100 C and a relative humidity of
100%. The change in length of each specimens was tested daily.
From the picture of the specimens for high-temperature rapid
catalytic test which shown in Fig. 11a, fracture occurred to a part
of the SSA mortar bar within 24 h, and all of SSA mortar were com-
pletely destroyed in 120 h. From Fig. 11b, the change in length of
the SSA mortar increased rapidly and unstably within 48 h. The
expansion of CSA mortar was within 0.02%, while a certain amount
of retraction occurred in the length of the NA mortar.
In order to prove the actual situation of the steel slag expanding
in concrete and the accuracy of the conclusion throughout the
experiment, paste cubes of 2 cm3 lled with each aggregate and
the concrete with each aggregate were made and placed under
the same circumstances to the high-temperature rapid catalytic
test. The results shown in Fig. 11c and d exhibited that direct use
SSA as aggregate in concrete caused serious damage within 48 h,
and after the extension of the test, the concrete with CSA did not
crack or cleft in the steam curing after 240 h which represents that
CSA could be used as aggregate with good volume stability, which
also prove the possibility of using large number of steel slag as
aggregate to instead of nature aggregate.
The use of CSA effectively eased the bleeding and segregation of
fresh concrete from the Table 3, and from the SEM images of the
ITZ around the SSA, CSA and NA which were shown in Fig. 12,
the pores and air-voids were obviously reduced by using CSA,
and the matrix was more compact than the other two. Whats
more, a crack about 13 lm were observed in the interfaces
between the SSA, NA and the cement matrix, while CSA was
attached to the cement matrix closely and hydration products
interspersed with each other. Thus, the CSA enhanced ITZ of aggre-
gate and cement matrix.
4. Conclusions and recommendations
Based on the study presented in this paper, the following con-
clusions are drawn:
1. The water absorption values of coarse CSA, NA and SSA were
8.91, 1.69 and 3.31, respectively. Although the absorption of
CSA is 5.27 times of NA and 2.69 times of SSA, the workability
of concrete with CSA was not signicantly inuenced.
2. After accelerated carbonation, the harmful pores in aggregate
reduced by 24.4% while the harmless pores increased 67.9%.
The reaction products were spindle and rod-like columnar cal-
cium carbonate particles formed by calcium hydroxide.
3. Using CSA instead of NA could improve the compressive
strength of the concrete by 20% in 28 days, and the strength
of CSC exceeded SSC in 60%. From the cross section of concrete
samples, the CSA enhanced ITZ of aggregate and cement matrix.
4. Relatively speaking, the mechanical property of the CSA was
lower but more stable than the other two and the service of
concrete with CSA was also better.
5. Carbonation could greatly reduce the free CaO content from
about 7% wt. to less than 1% wt. in 3 h.
6. The use of CSA in concrete mixes was useful in reducing the
environmental problems created by dumping steel slag, and
was also helped to maintain sustainability of the environment
by reducing the carbon dioxide levels in the atmosphere.
7. Direct use of SSA as aggregate in concrete would cause serious
damage within 48 h, and after the extension of the test, the con-
crete with CSA did not crack or cleft in the steam curing after
240 h which represents that CSA could be used as aggregate
with good volume stability, which also proved the possibility
of using large number of steel slag as aggregate to instead of
nature aggregate.
In short, the successful utilization of CSA as aggregate in pave-
ment construction provided a new and more cost effective
approach for aggregate resources, and decrease the threats of solid
wastes to environment. However, more studies should be carried
out on its durability in future.
Acknowledgements
This research work has been made possible thanks to nancing
from National Natural Science Foundation of China (No. 51172098)
and Shandong Province Science and Technology Development Plan
(2014GSF117017). It could not have gone ahead without the mate-
rials supplied by Zonghui Zhou.
Appendix
Abbreviations
CSA carbonated granulated steel slag aggregate
NA natural aggregate
SSA steel slag aggregate (crushed steel slag)
CSC carbonated granulated steel slag concrete, which is
concrete made with CSA
NAC natural aggregate concrete, which is concrete made
with NA
SSC steel slag concrete, which is concrete made with SSA
References
[1] Maslehuddin M, Sharif AM, Shameem M, Ibrahim M, Barry M. Comparison of
properties of steel slag and crushed limestone aggregate concretes. Constr
Build Mater 2003;17(2):10512.
[2] Yi H, Xu G, Cheng H, Wang J, Wan Y, Chen H. An overview of utilization of steel
slag. Proc Environ Sci 2012;16:791801.
[3] Mehta PK, Monteiro PJ, Education M-H. Concrete: microstructure, properties,
and materials. New York: McGraw-Hill; 2006.
