Download as pdf or txt
Download as pdf or txt
You are on page 1of 12

Cement and Concrete Research 125 (2019) 105893

Contents lists available at ScienceDirect

Cement and Concrete Research


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

Carbonation of steel slag and gypsum for building materials and associated T
reaction mechanisms

Xue Wanga, Wen Nia, , Jiajie Lia, Siqi Zhanga, Michael Hitchb, Rodrigo Pascualc
a
Beijing Key Laboratory of Resource-oriented Treatment of Industrial Pollutants, University of Science and Technology Beijing, Beijing 100083, China
b
Department of Geology, Tallinn University of Technology, Ehitajate tee 5, Tallinn 19086, Estonia
c
Mechanical Engineering Department, Universidad de Concepcion, Concepcion, Chile

A R T I C LE I N FO A B S T R A C T

Keywords: The carbonation of steel slag to produce building material is a useful way to increase the utilization of steel slag
Steel slag and absorb carbon dioxide. In this study, gypsum, steel slag, and water were mixed, compaction-shaped, and
Carbonation carbonation-cured as a means of improving the strength of the steel slag. It was observed that gypsum promoted
Gypsum an increase in both the compressive strength and the CO2 uptake of steel slag. CO2 uptake was positively cor-
Monocarbonate
related with strength. Microanalysis indicated that the main hydration product were C-S-H phases and ettringite,
Construction materials
while the main carbonation products were calcite and monocarbonate (C3A. CaCO3.11H2O). Gypsum is
speculated to promote the rapid hydration of steel slag to form ettringite (C3A.3CaSO4.32H2O), which then
reacts with CO2 to produce monocarbonate; thus, gypsum plays a catalytic role in this system. The results of this
study therefore provide theoretical guidance for the preparation of steel slag–gypsum carbide building materials.

1. Introduction waste is usually the main source of CO2 emissions. In-situ carbon
fixation not only saves transportation costs, but also makes full use of
Steel slag is a solid waste product of the steelmaking process, and its the thermal energy of the exhaust gas to accelerate reactions [15,16].
output accounts for 15–20% of crude steel [1,2]. The annual steel slag According to previous studies [17–19], when steel slag is mixed
production of China is more than 100 million tons, while its utilization with water and placed within a CO2-rich environment, components in
rate is only 29.5% [3]. Steel slag contains hydraulic components, such the steel slag, such as dicalcium (C2S) and tricalcium (C3S) silicates,
as dicalcium (C2S) and tricalcium (C3S) silicates that can react with H2O react with water to form C-S-H gels. These then react with CO2 to form
to form calcium-silicate-hydrate (C-S-H). Therefore, steel slag has the CaCO3 like other minerals in steel slag, such as portlandite (Ca(OH)2)
potential to be used as cementitious material [4]. However, steel slag and free oxide (f-CaO and f-MgO). Few studies have reported on the
has limited recyclability, primarily because (1) its hydration is slow and joint carbonation mechanisms of gypsum and steel slag.
(2) it contains a large amount of free lime (f-CaO) and magnesium oxide In this study, the compressive strength and CO2 uptake of blocks of
(f-MgO), which bring about deleterious expansion and volume in- steel slag and gypsum are investigated. The effects of gypsum on the
stability [5–9]. Carbonation is an excellent way to solve these problems, carbonation of steel slag for use as a construction material are studied
because the carbonation reaction is much faster than the hydration and microstructural analysis is used to explore the hydration and car-
reaction in a humid environment, and carbonation of free oxide will bonation mechanisms involved.
improve volume stability [8,9].
Global warming has attracted worldwide attention and CO2 emis- 2. Materials and methods
sion-reduction has become a topic of increasing concern. The cement
industry [10,11] and steel industry [12,13] are both responsible for 2.1. Materials
global warming due to their high CO2 emissions. Absorbing CO2 in the
exhaust gas from the steel industry by steel industry waste, such as steel The steel slag used in this test was provided by Anshan Iron and
slag to produce qualified building materials, is an effective way to Steel Co., Ltd., China and was ground to a specific surface area of 460
promote the utilization of solid waste, reduce carbon emissions, and m2/kg. The chemical compositions of the steel slag and gypsum were
reduce the use of cement clinker [8,14,15]. Moreover, industrial solid determined via X-ray Fluorescence and carbon/sulfur analyses. Table 1


Corresponding author.
E-mail address: niwen@ces.ustb.edu.cn (W. Ni).

https://doi.org/10.1016/j.cemconres.2019.105893
Received 13 August 2018; Received in revised form 20 July 2019; Accepted 2 September 2019
0008-8846/ © 2019 Published by Elsevier Ltd.
X. Wang, et al. Cement and Concrete Research 125 (2019) 105893

Table 1 to-solid ratio (w/s) of 0.20 was maintained, and gypsum/steel slag (G/
Chemical compositions of the steel slag and gypsum studied./Wt%. SS) ratios ranging from 1:16 to 1:2 were adopted. The proportions and
Chemical compositions Steel slag Gypsum Components Steel slag curing conditions of samples are shown in Table 2. The N2 group was
cured in a standard curing chamber and acted as the control group.
CaO 44.78 48.06 Larnite 23.88
Al2O3 5.70 1.18 Hatrurite 14.04
SiO2 15.28 2.95 Mayenite 7.85 2.3. Curing
Fe2O3 22.43 0.58 Portlandite 5.73
SO3 0.30 43.57 Calcite 5.90
MgO 7.04 1.48 Quartz 3.97
Immediately after de-molding, specimens C1–C7, which required
MnO 1.93 1.18 Fayalite, magnesian 5.69 carbonation, were placed inside a carbonation chamber in which the
P2O5 1.00 0.03 RO phase 19.55 temperature was 20 ± 3 °C, the relative humidity was kept at
TiO2 0.86 0.39 Srebrodolskite 13.40 70 ± 2%, and the CO2 concentration was 20 ± 3 vol%. Non-CO2
V2O5 0.20 –
curing for N2 samples was achieved by placing the cylindrical compacts
Cr2O3 0.17 –
Others 0.28 0.58 in a standard curing box within which the temperature was 20 ± 3 °C,
Total-C 0.865 – relative humidity was kept at 70 ± 2%, and the CO2 concentration was
0.038 vol% (the atmospheric value). For carbonated groups, tests with
different durations (1, 3, 14, and 28 days) were carried out.
and Fig. 1 show the results for the chemical and mineral compositions
of the steel slag powder, respectively. The crystal structures of the
samples were characterized using a Bruker D8 ADVANCE X-ray dif- 2.4. Test methods
fractometer using CuKα radiation (40 kV, 40 mA). Data were collected
from 5° to 70° (2θ) with a step size of 0.02° in the step-scanning mode 2.4.1. Carbonation depth, strength gain, and CO2 uptake
(FT1.5 s). The QXRD analysis was conducted by Rietveld refinement The carbonation depth of the block was examined by spraying the
using TOPAS software (ver. 5.0, Bruker AXS GmbH). The main crys- samples with a phenolphthalein indicator, and then the compressive
talline phases in the steel slag were portlandite (Ca(OH)2), Larnite strength of the block was tested. The compressive strength at each
(Ca2SiO5), Hatrurite (Ca3SiO5), Srebrodolskite (Ca2Fe2O5), Mayenite curing age was determined by calculating the mean value of three tests.
(Ca12Al14O33), and RO phase (solid solution of CaO, MgO, FeO, and The CO2 contents of samples before and after carbonation were
MnO) [20,21] are indicatively main steel slag components. The gypsum measured using a carbon/sulfur combustion analyzer (EMIA-820 V,
used in this study was desulfurization gypsum, which is a by-product of Horiba). If the mass percentage of CO2 absorbed by the sample (the CO2
the steelmaking process. The composition of gypsum is broadly Ca- uptake of the sample) is x%, then the following equation can be ob-
SO4.2H2O. tained according to the principle of carbon conservation:

mbc × CO2uptake + mbc × CO2initial = mdc × CO2final (1)


