Available online at www.sciencedirect.

com

Cement & Concrete Composites 30 (2008) 239–244

www.elsevier.com/locate/cemconcomp

Bonding and abrasion resistance of geopolymeric repair

material made with steel slag
Shuguang Hu a, Hongxi Wang a,b,*
, Gaozhan Zhang a, Qingjun Ding a

a
School of Materials Science and Engineering, Wuhan University of Technology, Wuhan 430070, China
b
Key Laboratory of Nonferrous Materials and Processing Technology, Guilin University of Technology, Ministry of Education, Guilin 541004, China

Received 30 April 2006; received in revised form 23 March 2007; accepted 11 April 2007
Available online 18 April 2007

Abstract

Three repair materials were prepared using cement-based, geopolymeric, or geopolymeric containing steel slag binders. Their mechan-
ical performances such as compressive strength, bond strength and abrasion resistance were examined experimentally. The test results
showed that the geopolymeric materials had better repair characteristics than cement-based repair materials, and the addition of steel
slag could improve significantly the abrasion resistance of geopolymeric repair. By means of scanning electron microscopy (SEM) it
can also be concluded that the steel slag was almost fully absorbed to take part in the alkali-activated reaction and be immobilized into
the amorphous aluminosilicate geopolymer matrix.
 2007 Elsevier Ltd. All rights reserved.

Keywords: Geopolymer; Steel slag; Repair material

1. Introduction durability, better acid resistance [4]. Besides, it is also a

‘‘Green Material’’ for its low manufacturing energy con-
Pioneered by Davidovits in the late 1970s [1], geopoly- sumption and low waste gas emission [5]. Because of these
mers are a novel class of materials that are formed by the prominent characteristics, geopolymer was considered as
polymerization of silicon, aluminum, and oxygen species one of the potential candidates to solve the conflict between
to form an amorphous three-dimensional framework struc- social development and environmental pollution from bin-
ture. Geopolymeric reactants could range from kaolinite or der [6]. However, little work has been reported about geo-
metakaolin to a group of materials containing rich SiO2 polymeric repair materials.
and/or Al2O3 oxides, e.g., fly ash, slag, construction waste On the other hand, steel slag is a by-product of the con-
and natural minerals. As we know, chemical bonds of Si–O version of iron to steel, which results in large environmen-
and Al–O are among the most stable covalent bonds in nat- tal pollution in China. Because olivine, merwinite, C3S,
ure. In addition, the polycondensation degree of geopoly- C2S, C4AF, C2F, RO phase (CaO–FeO–MnO–MgO solid
mer is much higher than cement-based materials [2]. solution) and free-CaO are common minerals in steel slag,
Therefore, geopolymer materials possess many advanced the presence of C3S, C2S, C4AF, C2F endows steel slag its
properties such as the ease with which it can be recycled, cementitious properties. But the reactivity of steel slag is
excellent compressive and bond strength [3], long-term small because of its thermal history, as a consequence, its
large application in the concrete and cement as a kind of
additive is almost impossible. However, the authors [7]
*
investigated water-glass-activated steel slag/blast furnace
Corresponding author. Address: School of Materials Science and
Engineering, Wuhan University of Technology, Wuhan 430070, China.
slag (BFS) and concluded that the alkali-activated mixture
Tel.: +86 27 87668602; fax: +86 27 87392350. possess much higher strengths than Grade-425 ordinary
E-mail address: [email protected] (H. Wang). Portland cement.

0958-9465/$ - see front matter  2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.cemconcomp.2007.04.004
240 S. Hu et al. / Cement & Concrete Composites 30 (2008) 239–244

Good bond and wear resistance are the two main Table 3
requirements for successful repair [8,9]. The geopolymers The mixture proportion of the repair materials (wt.%)
possess good bond strength and the steel slag is good Mixture number Water:cement:UEA Metakaolin Steel Water-
wear resistance because of the Fe phase existing in it. In slag glass
this paper, we focused on reporting the influences of Cement repair (CR) 0.4:0.95:0.05
the steel slag to the properties of geopolymeric repair Geopolymeric repair 1 0.56
(GR)
materials. Geopolymeric repair 0.8 0.2 0.56
with steel slag
(GRS)
2. Experimental

