Construction and Building Materials 272 (2021) 121971

Contents lists available at ScienceDirect

Construction and Building Materials

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

Effect of chloride attack on the bonded concrete system repaired by

UHPC
Qing Chen a,b, Rui Ma b, Haoxin Li a,b,⇑, Zhengwu Jiang a,b, Hehua Zhu c, Zhiguo Yan c
a
Key Laboratory of Advanced Civil Engineering Materials (Tongji University), Ministry of Education, 4800 Cao’an Road, Shanghai 201804, China
b
School of Materials Science and Engineering, Tongji University, Shanghai 201804, China
c
State Key Laboratory for Disaster Reduction in Civil Engineering, Tongji University, 1239 Siping Road, Shanghai 200092, China

h i g h l i g h t s

Effects of the chloride attack are studied on the bonded system repaired by UHPC.

The interfacial properties decrease after the chloride attack for the repaired system.

The chloride contents in the matrix layer is much higher than that of UHPC layer.

The UHPC bonded system demonstrate higher polarization resistances and free-corrosion potential.

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

Article history: UHPC is becoming a promising repairing material. However, little attention is paid on the effect of the
Received 1 August 2020 chloride attack on the bonded system repaired by UHPC. In this study, UHPC with different fiber contents
Received in revised form 29 November 2020 are prepared to form the repaired systems, whose compressive, bonding and flexural behaviors are inves-
Accepted 7 December 2020
tigated after different periods of chloride attack. Meanwhile, the UHPC bonded system in the fresh water
Available online 23 December 2020
is taken as the reference. The polarization curve and polarization resistance evolution are reached to
access the electrochemical behavior of the repaired systems after the chloride attack. Furthermore, the
Keywords:
scanning electron microscope (SEM) and energy dispersive spectroscopy (EDS) are employed to analyze
Ultra-high performance concrete
Repaired system
the bonding area’s microstructures and the chloride distributions. Compared with the results of the ref-
Chloride attack erence group in the fresh water, the repaired systems’ compressive strength after 56d chloride attack
Mechanical decreased by about 5% to 10% and the bonding strength demonstrates a reduction of 16.1% to 23.9%.
Electrochemical The decreases of the maximum load after the chloride attack varied from 0.51KN to 1.71KN in the flexural
Microstructure test herein. The polarization resistances and free corrosion potential reduced to 1060 Xcm2 and
492 mV for the repaired system with the normal mortar material, respectively. The values are about
6298 Xcm2 (or 15056 Xcm2) and 189 mV (or 168 mV) for different UHPC repaired system. The
microstructures of matrix layer are looser than those of the UHPC layer and the chloride contents in
the matrix layer is about 4 times than that of the UHPC layer, which further verifies that the UHPC layer
has a better resistance to the chloride attack than the matrix layer.
Ó 2020 Elsevier Ltd. All rights reserved.

1. Introduction improve the capacity of the damaged RC beams [9–12]. It can

remarkably enhance the flexural strength of the strengthened
Repairing damaged concrete is important for the sustainable beams at the early age [13–16]. Meanwhile, the existing RC beam’s
development of civil engineering [1–4]. Recently, the ultra-high stiffness can be significantly increased with the UHPC layer [17–
performance concrete (UHPC) is developed and emerged as a 20].
new repairing material due to the excellent performance under For a repaired system composed of the UHPC materials and the
the different conditions [5–8]. The UHPC material is adopted to normal concrete (NC), the bonding surface between the two phases
is usually the vulnerable part [21–26] and the interfacial behavior
⇑ Corresponding author at: Key Laboratory of Advanced Civil Engineering is critical for the performance of the whole repairing system. Sev-
Materials (Tongji University), Ministry of Education, 4800 Cao’an Road, Shanghai eral researchers have investigated the interfacial behavior of the
201804, China. concrete repaired system. The bond behavior was assessed from
E-mail address: [email protected] (H. Li).