2.2. Sample preparation
In the equation
The steel slag powder, gypsum, and water were mechanically mixed mbc – weight of the compact before carbonation; 8.00 g here;
for 2 min using a paste mixer. The moistened powder was then weighed mdc – weight of the dried carbonated compact; g.
(ensuring a weight of 8.00 g) and compacted into individual 20-mm CO2initial – CO2 content of the compact before carbonation obtained
diameter cylindrical specimens using a stainless steel mold at a uniaxial using ωss × CO2ss.
load of 9 MPa for 1 min. The compacts were then demoulded and placed ωss – steel slag content in the initial sample, wt%, see Table 2.
within a carbonation chamber. Fig. 2 shows the mold and compacts. To CO2final – CO2 content of the dried carbonated compact;
study the effects of gypsum on the properties of the compacts, a water- This can be rearranged to obtain the equation

3500 5 : Hatrurite
3
2: Larnite
3000 3: Portlandite
4: RO phase
2500 5: Mayenite
6: Srebrodolskite
Intensity counts

7: Calcite
2000

15
1500 21
2 4
4
7 14 1
1000 3 4 2
1 7
2 5 4
6 42 4 53
5 1 7 4
500 4
16 2 2
21 23 2 12 2
3 4

0
10 20 30 40 50 60 70
2 ()
Fig. 1. X-ray diffraction (XRD) pattern for steel slag.

2
X. Wang, et al. Cement and Concrete Research 125 (2019) 105893

Fig. 2. Stainless steel mold (left), and cylindrical compacts after de-molding (right).

Table 2 determined by using a NEXUS670 FTIR infrared spectrometer. The


Proportions and curing conditions of materials in the samples prepared. wave number ranged from 400 cm−1 to 4000 cm−1, and the resolution
Curing condition G/SS w/s ωss/Wt%
was 3 cm−1. The instrument used for X-ray photoelectron spectroscopic
analysis was a Thermo Scientific Escalab 250Xi. MIP analysis (Autopore
C1 Carbonation curing 0 0.2 83.3 IV 9500, Micromeritics) was conducted to determine the porosity and
C2 Carbonation curing 1:16 0.2 78.4 pore size distributions of the specimens.
C3 Carbonation curing 1:9 0.2 75.0
C4 Carbonation curing 1:6 0.2 71.4
C5 Carbonation curing 1:4 0.2 66.7 3. Results
C6 Carbonation curing 1:3 0.2 62.5
C7 Carbonation curing 1:2 0.2 55.6
3.1. Strength and CO2 uptake
N2 Non‑carbonation curing 1:16 0.2 78.4

ωss: SS content in the initial sample. 3.1.1. Effect of gypsum on compressive strength and CO2 uptake
In order to explore the influence of gypsum on the mechanical
mdc × CO2final − mbc × CO2initial properties and carbon sequestration capacity (CO2 uptake) of the car-
CO2uptake =
mbc (2) bonated steel slag blocks, the strength and total carbon of carbonated
blocks with different gypsum contents (C1–C7) were tested.
which was used to calculate the CO2 uptake. Fig. 3 shows the development of strength with continued curing. It
In order to study the effect of gypsum on the carbon sequestration can be seen that the compressive strengths of all of the test blocks were
capacity of steel slag, the mass percentage of CO2 absorbed by the steel over 15 MPa. The strength of the test block was highest when the mass
slag (i.e., the CO2 uptake of the steel slag) was calculated thus: ratio of gypsum:steel slag was 1:16; it reached 32 MPa after one day of
CO2uptake curing, which is 15.10% higher than the strength of blocks without
CO2 uptake of steel slag (%) = gypsum. This proves that the addition of an appropriate amount of
ωss (3)
gypsum helps increase strength. However, where a larger proportion of
gypsum was used, the strength of the test blocks was lower might
2.4.2. Microstructural analysis
60
The block samples were dried at 50 °C to remove free water. Several G/SS=0
techniques were then used to systematically investigate the micro- 55 G/SS=1/16
structure of the cured samples, namely X-ray diffraction (XRD), scan- G/SS=1/9
Compressive strength /MPa

ning electron microscopy-energy dispersive spectrometry (SEM-EDS), 50 G/SS=1/6


thermogravimetry-differential scanning calorimetry (TG-DTG), Fourier- G/SS=1/4
45 G/SS=1/3
transform infrared spectroscopy (FTIR), mercury intrusion porosimetry G/SS=1/2
(MIP), and X-ray photoelectron spectroscopy (XPS). 40
The crystal structures of the samples were characterized using a
Bruker D8 ADVANCE X-ray diffractometer using CuKα radiation 35
(40 kV, 40 mA). Data were collected from 5° to 90° (2θ) with a step size
of 0.02° in the step-scanning mode (FT1.5 s). The lattice parameters and 30
atomic positions and occupations were calculated by Rietveld refine- 25
ment using TOPAS software (ver. 5.0, Bruker AXS GmbH). Please note
that we show the QXRD results only for the crystalline part of the 20
sample. SEM backscattered electron (SEM-BSE) analysis of samples was
carried out with a SUPRA 55 Scanning Electron Microscope. The 15
0 3 6 9 12 15 18 21 24 27 30
thermal stability of the products was evaluated by TG-DSC synchronous
thermal analysis, which was carried out with a NETZSCH STA 449F3 at Curing duration /d
temperatures ranging from 50 °C to 1000 °C and at a heating rate of
10 °C/min. The chemical structure of the hydration products was Fig. 3. Compressive strength of carbonated blocks of different ages.