[4] Anastasiou E, Filikas KG, Stefanidou M. Utilization of ne recycled aggregates
in concrete with y ash and steel slag. Constr Build Mater 2014;50:15461.
[5] Bilir T. Effects of non-ground slag and bottom ash as ne aggregate on concrete
permeability properties. Constr Build Mater 2012;26(1):7304.
[6] Qasrawi H. The use of steel slag aggregate to enhance the mechanical
properties of recycled aggregate concrete and retain the environment. Constr
Build Mater 2014;54:298304.
[7] Yu S, Wang X, Wu D. Microencapsulation of n-octadecane phase change
material with calcium carbonate shell for enhancement of thermal
conductivity and serving durability: synthesis, microstructure, and
performance evaluation. Appl Energy 2014;114:63243.
[8] Shaikh FU, Supit SW. Mechanical and durability properties of high volume y
ash (HVFA) concrete containing calcium carbonate (CaCO3) nanoparticles.
Constr Build Mater 2014;70:30921.
[9] Wang Z-B, Li C-Z, Gu D-M, Yin G-P. Carbon riveted PtRu/C catalyst from glucose
in-situ carbonization through hydrothermal method for direct methanol fuel
cell. J Power Sources 2013;238:2839.
[10] Liang J, Liu Y, Zhang J. Effect of solution pH on the carbon microsphere
synthesized by hydrothermal carbonization. Proc Environ Sci 2011;11:13227.
[11] Morandeau A, Thiry M, Dangla P. Impact of accelerated carbonation on OPC
cement paste blended with y ash. Cem Concr Res 2015;67:22636.
[12] Zhang T, Yu Q, Wei J, Zhang P, Chen P. A gap-graded particle size distribution
for blended cements: analytical approach and experimental validation.
Powder Technol 2011;214(2):25968.
 B. Pang et al. / Construction and Building Materials 84 (2015) 454467
[13] Zhang T, Yu Q, Wei J, Li J, Zhang P. Preparation of high performance blended
cements and reclamation of iron concentrate from basic oxygen furnace steel
slag. Resour Conserv Recycl 2011;56(1):4855.
[14] Zhang T, Yu Q, Wei J, Zhang P. Effects of size fraction on composition and
fundamental properties of Portland cement. Constr Build Mater
2011;25(7):303843.
[15] Liu L, Shen D, Chen H, Xu W. Aggregate shape effect on the diffusivity of
mortar: a 3D numerical investigation by random packing models of ellipsoidal
particles and of convex polyhedral particles. Comput Struct 2014;144:4051.
[16] Fuller WB, Thompson SE. The laws of proportioning concrete. Trans Am Soc
Civil Eng 1906;57(2):67143.
[17] Andreasen AM, Andersen J. Relation between grain size and interstitial space
in products of unconsolidated granules. Kolloid Z 1930;50:271328.
[18] Andreasen A. Validity of Stokes law for nonspherical particles. Kolloid-Z
1929;48:179.
[19] Kim T, Olek J. Effects of sample preparation and interpretation of
thermogravimetric curves on calcium hydroxide in hydrated pastes and
mortars. Trans Res Record: J Transp Res Board 2012;2290(1):108.
[20] Berndt M. Properties of sustainable concrete containing y ash, slag and
recycled concrete aggregate. Constr Build Mater 2009;23(7):260613.
[21] Chang C-F, Chen J-W. The experimental investigation of concrete carbonation
depth. Cem Concr Res 2006;36(9):17607.
[22] Panesar DK, Francis J. Inuence of limestone and slag on the pore structure of
cement paste based on mercury intrusion porosimetry and water vapour
sorption measurements. Constr Build Mater 2014;52:528.
[23] Leech C, Lockington D, Hooton RD. Estimation of water retention curve from
mercury intrusion porosimetry and van Genuchten model. ACI Struct J
2006;103.
[24] Chen X, Wu S, Zhou J. Inuence of porosity on compressive and tensile
strength of cement mortar. Constr Build Mater 2013;40:86974.
[25] Wu S, Chen X, Zhou J. Inuence of strain rate and water content on mechanical
behavior of dam concrete. Constr Build Mater 2012;36:44857.
[26] Wu S, Xue Y, Ye Q, Chen Y. Utilization of steel slag as aggregates for stone
mastic asphalt (SMA) mixtures. Build Environ 2007;42(7):25805.
[27] Wu ZW, Lian HZ. High performance concrete. High performance concrete
1999.
[28] Chen X, Wu S. Inuence of water-to-cement ratio and curing period on pore
structure of cement mortar. Constr Build Mater 2013;38:80412.
[29] Li Z, Zhao S, Zhao X, He T. Cementitious property modication of basic oxygen
furnace steel slag. Constr Build Mater 2013;48:5759.