https://doi.org/10.1016/j.conbuildmat.2020.121971
0950-0618/Ó 2020 Elsevier Ltd. All rights reserved.
Q. Chen, R. Ma, H. Li et al. Construction and Building Materials 272 (2021) 121971

pull-off tests and the influence of the concrete substrate’s strength sand (particle size 0.15 ~ 0.6 mm) for UHPC and river sand (fine-
was analyzed [27]. Momayez et al. studied the bond strength ness modulus 2.48) for normal cement mortar matrix, provided
between concrete substrate and the repair material with different by Shengkuo Building Materials Co. Ltd. (5) Steel fiber: straight,
evaluating approaches, including the pull-off, slant shear and the its properties are presented in Table 2. (6) Superplasticizer:
Bi-Surface shear test [28]. The UHPC interface strength was also PC200, water reducing rate 23%, provided by Inshine New Material
studied by Munoz et al. [29]. Different roughness degrees in the Technology Co. Ltd. (7) Steel bar: HRB400, 12 cm length and
concrete substrate and two different interface angles were investi- 10 mm diameter, provided by Shanghai Gongci Industrial Co. Ltd.
gated with the slant shear testing using the prismatic composite (8) Sodium chloride: analytically pure, provided by Sinopharm
specimens comprised of NC and UHPC [29]. Similar works on the Group Chemical Reagent Co. Ltd.
interfacial behavior can be found by Cohen and Kuroda [30–31]
who demonstrated that the use of key components such as the sil-
2.2. Sample preparation
ica fume and the fly ash could change the microstructure of the
transition zone, thereby improving the interfacial bonding perfor-
The UHPC repaired system is consisted of two parts: the ordi-
mance. In addition, Farzad et al. [32] established a theoretical anal-
nary cement mortar matrix and the UHPC. It should be mentioned
ysis model for the interfacial performance of the repaired system.
that the chloride penetration behavior in the matrix layer will be
Recently, Hussein el al. [33] studied the effect of very severe sul-
different when the concrete material is adopted [42–43]. More
fate environment on the bonded composite concrete system with
attention is paid on the behavior of the bonded system instead of
the UHPC, and the results showed that the compressive and flexu-
the matrix material herein. The matrix and the UHPC were pre-
ral strengths of the repaired system decreased considerably after
pared according to the mixing ratio shown in Table 3, and their
the long-term exposure to the sulfate attack. Apart from the sulfate
28d compressive and flexural strength were shown in Table 4.
attack, the other factors, such as the temperature, pH value and the
The compressive strength does not increase with the steel fiber
chloride attack, can also degrade the performance of the repaired
contents. The reason is that the flowability of UHPC may be slightly
system [34–40]. For example, under the long-term action of the
lower when the steel fiber content becomes higher, which can lead
natural environment, the concrete structure will undergo complex
to the lower density and compressive strength [46].
physical and chemical changes, resulting in the gradual degrada-
As Fig. 1 shows, different bonded samples are prepared to inves-
tion of the various performances of the structure, especially in
tigate the different performance of the repairing system. The cubic
chloride attack environment [41]. Hussein et al. [42] also investi-
specimens of 100 mm  100 mm  100 mm were used in compres-
gated the effect of severe chloride environment on the flexural
sive strength tests. The casting process is similar with that of [33].
behavior of hybrid concrete systems and their results revealed that
The prefabricated 50 mm-thick matrix material was placed in the
severe chloride exposure influenced the flexural capacity of the
bottom of the cube mold and the UHPC was cast at the top, without
concrete composites. According to Midgley et al. [43], chloride ions
any surface preparation or vibration between two layers.
could react with hydrated tricalcium aluminate to form Friedel’s
40 mm  40 mm  160 mm prism specimens were prepared to
salt. The mechanical performance such as the compressive
test the bonding and flexural behaviors. For bonding strength tests,
strength of concrete might deteriorate under the influence of chlo-
the left half of the prisms were precast matrices while they were at
rine salts [44]. Yin et al. [45] studied on the durability of the con-
the bottom half for flexural behavior tests, and the other half of the
crete structure strengthened with CFRP under the chloride
specimens were UHPC material layers. No surface preparation or
environment.
vibration were performed when casing the prism specimens as
From the above, many researchers studied the chloride effects
well.
on the performance of concrete composite and the bonded system
As Fig. 2 shows, the cubic specimens of 100 mm  100 mm
using CFRP. However, few relative reports are available on the
 100 mm were used to conduct the electrochemical test. Three
behavior of the UHPC bonded system after the chloride attack. In
types of specimens were adopted, including the MM, MUHPC
this paper, to investigate the behaviors of the UHPC bonded system
and UHPC systems, where M and UHPC respectively denotes ordi-
in chloride attack environment, the bonded repairing specimen is
nary cement mortar layer and the UHPC layer (with 1.5% fiber vol-
prepared using the matrix material and the UHPC material with
ume content). The working electrode ran through the whole
different fiber contents. The compressive strength, bonding
cementitious materials. The thickness of the protective layer (the
strength and flexural performance are measured after different
distance from the working electrode to the right surface in Fig. 2)
periods of chloride attack. The electrochemical behavior and the
was 20 mm. Five surfaces of the cube were sealed with epoxy resin.
microstructural characterizations are conducted to show some
The surface that was one protective layer away from the rebar was
lights on the mechanism of the chloride attack. The rest of the
reserved as the working surface. When the epoxy resin was com-
paper is organized as below: Section 2 introduces the raw materi-
pletely solidified, the composite system is put into the chloride
als and the sample preparations. The experimental program and
attack environment for the electrochemical performance tests.
the testing methods are illustrated in Section 3. Section 4 provides
In order to analyze the change of microstructure at the interface
with the experimental results and the discussion. The conclusions
of the UHPC repaired system, a sample with a volume of about 1 cc
are reached in the final section.
containing materials of two layers was extracted from the interface
of the bonded specimen, and the hydration was terminated with
anhydrous ethanol after the sample was taken out. The sample
2. Materials and sample preparation
was then studied using scanning electron microscope (SEM).