3
X. Wang, et al. Cement and Concrete Research 125 (2019) 105893

14 55
1 d Carb
3 d Carb 50
12
14 d Carb
28 d Carb 45
CO2 uptake of blocks /%

Compressive strength /MPa


10
40

8
35

6 30

25
4 y=4.84045x 18.0507
2
20 R =0.810
2
15

0
10
0 1/16 1/9 1/6 1/4 1/3 1/2
7 8 9 10 11 12 13 14
Mass ratio G/SS
CO2 uptake /%
Fig. 4. CO2 uptake of carbonated blocks with different amounts of gypsum.
Fig. 6. Relationship between CO2 uptake and strength.

because steel slag is the main reactant of hydration and carbonation


reactions, and more gypsum means less steel slag. increased almost linearly with an increase in the amount of gypsum.
From Fig. 3, it can be seen that the strength of the test blocks in- This illustrates that the addition of gypsum promoted the carbonation
creased during the first 14 days of carbonation curing. However, the of steel slag to some extent. From Fig. 4, it can be observed that the CO2
strengths of all of the test blocks except those without gypsum added uptake of a block at one day reached 75.60–82.92% of that at 28 days,
decreased between 14 days and 28 days of curing, which illustrates that which means that the carbonation reaction within the test block was
excessive carbonation curing is not good for strength development. nearly complete after one day of carbonation.
Therefore, short-term curing is necessary to guarantee the strength The relationship between the CO2 uptake and the strength of the test
performance of blocks in this system, and the strength requirement can blocks was assessed via a scatter plot of the CO2 uptake versus strength
be achieved through one-day curing or less. at one day, three days,14 days, and 28 days under different gypsum
In industrial production, the ability to sequester CO2 (CO2 uptake) is contents (Fig. 6). The scattered points are distributed around the
another critical index for block evaluation. In this test, the carbon se- function y = 4.84045× − 18.0507with R2 of 0.810, which indicates
questration capacity of a test block is characterized by two parameters: that CO2 uptake is positively correlated with compressive strength. This
the CO2 uptake of the block, and the CO2 uptake of the steel slag within can be attributed to carbonation playing a key role in promoting the
it. The former provides guidance for actual production, while the latter early strength of the test block because the hydration reaction of the
plays a major role in research into the uptake mechanism. Figs. 4 and 5 steel slag is slow and rarely contributes to early strength. This conclu-
show the CO2 uptake of the blocks and the CO2 uptake of the steel slag, sion is supported by the data in Table 3.
respectively, with different amounts of gypsum. It can be seen from
Fig. 4 that with an increase in the amount of gypsum, the CO2 uptake of 3.1.2. Effect of carbonation on the properties of blocks
the block decreased. This is because of the reduction in the amount of The results for the strength and CO2 uptake of the C2 and N2 blocks
steel slag, which contains the main carbonation reactants, such as C2S, are shown in Table 3. It is obvious that the carbonated blocks were
C3S, CH and C12A7, and thus acts as the main carbon-catcher in the test much stronger than the non‑carbonated blocks. The compressive
block. However, Fig. 5 shows that the CO2 uptake of the steel slag strengths of carbonated blocks at one and three days were 8.56 and
4.84 times that of the non‑carbonated blocks, respectively. Thus, car-
13.0
bonation significantly enhanced the early strength of the test blocks.

12.8 3.1.3. Carbonation depth of the samples


The depth to which carbonation penetrated into C2 compacts was
examined by spraying them with a phenolphthalein indicator after they
1d CO2 uptake of steel slag /%

12.6
had been carbonated for 0 h, 0.5 h, 1 h, 2 h, and 3 h and measuring the
depth to which the indicator color extended. It can be seen in Fig. 7 that
12.4
specimens cured under non‑carbonation conditions showed an obvious
y=2.01422x+11.77314 pink color across the whole section, whereas exposure to CO2 caused
2
12.2 R =0.912 the pink area to reduce in extent, and it gradually disappeared with an
increase in exposure time to 3 h. This indicates gradual CO2 penetration
12.0 throughout the entire section of the specimen. The measured

11.8 Table 3
Results for the properties of carbonated and non‑carbonated blocks.

11.6 Compressive strength/MPa CO2 uptake of block/%


0.0 0.1 0.2 0.3 0.4 0.5
1d 3d 1d
Mass ratio G/SS
N2 3.71 7.36 3.56
Fig. 5. One-day CO2 uptake of steel slag with different amounts of gypsum,
C2 31.78 35.61 11.30
with trendline.

4
X. Wang, et al. Cement and Concrete Research 125 (2019) 105893

100

Gypsum
Ettringite
Srebrodolskite
RO phases
Quartz
Calcite
Portlandite
Mayenite

Wt/%
50
Hatrurite
Monocarbonate

C3-14d
C3-1d
C2-28d
C2-14d
C1-28d
C1-14d
Larnite

C2-6h

C2-1d
C1-1d
N2-1d
SS
0
Specimens

Fig. 9. Quantitative mineral phase analysis of various carbonated samples


compared with an air-cured sample.

(C3S), portlandite (Ca(OH)2), and mayenite (C12A7) decreased during


both hydration and carbonation curing. The main hydration products in
Fig. 7. Optical images of C2 compact specimens sprayed with a phenolphtha- sample N2-1d are some amorphous C-S-H gels that cannot be detected
lein indicator after different curing durations. though XRD, and a small amount of ettringite (C3A·3CaSO4·32H2O), all
of which are prone to take part in the carbonation reaction and mainly
carbonation depths of samples exposed to CO2 for 0 h, 0.5 h, 1 h, 2 h, produce carbonation products, such as calcite (CaCO3) and mono-
and 3h were 0 mm, 3.24 ± 0.18 mm, 5.95 ± 0.90 mm, carbonate (C3A. CaCO3.11H2O), the amounts of which increased gra-
9.24 ± 0.15 mm, and more than 10 mm, respectively. Once a particle dually as the curing time increased. The N2 sample contained more
had reacted with CO2 and CaCO3 had formed on its surface, the pH calcium silicates and portlandite than did the C2 sample cured for one
value dropped below 10, and the phenolphthalein solution therefore day, possibly indicating that a CO2-rich environment promotes the re-
did not turn red. The pH tests revealed that the carbonation depth of all activity of these substances.
of the samples reached the sample core after 3 h of carbonation curing.
However, from CO2 uptake results in Fig. 4, CO2 uptake gradually in- 3.2.2. Microstructure as characterized with SEM and BSE
creased after 3 h and even up to 28 days due to the gradual penetration Fig. 10 shows the SEM and EDS results for non‑carbonated (N2-1d)
of the carbonation reaction into the cores of the steel slag particles. and carbonated (C2-1d and C2-28d) samples. It can be seen from
Fig. 10(a) that the non‑carbonated test block contained a large number
3.2. Microstructural analysis results of C-(A-)S-H phases with a fibrillar morphology. Some researchers
[22–27] found that some C-S-H products, have a fibrillar morphology,
3.2.1. XRD analysis which is consistent with the results of this study. Calcite and mono-
Fig. 8 shows the XRD patterns of C1, C2 and N2 specimens at one carbonate can be clearly identified in Fig. 10(b). The green plot in
day, and Fig. 9 shows the outcome of quantitative X-ray diffraction Fig. 10(b) is very likely to be monocarbonate on the basis of the EDS
(QXRD) analysis using the Rietveld refinement method. Fig. 9 indicates results. Fig. 10(c) shows a large amount of calcite and a small amount of
that the initial amounts of different minerals in the steel slag, such as C-S-H phases, demonstrating the evolution of the carbonation process.
the RO phase, srebrodolskite (C2F), quartz (SiO2), and fayalite ((Fe, Fig. 11 shows BSE images of a test block after carbonation curing.
Mg)2SiO4), stayed almost the same after either the hydration or car- An uncarbonated calcium silicate core, a Ca-leached and Si-rich zone
bonation process, while the percentages of larnite (C2S), hatrurite surrounding it, and calcium carbonate layers can be seen in Fig. 11(a).
It can also be seen that the Fe-containing substances in this system, such
7500 as the RO phase and brownmillerite, did not enter into the carbonation
1 1: Calcite : Monocarbonate
7000 reaction.
: Gypsum 4: Quartz
6500
6000 5: Errringite : Srebrodolskite
: RO phase : Mayenite
3.2.3. FTIR analysis
5500
The results of FTIR analysis of test blocks C1, C2, and N2 at the age
5000 : Portlandite : Larnite, Hatrurite
of one day and C2 at 14 days and 28 days are shown in Fig. 12. The
4500
absorption band from 3405 cm−1 to 3548 cm−1 and the absorption
Intensity