2.1. Materials
2.2. Methods
Metakaolin used in this study was obtained by calcining
2.2.1. Mixture proportion
kaolin at 750 C for 6 h. It was used for reactive Si–Al
In this study, three kinds of repair materials were pre-
cementitious materials. Steel slag powder used in this study
pared. The details of mixture proportion, compressive
was obtained from Wuhan Iron and Steel Corp., Wuhan,
strength of the repair materials and old substrate are given
China. Solids from different batches were bulk mixed prior
in Table 3. The NaOH and sodium silicate solution, and
to their use and the compositions as determined by fusion
water were first mixed in a beaker to form the alkali-
analysis using a Siemens SRS3000 sequential X-ray fluores-
activated solution (water-glass) that the SiO2/Na2O molar
cence (XRF) spectrometer are listed in Table 1. Their par-
ration was 1.14 and the mass concentration was 38%,
ticle size distribution is shown in Table 2. The cement used
and cooled up to room temperature. Powder metakaolin
as the control sample was Grade-42.5 ordinary Portland
and steel slag was mixed for 2 min, and then they were
cement. The main component of U-type expansive agent
added into the cooled alkaline solution and mixed for
(UEA) is calcium sulphoaluminate.
another 3 min.
Analytical grade NaOH and sodium silicate solution
with the molar ratio of SiO2 to Na2O of 3.19 were used
2.2.2. Compressive strength testing
as reagents, respectively. Distilled water was used
The fresh geopolymer with or without steel slag and
throughout.
cement pastes were rapidly poured into steel moulds of
dimension with 40 · 40 · 40 mm. The specimens were
demoulded after 24 h curing at 20 ± 0.5 C and 95% rela-
Table 1
tive humidity, and then submitted to compressive strength
Composition of kaolinite as determined by fusion and XRF analysis
(wt.%) test at curing ages of 8 h, 1 d, 3 d, 7 d, 28 d.
Element Kaolinite Steel slag Cement
as oxide (wt.%) (wt.%) (wt.%) 2.2.3. Bond strength testing
Samples for interfacial bonding strength determination
SiO2 53.1 15.0 22.2
Al2O3 30.9 6.7 4.8 between the new repair materials and the old concrete sub-
CaO 0.1 44.2 62.5 strate were prepared as shown in Fig. 1. The old substrates
Fe2O3 1.6 15.4 6.2 with the proportion of water:cement:sand (0.4:1:3) were
MgO 0.3 10.9 1.5 cast as indicated in Fig. 1, then cured at 20 C and 95%
TiO2 2.0 3.1
Na2O 0.3 0.2 0.3
relative humidity for 28 d. As indicated in Fig. 1, freshly
K2O 0.1 0.1 prepared repair layer (5 mm thick) was then glued to the
SO3 0 0.7 1.1 middle of the old substrates for the bond strength test.
f-CaO 1.7
Fe 1.0
Other minor oxides 0.3 0.4
Loss on ignition 11.3 0.6 1.4
Total 100.00 100.00 100.00

Table 2
Particle size distribution of raw materials
Diameters 0.5 1 3 5 10 15 20 30 40
(lm)
Kaolinite 2.85 20.3 65.7 89.9 97.3 99.3 99.6 99.8 100
Fig. 1. Test set-up of the bond strength between the substrate mortar-
Steel slag 0 0.67 15.9 40.1 72.9 89.6 92.1 94.6 100
repair materials in sandwich specimens.
 S. Hu et al. / Cement & Concrete Composites 30 (2008) 239–244 241

Table 4
Compressive strength of the repair materials at various ages
Mixture number Compressive strength (MPa)
8h 1d 3d 7d 28 d
Cement repair (CR) – 8.7 23.2 33.1 46.1
Geopolymeric repair (GR) 10.4 17.2 30.4 37.8 40.9
Geopolymeric repair with steel slag 14.3 22.1 35.6 40.4 44.5
(GRS)