2.1. Materials
3. Experimental program and methods
The main materials used in this study are as follows. (1)
Cement: P.Ⅱ 52.5 cement, provided by Jiangnan Onoda Cement The above repaired systems were demoulded after one day of
Co. Ltd., its chemical compositions are shown in Table 1. (2) Silica casting, followed by a standard curing for 7 days. Subsequently,
fume: 970U silica fume from Topken enterprise with 96.2% SiO2 the experimental group was immersed in chloride attack environ-
content. (3) Quartz powder: mean diameter 12 lm, provided by ment for further curing, while the reference group was placed in
Shengkuo Building Materials Co. Ltd. (4) Fine aggregate: quartz the fresh water. The chloride attack environment in this paper
2
Q. Chen, R. Ma, H. Li et al. Construction and Building Materials 272 (2021) 121971

Table 1
Chemical compositions of cement (wt%).

Composition Na2O MgO Al2O3 SiO2 P2 O5 SO3 Cl K2O CaO TiO2 Cr2O3 MnO Fe2O3 CuO ZnO SrO
Content 0.08 0.65 4.56 20.90 0.12 2.65 0.05 0.87 65 0.22 0.01 0.09 3.23 0.01 0.05 0.03

Table 2
Properties of steel fibers.

Diameter(mm) Length(mm) Aspect ratio(L/D) Tensile strength(MPa) Elastic modulus(GPa) Density(g/cm3)

0.2 13 65 2850 200 7.85

Table 3
Mix ratio of materials (kg/m3).

Material Cement Silica Fume Quartz Powder Quartz Sand Sand Superplasticizer Water
UHPC 743 90 250 1070 7.15 193
Matrix 500 1500 0.6 220

Table 4 used to access the bonding and the flexural performance,

Compressive and flexural strength of materials.
which is shown in Fig. 1(b) and (c). The loading rate was
Material Fibers Flexural strength Compressive strength 0.08 mm/min. In the flexural tests, the mid-span deflection
(%) (MPa) (MPa) was read by an electronic extensometer and the load–deflec-
Matrix 0 9.5 34.6 tion curve was recorded automatically. The bonding flexural
UHPC 0 21.5 127.2 strength was determined by Eq. (1)
1 28.8 142.2
1.5 31.4 163.1
3
2 33.6 157.4 F ¼ 1:5P  L=b ¼ 2:34P ð1Þ
2.5 32.2 154.3
where L = 100 mm; b = 40 mm; P is the peak force; F is the
bonding flexural strength of the system.