4000
C1-1d band at 1621 cm−1 are the stretching and bending vibration bands of
3500 2
3 OeH in the water molecules of the hydration products, respectively.
3000
8 The appearance of the stretching band is attributed to the formation of
2500 76 6
4 8 4 2 66 7 1 7 hydrogen bonding with a wide range of strengths [28]. The intensity of
2000 6 4 3 13 4 1 1 11
6 3 63 16 16 the OeH bond absorption bands is higher in C2 than in C1, which in-
C2-1d
1500 dicates that more hydration products were formed in the system after
1000 5 5 the addition of gypsum. Therefore, gypsum promotes the hydration of
N2-1d
500 steel slag to some extent. The “shoulder” that appears at 3548 cm−1
10 20 30 40 50 60 70 may be the vibrational band of water molecules in the gypsum. The
2 depths of the OeH bond absorption bands are greater in C2 than in N2,
proving that CO2 promotes the formation of hydration products. In
Fig. 8. XRD patterns for C1, C2, and N2 samples at one day. summary, both gypsum and CO2 can promote hydration. As seen in

5
X. Wang, et al. Cement and Concrete Research 125 (2019) 105893

(a) Paste N2, 1 d of air curing

(b) Paste C2, 1 d of CO2 curing

(c) Paste C2, 28 d of CO2 curing


Fig. 10. SEM images of non‑carbonated and carbonated samples. C: Calcite, Mc: Monocarbonate.

Fig. 12(b), there was no significant change in the peaks characterizing addition of gypsum in C2 stimulated the rapid carbonation of the
the OeH bond in C2 over the course of 28 days, indicating a lack of phases in the steel slag. It is also apparent from a comparison of the
significant growth in the hydration product during this time. results for C2 and N2 that a carbon-rich environment greatly promoted
The absorption bands at the bands at 1420–1480 cm−1 could be the formation of carbonized products. Furthermore, Fig. 12(b) shows
attributed to the symmetric stretching vibration of C–O; the peaks at that the amount of carbonation product formed gradually increased
873 and 712 cm−1 are corresponding to the bending vibration of with curing time.
CO32−. C1 contained more steel slag than did C2 but produced a An absorption band ranging from 900 cm−1 to 1200 cm−1 is typical
smaller amount of the carbonation product, which indicates that the of the asymmetric stretching vibration of SieO, and absorption peaks at

6
X. Wang, et al. Cement and Concrete Research 125 (2019) 105893

(a) Block C2, 1 d of CO2 curing (b) Block C2, 28 d of CO2 curing

(c) Closeup of Zone A in (b)


Fig. 11. Typical BSE images of samples subjected to CO2 curing. CS: uncarbonated calcium silicate core, B: Brownmillerite, C: Calcite, R: RO phase, Mc:
Monocarbonate.

993 cm−1 and 516 cm−1 are characteristic of asymmetric stretching carbonized samples (C1, C2); this can be attributed to the formation of
vibration and bending vibration of Si–O–Si in C-S-H gel, respectively. silica-like products and an increase in the degree of polymerization.
The broad bands around 516 cm−1were due to the out-of-plane skeletal [28]
vibration of SieO bond. Fig. 9 (a) shows that these characteristic peaks The characteristic absorption peak of SO42− is between 1112 cm−1
only appeared in the uncarbonized test block (N2). The FTIR spectrum and 1141 cm−1. The characteristic absorption peak of the AlO6 octa-
also shows that an absorption band at 1041 cm−1 only appeared in the hedron in the monocarbonate, ettringite is at 669 cm−1. It can be seen

C2-28d

C1-1d
C2-14d
C2-1d

C2-1d
N2-1d
1621

712
669
712
669

3548

574
1041

3405

873
1041
1141
1141

458
1112
3548

1112
3405

993

1481
1420
1481

574
1420

458
873

516

4000 3500 3000 1500 1000 500


4000 3500 3000 1500 1000 500
-1 -1
Wavenumbers(cm ) Wavenumbers(cm )

(a) (b)
Fig. 12. Results of FTIR analysis of C1, C2, and N2 test blocks after one day of curing (a) and C2 samples with different curing ages (b).

7
X. Wang, et al. Cement and Concrete Research 125 (2019) 105893

(a) 100 0.06

90
0.03
80
0.00
70
-0.03
60
Gypsum

DTG(%/ C)
C-S-H phases
TG (%)

Mc
50 C-S-H Ca(OH)2 -0.06
AFt
40
Gypsum Carbonate group -0.09
30
-0.12
20
C1-1d -0.15
10 C2-1d
0
N2-1d -0.18
0 100 200 300 400 500 600 700 800 900 1000
Temp.( C)
(b) 100 0.06

90 0.03

80
0.00
70
-0.03
60 C-S-H phases Gypsum
C-S-H Mc -0.06
Ca(OH)2
TG (%)

DTG(%/ C)
50 AFt
-0.09
40 Gypsum
-0.12
30
-0.15
20
C2-1d
10 -0.18
C2-14d
C2-28d Carbonate group
0 -0.21
0 100 200 300 400 500 600 700 800 900 1000

Temp.( C)
Fig. 13. Results of TG analysis of a) C1, C2 and N2 samples cured for one day and b) C2 samples of different curing ages.

that this peak was detected only in the C2 sample, which had added Table 4
gypsum and was CO2-cured. CO2 uptake and bound water, as determined by TG and total-C analysis.
CO2 content/% CO2 uptake of CO2 uptake of Bound
3.2.4. TG-DTG analysis block/% steel slag/% water/%