than that of the geopolymeric repair material with or with-

out steel slag at the 8 h, 1 d, 3 d, 7 d, but at 28-d, the results
were the opposite. The higher early compressive strength of
geopolymeric repair materials could be attributed to their
activated character. Comparing GRS with GR, it was
Fig. 2. The sketch of the test set-up for abrasion resistance.
found that the 8 h, 1 d, 3 d, 7 d, 28 d compressive strengths
increased 43%, 28%, 17%, 6.9%, 7.6%, respectively. The
When the cement repair material was glued, the resultant addition of steel slag could accelerate the setting time and
sandwich specimen was named Cb. When the geopolymer significantly improve compressive strength, which could
was glued, the resultant sandwich specimen was named be due to its latent hydraulic cementitious character.
Gb. When the geopolymer with steel slag was glued, the
resultant sandwich specimen was named GSb.
3.2. Bond strength
2.2.4. Abrasion resistance
The results of the 8 h, 1 d, 3 d, 7 d, 28 d bond strength
Abrasion resistance was tested in accordance with
tests on the various repair materials are summarized in
GB16925-1997, which was published by the Chinese Build-
Table 5. As shown in Table 5, the bond strengths of Gb
ing Material Industry, which the test set-up shown in
and GSb at 1 d were both higher than of Cb at 7 d. Simi-
Fig. 2. The 100 mm diameter cylinders with a 2:1 diame-
larly, the bond strength of Gb and GRb at 3 d were both
ter-to-length ratio were fixed in the holding device of the
higher than of Cb. The bond strength of GSb at 3 d was
abrasion machine and a load of 154 ± 2.5 N was applied.
2.6% and 600% higher than of Gb, Cb, respectively. Simi-
The grinding machine was then put in motion at a speed
larly, the bond strength of GSb at 28 d was 4.4% and
of 1000 rpm. Each specimen was abraded for 5 min and
55.9% higher than of Gb, Cb, respectively. It could be seen
the depth of grinding trace on the specimen surface was
that the geopolymeric materials possessed better bonding
measured at the fifth minute. The results for the grade of
character than cement materials.
abrasion resistance were evaluated using the following
The noticeable increase in bond strength was due to rea-
formula:
pffiffiffi sons: the improvement of the interfacial transition zone
I a ¼ R=P ð1Þ between the old substrate and repair materials and the
increasing of the tensile strength of repair material. Fig. 3
where Ia is the grade of abrasion resistance; R is the num- shows the failure modes of specimens for bond strength
ber of revolution of grinding machine in 1000 rpm; P is the test. Failures occurred in the three different regions: (1) in
depth of grinding trace in mm. the interfacial transition zone between the substrate and
repair materials, named mode A; (2) in the repair materials,
2.2.5. Microstructure analysis named mode B; and (3) throughout the substrate and the
Scanning electron microscopy (SEM) was used to inves- repair materials, named mode C. The failure mode of the
tigate the images of the geopolymeric repair materials with repair materials at different age can be seen in Table 5.
steel slag. The sandwich samples used in the bond strength The failure mode of cement repair materials was mainly
testing were destroyed and then lightly polished at the mode A. The failure mode of the geopolymeric repair
cracking before micrographs were taken for SEM test. materials was mainly mode B at the age of 8 h, 1 d, and

3. Results and discussion

Table 5
Bond strength and failure mode of the repair materials
3.1. Compressive strength
Mixture Bond strength (MPa) Failure mode
number 8h 1d 3d 7d 28 d 8h 1d 3d 7d 28 d
The results of the 8 h, 1 d, 3 d, 7 d, 28 d compressive
strength tests on the various repair materials are summa- Cb – – 0.34 1.17 1.95 B A A A A
rized in Table 4. From Table 4, it was found that the com- Gb 0.42 1.23 2.34 2.43 2.91 B B C C C
GSb 0.45 1.27 2.40 2.47 3.04 B B C C C
pressive strength of the cement repair material is lower
242 S. Hu et al. / Cement & Concrete Composites 30 (2008) 239–244

Fig. 3. Failure modes of specimens in bond strength test.

was mainly mode C at the age of 3 d, 7 d, 28 d. Failure Table 6

mode A was attributive to the week interface, and failure Depth of grind track of the repair materials
mode B or C was attributive to the week tensile strength Mixture number P (mm)
of the geopolymeric repair materials or the old substrate. 3d 7d 28 d 56 d 90 d
It could be inferred that the interfacial transition zone Cement repair (PC) 6.41 5.42 4.21 4.01 3.89
between the old substrate and geopolymeric repair materi- Geopolymeric repair (PG) 3.32 3.03 2.98 2.87 2.76
als could be improved because of the reaction between the Geopolymeric repair with steel slag 3.02 2.76 2.34 2.23 2.14
alkali activator and the surface product of the old sub- (PGS)
strate, thus leading to an even denser interface zone with
better bond strength.
PG, the P values at 3, 7, 28, 56 and 90 d decreased 9%,
3.3. Abrasion resistance 8.9%, 21.4%, 22.2%, 22.5%, respectively. This showed that
the depth of wear for geopolymeric repair materials was
The abrasion resistance of repair materials was deter- smaller than for the cement repair materials, which means
mined at the ages of 3, 7, 28, 56 and 90 d. It was measured that the abrasion resistance for geopolymeric repair mate-
in term of depth of wear as shown in Table 6. It was rials was better. The decreasing extent for the PG and
observed that depth of wear decreased with the increase PGS value from 3 d to 28 d is smaller than for the PC, which
of the age for the repair materials. Comparing PG with is similar to the relation in their compressive strengths. This
PC, the P values at 3, 7, 28, 56 and 90 d decreased 48%, is mainly because the abrasion resistance of repair materi-
44%, 29%, 28%, 29%, respectively. Comparing PGS with als is determined by the density of the structure. The geo-

Fig. 4. The grade of abrasion resistance of the repair materials at various ages.
 S. Hu et al. / Cement & Concrete Composites 30 (2008) 239–244 243

Fig. 5. SEM tested of geopolymeric repair with steel-making slag at 1 d, 20 ± 0.5 C, 95% relative humidity: (a) 1000· and (b) 5000·.