[30] Zhang T-s, Liu F-t, Wang J-w, Li Y-k, Zhou Z-h, Cheng X. Recent development of
steel slag stability and activating activity. Bull Chin Ceram Soc 2007;5:028.
[31] Back M, Bauer M, Stanjek H, Peiffer S. Sequestration of CO2 after reaction with
alkaline earth metal oxides CaO and MgO. Appl Geochem 2011;26(7):
1097107.
[32] Jeon H, Kim DJ, Kim SJ, Kim JH. Synthesis of mesoporous MgO catalyst
templated by a PDMSPEO comb-like copolymer for biodiesel production. Fuel
Process Technol 2013;116:32531.
[33] Qing Y, Chen HX, Wang YQ, Wang SX, Lou ZG. Effect of MgO and gypsum
content on long-term expansion of low heat Portland slag cement with slight
expansion. Cem Concr Compos 2004;26(4):3317.
View publication stats
467
[34] Zheng Y, Gu X, Xi Y, Chai D. In vitro degradation and cytotoxicity of Mg/Ca
composites produced by powder metallurgy. Acta Biomater
2010;6(5):178391.
[35] Salman M, Cizer , Pontikes Y, Vandewalle L, Blanpain B, Van Balen K. Effect of
curing temperatures on the alkali activation of crystalline continuous casting
stainless steel slag. Constr Build Mater 2014;71:30816.
[36] Zhao L, Li Y, Zhou Y, Cang D. Preparation of novel ceramics with high CaO
content from steel slag. Mater Des 2014;64:60813.
[37] Lei F, Yan B. Morphology-controlled synthesis, physical characterization, and
photoluminescence of novel self-assembled pomponlike white light phosphor:
Eu3+-doped sodium gadolinium tungstate. J Phys Chem C
2008;113(3):107482.
[38] Jiang J, Xu D, Zhang Y, Zhu S, Gan X, Liu J. From nano-cubic particle to micro-
spindle aggregation: the control of long chain fatty acid on the morphology of
calcium carbonate. Powder Technol 2015;270:38792.
[39] Ukrainczyk M, Kontrec J, Kralj D. Precipitation of different calcite crystal
morphologies in the presence of sodium stearate. J Colloid Interface Sci
2009;329(1):8996.
[40] Wang Q, Yan P, Feng J. A discussion on improving hydration activity of steel
slag by altering its mineral compositions. J Hazard Mater 2011;186(2):10705.
[41] Lun YX, Zhou MK, Cai X, Xu F. Methods for improving volume stability of steel
slag as ne aggregate. J Wuhan Univ Technol-Mat Sci Edit 2008;23(5):73742.
[42] Zhao QL, Zhu L, Li L, Wei M, Zhou MK. Chinese steel slag for dry-mix mortar.
ZKG Int 2008;61(2):7786.
[43] Kim H, Jeon J, Lee H. Workability, and mechanical, acoustic and thermal
properties of lightweight aggregate concrete with a high volume of entrained
air. Constr Build Mater 2012;29:193200.
[44] Mattila H-P, Zevenhoven R. Production of precipitated calcium carbonate from
steel converter slag and other calcium-containing industrial wastes and
residues. In: van Eldik R, Aresta M, editors. Adv Inorg Chem 2013:34784.
[45] Kuo W-T, Shu C-Y. Application of high-temperature rapid catalytic technology
to forecast the volumetric stability behavior of containing steel slag mixtures.
Constr Build Mater 2014;50:46370.
[46] Qasrawi H, Shalabi F, Asi I. Use of low CaO unprocessed steel slag in concrete as
ne aggregate. Constr Build Mater 2009;23(2):111825.
[47] Olawuyi B, Olusola K, Babafemi A. Inuence of curing age and mix composition
on compressive strength of volcanic ash blended cement laterized concrete.
Civil Eng Dimension 2012;14(2):8491.
[48] San-Jos JT, Vegas I, Arribas I, Marcos I. The performance of steel-making slag
concretes in the hardened state. Mater Des 2014;60:6129.
[49] Wang XH, Jacobsen S, He JY, Zhang ZL, Lee SF, Lein HL. Application of
nanoindentation testing to study of the interfacial transition zone in steel ber
reinforced mortar. Cem Concr Res 2009;39(8):70115.
[50] Yusuf MO, Johari MAM, Ahmad ZA, Maslehuddin M. Shrinkage and strength of
alkaline activated ground steel slag/ultrane palm oil fuel ash pastes and
mortars. Mater Des 2014;63:7108.
[51] Wang G. Determination of the expansion force of coarse steel slag aggregate.
Constr Build Mater 2010;24(10):19616.
[52] Motz H, Geiseler J. Products of steel slags an opportunity to save natural
resources. Waste Manage 2001;21(3):28593.