Fig. 13a shows the results of TG-DTG analysis of the C1, C2, and N2 Test TG Total-C TG Total-C TG Total-C
test blocks after one-day curing, and Fig. 13b shows that of C2 samples methods
after one-day, 14-day, and 28-day curing. The mass loss below 500 °C
was mainly due to the loss of interlayer water and hydroxyl groups in N2-1d 5.18 5.48 2.28 2.57 2.91 3.27 5.88
C1-1d 13.64 12.94 10.50 9.83 12.60 11.80 6.99
the structure of hydration products. Specifically, the evaporation of
C2-1d 13.11 12.39 10.08 9.39 12.85 11.97 7.74
interlayer water occurred below 240 °C and the loss of hydroxyl groups C2-14d 13.75 14.55 10.97 11.76 13.99 14.99 6.47
in the structure occurred between 240 °C and 500 °C. The mass loss that C2-28d 14.30 15.19 11.54 12.42 14.72 15.84 6.19
occurred between 500 °C and 800 °C is mainly associated with the
decarbonation of calcium carbonate and/or monocarbonate [29], while
that above 800 °C was mainly due to the loss of SO3 in gypsum. Table 4 curing were 13.64%, 13.11%, and 5.18%, respectively, whereas the
shows some CO2 uptake and bound water amount data obtained from carbon contents obtained by total carbon analysis were 12.94%,
both TG and Total-C analysis, measured by TG-DTG curves and carbon/ 12.39%, and 5.48%, respectively, giving an error value of less than 6%.
sulfur analyses, respectively. Both TG and total-C analysis are shown to indicate that the relative
Table 4 shows that the carbon contents in C1, C2, and N2 after 1-d magnitudes of CO2 uptake of the steel slag in the three samples are

8
X. Wang, et al. Cement and Concrete Research 125 (2019) 105893

102.3 347.2

350.8

Si-C1 347.6
102.7 Ca-C1
Ca-C2
351.3 347.7
Si-C2
102.2 Ca-N2
Si-N2

115 110 105 100 95 354 352 350 348 346 344 342 340

Binding Energy (eV) Binding Energy (eV)

(a) (b)
Fig. 14. XPS analysis of C1, C2, and N2 test blocks at the age of one day (a) Si 2p; (b) Ca 2p.

C2 > C1 > N2, indicating that both gypsum and CO2 promoted car- The change trend in Ca 2p is similar to that of Si 2p. and the Ca 2p
bonation of steel slag. Furthermore, the table shows that the relative binding energy of hydration products is larger than that of unhydrated
magnitudes of bound water content are C2 (7.74%) > C1 products. According to Black [29], the central peaks of Ca 2p binding
(6.99%) > N2 (5.88%), demonstrating that both gypsum and CO2 energy for dicalcium silicate and tricalcium silicate are 346.87 eV and
promoted hydration, with CO2 playing the more significant role (in- 346.55 eV, respectively. It can be seen from Fig. 14(b) that both the
creasing hydration by 31.63%). hydrated and carbonated samples surpassed 347 eV. However, a com-
Fig. 13b and Table 4 show the CO2 content of the test block in- parison of the C2 and N2 groups shows that they have similar Ca 2p
creasing as the carbonation curing time increases, but staying nearly binding energy results. This could be attributed to the Ca 2p binding
the same from 14 days to 28 days. This result is in line with the CO2 energies of calcium carbonate and C-S-H gel being similar [32].
uptake result. It can also be seen from Table 4 that the bound water C1 has a lower Ca 2p binding energy than do C2 and N2, indicating
content decreased with time due to the carbonation reaction of the C-S- that the addition of gypsum leads to an increase in Ca 2p binding en-
H gels or the instability of monocarbonate. ergy. This may be attributed to the fact that the Ca 2p binding energy of
gypsum is 347.70 eV [36] and that of Ca/Al-LDH is 347.5 eV [37], both
of which are higher than that of C2S and C3S. It can be clearly seen that
3.2.5. XPS analysis the breadth of the Ca 2p binding energy peak is larger for C2 than for
XPS analysis was performed on C1, C2, and N2 test blocks at the age N2 and C1. This may be because of a difference in the diversity of
of one day. The results are shown in Fig. 14. products: calcium carbonate, C-S-H gel, gypsum, and monocarbonate
The main forms of silicon in steel slag are dicalcium silicate and all coexisted in the C2 blocks, which will have contributed to a wider
tricalcium silicate. According to Black [30], the centers of the Si 2p range in Ca 2p binding energy.
binding energy peaks for dicalcium silicate and tricalcium silicate are at
100.80 eV and 100.57 eV, respectively. The SieN2 curve in Fig. 14(a)
indicates that the Si 2p binding energy had increased after one day of 3.2.6. MIP analysis
hydration curing. This may be because dicalcium silicate and tricalcium Pore size and pore size distribution measurements were performed
silicate are nesosilicate minerals, which tend to have low Si 2p binding for carbonated (C2-1d) and non‑carbonated pastes (N2-1d). The MIP
energy, while the Si 2p binding energy of C-S-H gel is higher due to the results for all mixtures and the main pore structure characteristics are
loss of non-bridging oxygen atoms during hydration, i.e., the loss of summarized in Table 5, while the cumulative intruded pore volume
Si–O–Ca moieties upon silicate polymerization. C-S-H decalcification curves and corresponding pore size distributions obtained for carbo-
upon carbonation leads to silicate polymerization, with the ultimate nated and non‑carbonated samples are presented in Fig. 15. The results
product being a Q4 silicate and/or silanol species [31] and calcium reveal that the one-day carbonated sample had not only a lower pore
carbonate [24]. Many researchers [32–34] have related Si 2p binding volume, but smaller pores.
energy to the extent of silicate tetrahedral polymerization. Black [35] Fig. 15(b) shows that the carbonized sample contained more small
also found that Si 2p binding energies show a strong negative correla- pores (those with a diameter of 30 nm or less) than the non‑carbonated
tion with C/S in C-S-H phases. Carbonation and the subsequent dec- sample. This conclusion can also be drawn from the greater steepness of
alcification of C-S-H lead to a decrease in C/S and thus an increase in the cumulative volumetric curve for the carbonized sample than that of
the Si 2p binding energy. the uncarbonized sample within the size range of less than 30 nm
The C2 test block had a higher Si 2p binding energy than the C1 (Fig. 12(a)). It is investigated that the interlayer distances of C-S-H gels
block, indicating that the addition of gypsum also leads to an increase
in the Si 2p binding energy. This may be because the gypsum consumes Table 5
Ca and Al in the C-S-H gel in the system, resulting in a C-S-H gel with Results of MIP analysis.
low C/S, thereby increasing the Si 2p binding energy in the gel. Sample N2-1d C2-1d
In addition to the shift to higher binding energies upon aging, the Si
2p binding energy peaks of carbonated specimens (C1 and C2) are Total intrusion volume 0.1730 mL/g 0.1289 mL/g
Total pore area 9.457 m2/g 16.270 m2/g
slightly broader than that of a non‑carbonated specimen (N2). Peak
Average pore diameter (4 V/A) 73.2 nm 31.7 nm
broadening may indicate a more disordered silicate structure compared Porosity 30.4976% 25.2706%
with non‑carbonated phases.