was simulated with a saturated sodium chloride solution. (3) Electrochemical performance tests: CS350H type electro-
UHPC0 ~ UHPC2.5 is used to represent the repaired systems com- chemical workstation was used in electrochemical perfor-
posed of matrix materials and UHPC with 0 ~ 2.5% fiber volume mance tests. The three-electrode system is employed,
content. which is shown in Fig. 2. The working, reference and counter
The accessing methods for the effect of chloride attack on the electrode during the electrochemical test were the rebar, the
bonded concrete system include: saturated calomel electrode (SCE) and the stainless steel
plate respectively. In linear polarization part, the method
(1) Compressive strength tests: Bairoe electronic universal test- of dynamic potential polarization was adopted. The scan
ing machine was used to test the compressive strength and was performed in between 0.1 and + 0.1 V relative to open
the loading rate was 0.5 MPa/s. There were two compression circuit potential. The polarization resistance Rp was calcu-
modes, namely the compression direction was parallel to the lated according to the polarization curve.
bonding surface (abbreviated ask) and perpendicular to this (4) Microstructural analysis: German ZEISS195 ULTRA 55 Type
area (abbreviated as \), which is shown in Fig. 1(a). Field Emission Scanning Electron Microscope was used to
(2) Bonding/Flexural strength tests: The electronic universal analyze the microstructure and element distribution of the
testing machine controlled by SANS microcomputer was bonding area.

Fig. 1. Different bonded samples for the mechanical tests, where (a), (b) and (c) represent the specimens utilized to assess the system’s compressive, bonding and flexural
performance.

3
Q. Chen, R. Ma, H. Li et al. Construction and Building Materials 272 (2021) 121971

Fig. 2. Experimental program and methods Electrochemical test device.

4. Experimental results and discussion may come from two points. One is related to the restriction effect
of steel fibers on the dry shrinkage deformation [49] and the other
4.1. Compressive strength point is attributed to the enhancement of the mechanical occlusal
effect generated from the steel fibers gathered at the interface [50].
Fig. 3 (a) and Fig. 3(b) displays the changes of the compressive It can also be observed that the bonding performance of the system
strength of the repaired system at different ages (7d, 28d and 56d) was greatly affected by the chloride attack environment, and the
exposed to chloride attack. It can be found that the repaired system degradation degree gradually increased with time. When the
with UHPC material demonstrate higher compressive strength results considering the chloride attack after 56d are compared with
than that of the matrix material. Meanwhile, the presence of steel those of the reference group, the bonding strength of the system
fiber in the UHPC layer can greatly improve the compressive per- demonstrates decrements of 23.9%, 17.4%, 16.1%, 18.5% and 17.9%
formance of the repaired system. After being immersed in the for UHPC0, UHPC1, UHPC1.5, UHPC2 and UHPC2.5, respectively.
water or the NaCl solution, the specimen’s compressive strength The decrement percentage of each system seems to decrease
improves. The improvement can be attributed to the cement slightly with the addition of steel fibers, likely because the pres-
hydration in both layers of the bonded concrete system. Mean- ence of steel fibers from the interface can improve the compact-
while, the physical and mechano-chemical bond occurs in the ness of the transition zone [33].
interface between two connected layers may also have influenced
the strength enhancement [47–48]. In the early stage of chloride 4.3. Flexural behavior
attack (7d, 28d), the compressive strength of the system was basi-
cally the same as that of the reference group. At this time, the cor- 4.3.1. Failure mode
rosive solutes can enter the interior of the system and the Fig. 5 shows the typical failure mode of UHPC repaired systems
expansive product only plays a role of filling. This phenomenon exposed to the chloride attack. It can be noted that all the systems
agreed with the conclusions of Fan et al. [44] who conducted suffered a flexural failure. For the UHPC0 system, sudden brittle
experimental studies on the properties of concrete corroded by failure occurred. However, for the other three systems with steel
NaCl. With the extension of time, compared with the reference fibers, the micro-cracks appeared first in the UHPC layer in tensile
group, the compressive strength of the system in both compression zone. As the load increased, the micro-cracks expanded upwards,
modes showed a small decrease. The 56d compressive strength gradually reached the bonding surface and then penetrated the
decreased by about 5% to 10%. During this period, the expansive whole systems. The specimens with steel fibers could still bear
product continues to fill the initial void. Meanwhile, the expansion the load after cracking. In addition, it can be seen that the bonding
stress is generated, which will lead to the transverse failure of con- area of all repaired systems remained intact after flexure failure.
crete [33]. Therefore, the compressive strength of concrete will be No obvious horizontal slip and peeling occurred at the interface,
slightly reduced [33]. indicating that the bonding ability of the UHPC repaired system
in this paper could resist severe chloride attack environment.
Meanwhile, the steel fibers acting as ‘‘bridge” could be observed
4.2. Bonding strength
in the cracks [24].
Fig. 4 shows the development of the bonding strength of the
UHPC repaired system. Most of the specimens’ bonding strength 4.3.2. Flexural capacity
is greater than 4.5 MPa. The use of the steel fibers could greatly Table 5 lists the changes in the initial crack load (Pic) and the
improve the bonding strength of the system. Especially at the ini- maximum load (Pmax) of the UHPC repaired system with different
tial stage, when the volume content of the steel fiber is 2%, the steel fiber volume contents at 0 day (before attack), 28 days and
bonding strength could be increased by 50%. This enhancement 56 days. According to the results at 0 day, when the volume con-
4
Q. Chen, R. Ma, H. Li et al. Construction and Building Materials 272 (2021) 121971