Fig. 6. SEM tested of geopolymeric repair with steel-making slag at 28 d, 20 ± 0.5 C, 95% relative humidity: (a) 1000· and (b) 5000·.

polymeric repair materials could form a dense structure the steel slag particles have almost fully reacted with the
earlier than the cement repair materials. sodium silicate solution and could not be observed from
The calculated value about the grade of abrasion resis- the SEM, which could be the reason why the compressive
tance for the repair materials at various ages is plotted and bond strengths are higher at 28 d age.
in Fig. 4. Fig. 4 shows the grade of abrasion resistance
increased with the increase of the age for the repair materi- 4. Conclusions
als. It could be seen that the Ia for the geopolymeric repair
is higher than for the cement repair, but lower than for the Within the indicated scope of this study, the particular
geopolymeric repair with steel slag. conclusions may be summarized as follows:

3.4. Microstructure analysis 1. The compressive strength of the cement repair material
is lower than that of the geopolymeric repair material
It could be observed in Fig. 4 that the geopolymeric with or without steel slag at the 8 h, 1 d, 3 d, 7 d, but
repair with steel slag had the best abrasion resistance, at 28-d, the opposite is true.
which could be due to the addition of steel slag. Figs. 5 2. The geopolymeric repair materials have better repair
and 6 show some microstructural characteristics of the characteristics than cement-based repair materials. The
geopolymeric repair materials with steel slag by means of bond strength of GSb at 3 d was 2.6% and 600% higher
scanning electron microscopy (SEM) methods. than those of Gb and Cb, respectively. Similarly, the
As can be observed in these figures, the geopolymeric bond strength of GSb at 28 d was 4.4% and 55.9% higher
repair materials studied through SEM have quite different than those of Gb, Cb, respectively.
microstructure at different ages. At 1 d age, 20 ± 0.5 C, 3. The geopolymeric repair materials had better abrasion
95% relative humidity, the resulting material is a porous resistance character than the cement repair. The PG
one (see Fig. 5a). In many cases, a surface reaction is values at 3, 7, 28, 56 and 90 d decreased 48%, 44%,
responsible for bonding the undissolved steel slag particles 29%, 28%, 29% than PC, respectively. The addition of
into the final aluminosilicate matrix structure. The reacted steel slag could improve significantly the abrasion
steel slag particles appear to be immobilized into the amor- resistance performance of the geopolymeric repair.
phous aluminosilicate matrix (see Fig. 5b). Fig. 6a shows Comparing PGS with PG, the P values at 3, 7, 28, 56
that the geopolymeric repair materials at 28 d age, and 90 d decreased 9%, 8.9%, 21.4%, 22.2%, 22.5%,
20 ± 0.5 C, 95% relative humidity had more compact respectively.
structure than the geopolymeric repair materials at 1 d 4. The steel slag was almost fully absorbed to take part in
age, 20 ± 0.5 C, 95% relative humidity. Fig. 6b shows that the alkali-activated reaction and be immobilized into the
244 S. Hu et al. / Cement & Concrete Composites 30 (2008) 239–244

amorphous aluminosilicate geopolymer matrix by [2] Yang NR. Physical chemistry basis for the formation of alkaline
means of scanning electron microscopy (SEM). cementitious material. J Chin Ceram Soc 1996;24(4):459–65.
[3] van Jaarsveld JGS, van Deventer JSJ, Lukey GC. The characterisation
of source materials in fly ash based geopolymers. Mater Lett
2003;57(7):1272–86.
Acknowledgements [4] Phair JW, Van Deventer JSJ. Effect of silicate activator pH on the
leaching and material characteristics of waste-based geopolymers.
The authors gratefully acknowledge the key technology Miner Eng 2001;14(3):289–304.
[5] Xu Hua, Van Deventer JSJ. The geopolymerisation of alumino-silicate
R&D program of 10th Five-Year Plan of China for finan- minerals. Int J Miner Process 2000;59(3):247–66.
cially supporting the Project No. 2003BA652C and the [6] Davidovits J. Method for bonding fiber reinforcement on concrete and
Opening Fund of Key Laboratory of Nonferrous Materials steel structures and resultant products, US Patent No. 5925449; 1999.
and Processing Technology. [7] Hu SG, Wei JX, Ding QJ. Research on excitation principle of sodium
silicate to steel slag cement. Cement Eng 2001;5:4–6 [in Chinese].
[8] Xiong GJ, Cui YQ, Chen LQ, Jiang H. Influence of hydrochloric acid
References etching on bond strength between concrete substrate and repair
materials. Cem Concr Compos 2004;26:41–5.
[1] Davidovits J. Geopolymers and geopolymeric new materials. J Therm [9] Siddique R. Performance characteristics of high-volume class F fly ash
Anal 1989;35(2):429–41. concrete. Cem Concr Res 2004;34:487–93.