9
X. Wang, et al. Cement and Concrete Research 125 (2019) 105893

100
0.10

dV/dlogD Pore Volume (mL/g)


80

% of Total Intrusion Volume


0.08
Non-carbonated
60 0.06
Non-carbonated
0.04
40
Carbonated
0.02
20
Carbonated
0.00
0
1 10 100 1000 10000 100000 1 10 100 1000 10000 100000
Pore Diameter (nm) Pore Diameter (nm)

(a) (b)
Fig. 15. Pore size distribution of carbonized (C2) and uncarbonized (N2) test blocks after one-day curing.

is close to 1 nm [38–40] and could not be determined by MIP results. space is provided for precipitation of C-S-H. In this situation, fibrillar C-
While C–S–H gel pore size is between 2.5 and 10 nm. The main capillary S-H is prone to form [25].
pore is thought to be in the range of 10–100 nm. From Fig. 15 (b), the In the CO2-cured block without gypsum, hydration and carbonation
C–S–H gel pore volume in the sample increases after it is subjected to proceeded at the same time. The hydration products portlandite and C-
CO2. It may be because that C-S-H gels are expended during carbona- S-H with high Ca/Si (C/S) are easily carbonized to form calcite and low
tion and the structure of C-S-H gels is destroyed to some extent, thus C/S C-S-H (Fig. 16). In a CO2-rich environment, the reduction of C-S-H
increasing C-S-H pore volume. In this particular study, structure is gel promoted the progress of the slag hydration reaction to some extent
different from that of traditional cement mainly because of low re- (Fig. 9). Fe-containing mineral phases, such as brownmillerite and the
activity of the slag. These factors would possibly leads to relative loose RO phase in steel slag, had lower carbonation reactivity than calcium
structure and bigger pore in the system as seen in Fig. 15. It is also silicate and underwent little carbonation according to the XRD (Fig. 8
obvious that the volume in the pore size range 30 nm to 7 μm domi- and Fig. 9) and BSE (Fig. 11(a)) results. This is possibly because they
nated the pore volume in either C2 or N2 sample. The volume per- have low reactivity in the presence of CO2 at room temperature and
centage of this size range was larger in the uncarbonized sample than in pressure.
the carbonized sample. This can be attributed to carbonized products A sample that is doped with gypsum and CO2-cured has a reduced
precipitating in the pores of the matrix [41,42] and small precipitated CO2 uptake compared with one that is not (Fig. 4) due to a decrease in
particles of varying sizes joining together, reducing the overall pore size the amount of steel slag, which contains the main carbonation reactant.
in the test block. Besides, the irreversible drying shrinkage that is However, the strength of the former is higher (Fig. 3). This is because
caused by the water-loss happened along with the carbonation [43]. the addition of desulfurization gypsum promoted carbonation reactivity
The refining of the pores and decrease of the porosity caused by car- in mayenite (C12A7) to form monocarbonate (XRD results in Fig. 8 and
bonation contributed to the increase of the compressive strengths of the Fig. 9, SEM-BSE results in Fig. 10 and Fig. 11, FTIR results in Fig. 12),
specimens. which acted as a skeleton and improved the strength of the material to
Table 5 shows that the porosity reduced from 30.50% before car- some extent. Therefore, the strength and CO2 uptake of gypsum-doped
bonation to 25.27% after carbonation, and the average pore diameter steel slag were higher than those of a test block without gypsum (Fig. 3
also reduced by approximately 50%. The above findings fully prove that and Fig. 4). The XRD patterns for non‑carbonated (N2) and carbonated
the carbonation process increases the density of the system, thus pro- (C2) blocks in Fig. 9 clearly show that C2 had a significantly higher
moting an improvement in strength. amount of monocarbonate, which indicates that the carbonation reac-
tion promoted the formation of monocarbonate. It is very likely that the
addition of gypsum may promote the rapid dissolution of mayenite
4. Discussion (C12A7) in steel slag to form ettringite (C3A.3CaSO4.32H2O), which then
reacts with CO2 to produce monocarbonate and/or calcite (Fig. 16).
4.1. Phase and microstructural evolution Therefore, in this system, gypsum promotes not only hydration but
also carbonation. The products of hydration reactions ettringite are the
The hydration and carbonation mechanisms in this system is shown reactants of the carbonation reaction, so the progress of the carbonation
in Fig. 16. The sample without carbonation curing contained near- reaction can also promote hydration (Fig. 13 and Table 4). From the
amorphous C-S-H [44] and some fibrillar C-S-H phases (Fig. 8, Fig. 9, perspective of chemical reaction equilibrium, the amount of gypsum
and Fig. 10(a)). The reason for formation of fibrillar C-S-H may be (1) that reacted and was produced was equal, which indicates that gypsum
the relative high Ca/Si ratio. It is seen from Fig. 10(a) that Ca/Si ratio of plays a catalytic role in this system. This is the reason why the optimum
C-S-H is around 2, which is because of the existence of C2S and C3S in amount of gypsum is small (mass ratio G/SS = 1/16) (Fig. 3).
steel slag. Richardson et al. [25,27] showed the overview of the nature The BSE results show the formation of calcium carbonate and the
of the hydration products formed in the calcium silicates systems (C3S accumulation of a silicon-rich phase on the surfaces of calcium silicate
and β-C2S). They proved that C-S-H has fibrillary, directional mor- particles after carbonation. During the carbonation process, the for-
phology by scanning and transmission electron microscope (SEM and mation of a densely aggregated CaCO3 coating and Ca-depleted silicate
TEM) and NMR, which is in consistent with result of this study. (2) rims will have hindered the diffusion of Ca2+ from the uncarbonated
loose structure of the system. Morphology of C-S-H depends on the core part of the calcium silicate particle, accounting for the declining
availability of free space (pore network) where hydration products can reaction rate [44] (Fig. 9).
precipitate. From MIP analysis results of N2-1d sample in Fig. 15,
structure of non‑carbonated sample is loose, which means that enough

10
X. Wang, et al. Cement and Concrete Research 125 (2019) 105893

Fig. 16. Hydration and carbonation mechanisms in this system.