Fig. 4. Development trend of bonding strength of the UHPC repaired system (r and s
mean the reference and experimental group; 0d, 7d, 28d and 56d are the chloride
attack periods).

improving the chemical resistance of the repaired system [33,53–

54].

4.3.3. Load-deflection response

Fig. 6 shows the load–deflection response of the UHPC repaired
systems subjected to the water immersion and the chloride envi-
ronment at 0d and 56d. The influence of the steel fibers on the flex-
ural performance of the systems can be seen from Fig. 6 (a)-Fig. 6
(c). With the addition of steel fibers, not only the flexural capacity
of the system is improved, but also the failure mode is changed.
The brittle failure mode is observed from the UHPC0 group without
steel fibers. For the other systems with the steel fibers, after the
initial crack occurs in the tension zone, the steel fibers could redis-
tribute the stress and the systems are still able to bear the load.
Later, as the crack continues to widen and expand upward, the sys-
tem reaches its ultimate flexural strength. At the same time, the
load gradually decreased and the mid-span deflection continued
Fig. 3. Development trend of compressive strength of the UHPC repaired system (r to increase. Fig. 6 (b) and (c) illustrates the load–deflection
means the reference group; s means the experimental group): (a) compressive
response of the systems in both water immersion and chloride
strength development (k); (b) compressive strength development (\).
environment after 56 days. Compared with those in the water,
the flexural capacities decreased slightly for the repaired systems
tents were 1%, 1.5%, 2% and 2.5%, its flexural capacity Pmax could immersed in the chloride environment. Meanwhile, the mid-span
increase by 30%, 65%, 66% and 95% compared with the system deflection corresponding to the maximum load also seemed to
without steel fiber. The amount of the steel fibers in the UHPC layer decrease to some extent after the chloride attack. The decreases
can significantly improve the flexural performance of the repaired of the maximum load after the chloride attack varied when differ-
system. This is mainly due to the excellent mechanical strength ent fiber contents are considered. The differences are 0.81KN,
and ductility of the steel fibers. When micro-cracks in UHPC 0.51KN,1.19KN,1.51KN and 1.17KN respectively for the repaired
expand under the action of external forces, the fiber bridging systems with UHPC0, UHPC1, UHPC1.5, UHPC2 and UHPC2.5.
effects will occur and the stress concentration at the crack tip is
relieved [51–52]. And the maximum load can reach 1.1 to 1.25 4.4. Electrochemical behavior
times of the initial crack load.
It can also be seen from Table 5 that compared with the refer- 4.4.1. Polarization curve analysis
ence group in the clean water, the flexural capacity of the repaired Fig. 7 (a) and (b) display the polarization curve of the steel bars
systems showed a small decrease after chloride attack. Taking the in the different repaired systems before and after the chloride
28d Pmax as an example, the decrements of 8.2%, 6.2%, 4.3%, 6.1% attack. From Fig. 7 (a), the polarization curves are relatively similar
and 7.8% was observed for UHPC0, UHPC1, UHPC1.5, UHPC2 and when different repaired systems are considered. They have a large
UHPC2.5 respectively. The deterioration degree appeared to be curvature radius for the anodic polarization curve and a small cur-
lower after the addition of steel fibers, which may be because the vature radius for the cathodic polarization curve, which is consis-
maximum stress under the flexural load is concentrated in the tent with the characteristics of the steel bar in the state of
UHPC layer. Therefore, the densification enhancement of UHPC passivation [55]. After 28 days of chloride exposure, the polariza-
with a certain amount of fibers may play an important role in tion curve of the MUHPC and UHPC systems did not change
5
Q. Chen, R. Ma, H. Li et al. Construction and Building Materials 272 (2021) 121971