4.2. Correlation between the phase assemblage and strength 5. Conclusions

The compressive strength of specimens without carbonation curing The joint effects of carbonation and the addition of gypsum on the
was low, with a value of only 3.71 MPa after one day (Table 3). The compressive strength and CO2 uptake of steel slag blocks were in-
increase in strength mainly depends on the hydration of dicalcium si- vestigated in this study. The hydration and carbonation mechanisms
licate and tricalcium silicate in the steel slag to form C-S-H [33] and a were explored using microstructural analyses such as XRD, SEM-EDS,
small amount of ettringgite (Fig. 9), which can form an intertwined FTIR, TG, XPS, and MIP. The following conclusions can be drawn.
structure that promotes some strength increase. In the non‑carbonsated
sample, some C-S-H are near-amorphous C-S-H gels like that in portland 1. When the mass ratio of water to solid is 0.2 and the compaction
cement, and some of the C-S-H phases are with a fibrillar morphology, strength is 9 MPa, the compressive strength of test blocks after one-
which is thought to be more loose than traditional gel-like C-S-H thus day carbonation increases then decreases with the addition of in-
leads to a relatively loose structure (Fig. 15) and low strength. But this creasing amounts of gypsum. A test block with a gypsum/steel slag
structure helps provide sufficient porosity for the ingress and diffusion mass ratio of 1/16 had the highest strength of all of the samples,
of CO2. reaching 32 MPa at one day of curing. However, the 28-day strength
In samples that were subjected to CO2 curing, the carbonation depth of CO2 cured samples with gypsum was weaker than their 14-day
gradually increased in the early stage of the carbonation reaction, strength because of decomposition of C-S-H gels and mono-
reaching the core of the test block in 3 h (Fig. 7), proving that the carbonate. Therefore, short-term curing is suitable for this system.
diffusion rate of CO2 was high. The carbonation reaction proceeded 2. The overall CO2 uptake of test blocks reduced with an increase in
from the outside to the inside of the particles. The mechanisms of the gypsum content; this is due to the decrease of the amount of steel
strength formation can be explained as follows. Firstly, carbonized slag, which plays the main carbon fixation role in the test block.
products precipitate in the pores of the matrix [42,43] and small pre- Meanwhile, the CO2 uptake of the steel slag within the block in-
cipitated particles of varying sizes join together, reducing the overall creased almost linearly with an increase in gypsum content in the
pore size in the test block (Fig. 15 and Table 5) at the early carbonation block, showing that the addition of gypsum promotes the carbona-
stage (0-14d). It is known that the volume of pores with diameters tion of steel slag to some extent.
between 50 nm and 10 μm is well correlated with the mechanical 3. CO2 uptake was positively correlated with compressive strength.
strength of cementitious materials, i.e., a lower pore volume leads to This is because carbonation plays a key role in promoting the early
higher strength and better durability [45]. Secondly, the irreversible strength of the test block, as the hydration reaction of the steel slag
drying shrinkage that is caused by the water-loss happened along with is slow and rarely contributes to early strength.
the carbonation [46]. The refining of the pores and decrease of the 4. The main hydration product were C-S-H phases and ettringite, and
porosity caused by carbonation contributed to the increase of the the main carbonation products were calcite and monocarbonate.
compressive strengths of the specimens. Thirdly, the CaCO3 crystals The addition of gypsum may promote the rapid hydration of steel
were bonded together and hence performed a binding role themselves slag to form ettringite, which reacts with CO2 and produces mono-
(Fig. 10(b), Fig. 10(c), and Fig. 11(a)). It is known that the na- carbonate and/or calcite. Gypsum plays a catalytic role in this
noindentation modulus and hardness of CaCO3 (38.9 ± 12.1 GPa and system.
1.79 ± 0.63 GPa, respectively) are higher than that of hydration C-S-H 5. Strength formation of carbonated samples at early stage is asso-
products (18.2 ± 4.2 GPa and 0.45 ± 0.14 GPa for low-density C-S-H ciated with (1) carbonized products precipitation in the pores of the
and 29.1 ± 4.0 GPa and 0.83 ± 0.18 GPa for high-density C-S-H, re- matrix to reduce pore size, (2) the irreversible drying shrinkage that
spectively) [47]. That is why the mechanical strength of blocks con- refine the pore, (3) the CaCO3 crystals bonding together to per-
taining a large amount of calcium carbonate was almost 5 to 10 times formed a binding role themselves.
stronger than of a hydration test block with the same curing time. 6. The extended carbonation of hydrates results in the decrease of the
Therefore, CO2-cured samples had higher strength than fresh samples solid volume and consequently in the increase of porosity, resulting
(Table 3). in a decrease in strength.
The compressive strength of samples cured for 28 d was lower than
that of 14-d cured samples. This may be because the extended carbo- While the preparation of carbonatized blocks using steel slag and
nation of hydrates results in the decrease of the solid volume and gypsum has been demonstrated in this study, the long-term environ-
consequently in the increase of porosity. The destruction of the C-S-H mental safety and durability of this material requires additional re-
gel and ettringite damages the framework, resulting in a decrease in search and will be reported in a forthcoming study.
strength.

11
X. Wang, et al. Cement and Concrete Research 125 (2019) 105893

Acknowledgments [23] I.G. Richardson, The calcium silicate hydrates, Cem. Concr. Res. 38 (2) (2008)
137–158.
[24] I.G. Richardson, The nature of C–S–H in hardened cements, Cem. Concr. Res. 29
The authors gratefully acknowledge the support of the National Key (1999) 1131–1147.
Research and Development Program of China (2018YFC1900603), [25] I. Richardson, The nature of the hydration products in hardened cement pastes,
Fundamental Research Funds for the Central Universities (FRF-TP-17- Cem. Concr. Compos. 22 (2000) 97–113.
[26] I.G. Richardson, Tobermorite/jennite- and tobermorite/calcium hydroxide-based
077A1), and project funding by the China Postdoctoral Science models for the structure of C-S-H: applicability to hardened pastes of tricalcium
Foundation (2018M631342). silicate, β-dicalcium silicate, Portland cement, and blends of Portland cement with
blast-furnace slag, metakaolin, or silica fume, Cem. Concr. Res. 34 (9) (2004)
1733–1777.
References [27] E. Kapeluszna, Ł. Kotwica, A. Różycka, et al., Incorporation of Al in C-A-S-H gels
with various ca/Si and Al/Si ratio: microstructural and structural characteristics
[1] F. Han, Z. Zhang, D. Wang, P. Yan, Hydration heat evolution and kinetics of blended with DTA/TG, XRD, FTIR and TEM analysis, Constr. Build. Mater. 155 (2017)
cement containing steel slag at different temperatures, Thermochim. Acta 605 643–653.
(2015) 43–51. [28] I. Pane, W. Hansen, Investigation of blended cement hydration by isothermal ca-
[2] C. Shi, Steel slag — its production, processing, characteristics, and cementitious lorimetry and thermal analysis, Cem. Concr. Res. 35 (2005) 1155–1164.
properties, J. Mater. Civil Eng. 16 (3) (2004) 230–236. [29] J. Chang, C. Xiong, Y. Zhang, et al., Foaming characteristics and microstructure of
[3] J. Guo, Y. Bao, M. Wang, Steel slag in China: treatment, recycling, and manage- aerated steel slag block prepared by accelerated carbonation, Constr. Build. Mater.
ment, Waste Manag. 78 (2018) 318–330. 209 (2019) 222–233.
[4] Y. Jiang, T.C. Ling, C. Shi, S.Y. Pan, Characteristics of steel slags and their use in [30] L. Black, A. Stumm, K. Garbev, P. Stemmermann, K.R. Hallam, G.C. Allen, X-ray
cement and concrete—a review, Resour. Conserv. Recycl. 136 (2018) 187–197. photoelectron spectroscopy of the cement clinker phases tricalcium silicate and β-
[5] M. Frias Rojas, M.I. Sanchez de Rojas, Chemical assessment of the electric arc dicalcium silicate, Cem. Concr. Res. 33 (2003) 1561–1565.
furnace slag as construction material: expansive compounds, Cem. Concr. Res. 34 [31] L. Black, C. Breen, J. Yarwood, K. Garbev, P. Stemmermann, B. Gasharova,
(2004) 1881–1888. Structural features of C-S-H(I) and its carbonation in air—a Raman spectroscopic
[6] J. Geiseler, Use of steelworks slag in Europe, Waste Manag. 16 (1996) 59–63. study. Part II: carbonated phases, J. Am. Ceram. Soc. 90 (3) (2007) 908–917.
[7] I. Akin Altun, I. Yilmaz, Study on steel furnace slags with high MgO as additive in [32] L. Black, K. Garbev, G. Beuchle, P. Stemmermann, D. Schild, X-rayphotoelectron
Portland cement, Cem. Concr. Res. 32 (8) (2002) 1247–1249. spectroscopic investigation of nanocrystalline calcium silicate hydrates synthesised
[8] L. Mo, F. Zhang, M. Deng, F. Jin, A. Al-Tabbaa, A. Wang, Accelerated carbonation by reactive milling, Cem. Concr. Res. 36 (2006) 1023–1031.
and performance of concrete made with steel slag as binding materials and ag- [33] L. Black, K. Garbev, P. Temmermann, K.R. Hallam, G.C. Allen, Characterisation of
gregates, Cem. Concr. Compos. 83 (2017) 138–145. crystalline C-S-H phases by X-ray photoelectron spectroscopy, Cem. Concr. Res. 33
[9] N.L. Ukwattage, P.G. Ranjith, X. Li, Steel-making slag for mineral sequestration of (2003) 899–911.
carbon dioxide by accelerated carbonation, Measurement 97 (2017) 15–22. [34] K. Okada, Y. Kameshima, A. Yasumori, Chemical shifts of silicon Xray photoelectron
[10] T. Gao, L. Shen, M. Shen, L. Liu, F. Chen, L. Gao, Evolution and projection of CO2 spectra by polymerization structures of silicates, J. Am. Ceram. Soc. 81 (7) (1998)
emissions for China's cement industry from 1980 to 2020, Renew. Sust. Energ. Rev. 1970–1972.
74 (2017) 522–537. [35] L. Black, K. Garbev, P. Stemmermann, K.R. Hallam, G.C. Allen, X-ray photoelectron
[11] B. Lin, Z. Zhang, Carbon emissions in China' s cement industry: a sector and policy study of oxygen bonding in crystalline C-S-H phases, Phys. Chem. Miner. 31 (2004)
analysis, Renew. Sust. Energ. Rev. 58 (2016) 1387–1394. 337–346.
[12] C. Feng, J.B. Huang, M. Wang, Y. Song, Energy efficiency in China's iron and steel [36] L. Black, K. Garbev, I. Gee, Surface carbonation of synthetic c-s-h samples: a com-
industry: evidence and policy implications, J. Clean. Prod. 177 (2018) 837–845. parison between fresh and aged c-s-h using x-ray photoelectron spectroscopy, Cem.
[13] X. Wang, B. Lin, How to reduce CO2 emissions in China's iron and steel industry, Concr. Res. 38 (6) (2008) 745–750.
Renew. Sust. Energ. Rev. 57 (2016) 1496–1505. [37] B. Demri, D. Muster, XPS study of some calcium compounds, J. Mater. Process.
[14] Z. Ghouleh, R.I. Guthrie, Y. Shao, Production of carbonate aggregates using steel Technol. 55 (3) (1995) 311–314.
slag and carbon dioxide for carbon-negative concrete, J. CO₂ Util. 18 (2017) [38] H. Chen, Z. Chen, G. Zhao, Z. Zhang, C. Xu, Y. Liu, et al., Enhanced adsorption of U
125–138. (VI) and 241 am(III) from wastewater using ca/Al layered double hydroxide@
[15] S.Y. Pan, R. Adhikari, Y.H. Chen, P. Li, P.C. Chiang, Integrated and innovative steel carbon nanotube composites, J. Hazard. Mater. 347 (2017) 67.
slag utilization for iron reclamation, green material production and CO2 fixation via [39] E. Kapeluszna, Ł. Kotwica, R. Agnieszka, et al., Incorporation of Al in C-A-S-H gels
accelerated carbonation, J. Clean. Prod. 137 (2016) 617–631. with various ca/Si and Al/Si ratio: microstructural and structural characteristics
[16] M. Salman, Ö. Cizer, Y. Pontikes, R.M. Santos, R. Snellings, L. Vandewalle, ... K. Van with DTA/TG, XRD, FTIR and TEM analysis, Constr. Build. Mater. 155 (2017)
Balen, Effect of accelerated carbonation on AOD stainless steel slag for its valor- 643–653.
isation as a CO2-sequestering construction material, Chem. Eng. J. 246 (2014) [40] E. L'Hôpital, B. Lothenbach, G. Le Saout, et al., Incorporation of aluminium in
39–52. calcium-silicate-hydrates, Cem. Concr. Res. 75 (2015) 91–103.
[17] Z. Ghouleh, R.I. Guthrie, Y. Shao, Production of carbonate aggregates using steel [41] A.C.A. Muller, K.L. Scrivener, A reassessment of mercury intrusion porosimetry by
slag and carbon dioxide for carbon-negative concrete, J. CO₂ Util. 18 (2017) comparison with, 1 H NMR relaxometry, Cem. Concr. Res. 100 (2017) 350–360.
125–138. [42] V.D. Pizzol, L.M. Mendes, L. Frezzatti, H. Savastano Jr., G.H.D. Tonoli, Effect of
[18] H. Jo, M.G. Lee, J. Park, K.D. Jung, Preparation of high-purity nano-CaCO3 from accelerated carbonation on the microstructure and physical properties of hybrid
steel slag, Energy 120 (2017) 884–894. fiber-cement composites, Mine. Eng. 59 (3) (2014) 101–106.
[19] S. Teir, T. Kotiranta, J. Pakarinen, H.P. Mattila, Case study for production of cal- [43] O.O. Metalssi, A. Aït-Mokhtar, P. Turcry, B. Ruot, Consequences of carbonationon
cium carbonate from carbon dioxide in flue gases and steelmaking slag, J. CO₂ Util. microstructure and drying shrinkage of a mortar with cellulose ether, Constr. Build.
14 (2016) 37–46. Mater. 34 (3) (2012) 218–225.
[20] Y. Zhang, S. Zhang, W. Ni, et al., Immobilisation of high-arsenic-containing tailings [44] Z. Ghouleh, R.I.L. Guthrie, Y. Shao, High-strength KOBM steel slag binder activated
by using metallurgical slag-cementing materials, Chemosphere 223 (2019) by carbonation, Constr. Build. Mater. 99 (2015) 175–183.
117–123. [45] A. Hartmann, M. Khakhutov, J.C. Buhl, Hydrothermal synthesis of csh-phases (to-
[21] Y. Kong, P. Wang, S. Liu, Microwave pre-curing of Portland cement-steel slag bermorite) under influence of ca-formate, Mater. Res. Bull. 51 (2) (2014) 389–396.
powder composite for its hydration properties, Constr. Build. Mater. 189 (2018) [46] H.F.W. Taylor, Cement Chemistry, second ed., Thomas Telford publishing, London,
1093–1104. 1997.
[22] Y. Yang, B. Yuan, Y. Wang, et al., Carbonation resistance cement for CO2 storage [47] A. Nonat, The structure and stoichiometry of C-S-H, Cem. Concr. Res. 34 (9) (2004)
and injection wells, J. Pet. Sci. Eng. 146 (2016) 883–889. 1521–1528.

12

You might also like