Fig. 5. Failure modes of the UHPC repaired system.

Table 5
Initial cracking load and maximum load of the repaired system.

Fibers(%) Category 0d 28d 56d

Pic(KN) Pmax(KN) Pic(KN) Pmax(KN) Pic(KN) Pmax(KN)
0 Reference Group 4.33 6.89 7.72
Chloride Environment 6.37 6.91
1 Reference Group 4.49 5.65 6.52 7.66 6.99 8.16
Chloride Environment 6.14 7.21 6.79 7.65
1.5 Reference Group 6.67 7.15 6.43 8.88 7.69 10.02
Chloride Environment 5.67 8.51 6.70 8.83
2 Reference Group 6.71 7.21 7.05 9.05 6.97 10.82
Chloride Environment 6.02 8.53 6.56 9.31
2.5 Reference Group 7.90 8.46 7.89 9.18 10.47 11.31
Chloride Environment 7.50 8.51 9.39 10.14

much. The polarization curve of the MM system was changed and state of the steel bars at different stages, the linear fitting is per-
consistent with the shape of a typical corroded steel bar [55], indi- formed according to the polarization process. The changes of the
cating that after 28 days of chloride immersion, the steel reinforce- polarization resistance Rp at 0 ~ 28d are shown in Fig. 8. In the fol-
ment had been corroded. lowing text, Rp stands for the average value of 3 polarization resis-
The potential value corresponding to the horizontal segment of tance values tested at each age.
the polarization curve represents the free corrosion potential of the From Fig. 8, before chloride immersion, the polarization resis-
system, which is important for estimating the corrosion probability tances of MM, MUHPC and UHPC repaired systems were
of the steel bars. According to the corrosive principle, when the 3052 Xcm2, 3555 Xcm2 and 5748 Xcm2 respectively. When
free corrosion potential increases, it means that the corrosion is chloride attack began, different evolutions can be reached when
difficult to occur, and vice versa [55]. The standard of rebar corro- the three systems are considered. For the MM system, the polar-
sion by corrosion potential method is listed in Table 6 according to ization resistance showed a continuous decreasing trend. The
ASTM and Ministry of Metallurgy of China. From Fig. 7 (a), the free polarization resistances at 0d, 2d, 6d, 12d, and 28d were
corrosion potential of all the repaired systems before chloride 3052 Xcm2, 1731 Xcm2, 1390 Xcm2, 1124 Xcm2 and
attack is at the level of 200 mV, which could be preliminarily 1060 Xcm2 respectively. These results indicate that the chloride
determined from Table 6 that the steel reinforcement is in the pas- ions can easily reach the surface of the steel bar through the
sivation state at this time, with almost no possibility of corrosion. cement-based protective layer in the MM system, resulting in
After 28 days of the chloride exposure, the free corrosion potential the continuous corrosion of the reinforcement.
of MM system changed to 492 mV, which was lower than the Different from MM system, the Rp values of the MUHPC and
standard of 400 mV. For MUHPC and UHPC systems, their 28d UHPC systems continued to increase with an obvious rising segment
free corrosion potentials were 189 mV and 168 mV respec- at the beginning of chloride attack. The Rp values of the MUHPC
tively, both higher than 250 mV. These results indicate the steel system and the UHPC system at 6 days were 5792 Xcm2 and
bars in the UHPC repaired systems could withstand the effects of 12156 Xcm2 respectively, which increased by 63% and 111% com-
chlorine corrosion. pared with the initial period. Eventually, the Rp value of the
MUHPC system changes to 6298 Xcm2 at 28d, which is nearly 6
4.4.2. Polarization resistance analysis times the Rp value of the MM group. The 28d Rp value of the UHPC
According to the electrochemical theory, when the polarization system could reach to as high as 15056 Xcm2, almost 15 times of the
overpotential DE is very small, the current I and the overpotential one of the MM group. Meanwhile, it should be mentioned that the
DE can be considered to have a linear relationship. The slope of the pore size distribution and permeability of the material play impor-
polarization curve at DE = 0 is defined as the polarization tant roles in determining the evolution of the polarization resis-
resistance Rp. The larger the value of Rp, the harder it is for the tance. The chloride ions are difficult to attack the steel in UHPC
corrosion to occur [55]. In order to further analyze the corrosion due to their dense microstructures and low permeabilities.
6
Q. Chen, R. Ma, H. Li et al. Construction and Building Materials 272 (2021) 121971

Fig. 7. Polarization curve of the steel in the UHPC repaired system: (a) 0d, (b) 28d.

Table 6
Standard of rebar corrosion by corrosion potential method.

Standard Potential* /mV Criterion

>-200 5% corrosion probability
ASTM-C876[56] 200 ~ -350 50% corrosion probability
<-350 95% corrosion probability
>-250 No corrosion
Standard of the Ministry 250 ~ -400 Possible corrosion
of Metallurgy of China[57]
<-400 corrosion
*
The saturated copper–copper sulphate electrode is adopted in ASTM-C876 and
the saturated calomel electrode is employed in Standard of the Ministry of Metal-
lurgy of China.

Fig. 6. Load deflection response of the UHPC repaired system (r means the
4.5. Microstructural analysis
reference group, s means the experimental group, UHPC0 ~ 2.5 represents the
repaired systems using UHPC with 0 ~ 2.5% fiber volume content): (a) 0d, (b) 56d
(UHPC0, 1, 1.5 group), (c) 56d (UHPC2, 2.5 group), where ▲L represents the Considering that the bonding area between the UHPC and the
difference between the maximum load under the condition of attack environment matrix material was the weakest part of the whole system [58–
and the clean water curing. 59], SEM-EDS analysis are conducted on the bonding surface,

7
Q. Chen, R. Ma, H. Li et al. Construction and Building Materials 272 (2021) 121971

and no sudden change can be observed. These results also indicate

that compared with the ordinary cement-based material, UHPC can
significantly improve the resistance to chloride ions.
The difference of chloride contents in the matrix layer and the
UHPC layer can further explain the better performance of the UHPC
repaired system. According to Shen and Hussein [42,60], the inter-
facial chemical bonding will reduce after the chloride attack. The
crystals with no cementing effect will be formed [42,60]. Further-
more, the micropores at the interface gradually increase in the
chloride environment [61]. These microstructural evolutions can
show some lights to the deterioration mechanism of the mechan-
ical performance at the macroscale level.

5. Conclusions

This paper investigates the effects of the chloride attack on the

bonded system repaired by the UHPC material with different fiber
contents. The mechanical and electrochemical tests are conducted
to access the performance of the repaired system after the chloride
Fig. 8. Polarization resistance development of the UHPC repaired system.
attack. The SEM and EDS tests are performed to evaluate the
repaired system’s microstructures. Some conclusions can be
which is 5 mm away from the outer surface. Fig. 9 ~ 10 shows the reached as follows:
SEM micrograph and the EDS data at 28d of this area. From Fig. 9 (1) The chloride attack can degrade the mechanical perfor-
(a), there was an obvious transition zone between the substrate mance of the UHPC repaired system. After 56d chloride attack,
and the UHPC layer. In comparison, the substrate layer is relatively the repaired systems’ compressive and bonding strength decrease
loose while the UHPC layer is dense. It was also clear that the bond by about 5% to 10% and 16.1% to 23.9%, respectively. The decreases
of the repaired system was conserved despite the very severe chlo- of the maximum load after the chloride attack varies from 0.51KN
ride attack exposure. From Fig. 9 (b), the main elements in bonding to 1.51KN in the flexural test. The steel fibers can reduce the degra-
area were oxygen, silicon, carbon, calcium, etc. A small amount of dation degree of the repaired system’s mechanical performances in
chlorine element could also be observed, which proves that the chloride attack environment.
chloride ions in the solution are able to penetrate the cement- (2) Compared with the normal mortar repaired system, the
based layer and reach a certain depth inside the system. UHPC repaired system demonstrates better electrochemical per-
In order to investigate the distribution of chlorine elements formance and can resist the influence of severe chloride attack
near the bonding area inside UHPC repaired system, EDS line scan- environment. The polarization resistances and free corrosion
ning is conducted at the same depth of the system. Fig. 10 presents potential reduce to 1060 Xcm2 and 492 mV for the repairing sys-
the EDS line scanning results containing both the UHPC layer and tem with the normal mortar material, respectively. The values are
the matrix. The yellow line in Fig. 10 (a) is the scanning trace. It about 6298 Xcm2 (or 15056 Xcm2) and 189 mV (or 168 mV)
can be seen that the chloride contents vary greatly on both sides for the different UHPC repairing system.
of the interface. For the UHPC layer on the left, its chlorine content (3) There is an obvious transition zone between the matrix and
is low, with the average value only 2 to 3 counts. For the matrix the UHPC layer. The microstructures of the matrix layer are looser
material layer on the right, the chlorine content is much higher, than those of the UHPC layer. The bond of the repaired system was
with an average of 8 to 10counts. Within the bonding zone, the conserved despite the severe chloride attack exposure. Within the
chlorine contents are becoming higher from the UHPC side to the interfacial transition zone, the chloride content near the matrix
matrix side. Meanwhile, the increasing rate is statistically stable layer is about 4 times than that on the UHPC side.

Fig. 9. SEM micrograph and EDS data of the UHPC repaired system after exposure to chloride attack (arrows: bond area, box: EDS analysis area).

8
Q. Chen, R. Ma, H. Li et al. Construction and Building Materials 272 (2021) 121971

Fig. 10. EDS line scanning results of the UHPC repaired system after exposure to chloride attack (yellow line: scanning trace, green dotted line: average chlorine content in
matrix layer, blue dotted line: average chlorine content in UHPC layer). (For interpretation of the references to colour in this figure legend, the reader is referred to the web
version of this article.)

Author contribution [9] F.J. Alaee, B.L. Karihaloo, Retrofitting of reinforced concrete beams with
CARDIFRC, J. Compos. Constr. 7 (3) (2003) 174–186.
[10] Q. Wang, D.Q. Wang, S.Y. Zhuang, The soundness of steel slag with different
Qing Chen prepares and revises the manuscript. Rui Ma pre- free CaO and MgO contents, Constr. Build. Mater. 151 (1) (2017) 138–146.
pares the manuscript. Haoxin Li and Zhengwu Jiang contributes [11] W.T. Wang, K.Y. Zhao, P. Zhang, J.W. Bao, S.B. Xue, Application of three self-
developed ECT sensors for monitoring the moisture content in sand and
to the experimental design. Hehua Zhu and Zhiguo Yan contributes
mortar, Constr. Build. Mater. 121008 (2020).
the manuscript editing and experimental design. [12] P. Zhang, Y.Q. Dai, W.T. Wang, J.B. Yang, L.W. Mo, W.N. Guo, J.W., Effects of
magnesium oxide expansive agents on the self-healing performance of
microcracks in Strain-Hardening Cement-based Composites (SHCC), Mater.
Declaration of Competing Interest Today Comm. (2020).
[13] K. Habel, P. Gauvreau, Behavior of reinforced and posttensioned concrete
The authors declare that they have no known competing finan- members with a UHPFRC overlay under impact loading, J. Struct. Eng. 135 (3)
(2009) 292–300.
cial interests or personal relationships that could have appeared [14] J.W. Bao, S.G. Li, P. Zhang, S.B. Xue, Y.F. Cui, T.J. Zhao, Influence of exposure
to influence the work reported in this paper. environments and moisture content on water repellency of surface
impregnation of cement-based materials, J. Mater. Res. Technol. 9 (6) (2020)
12115–12125.
Acknowledgements [15] S.S. Qin, D.J. Zou, T.J. Liu, A. Jivkov, A chemo-transport-damage model for
concrete under external sulfate attack, Cem. Concr. Res. 132 (2020).
[16] Y.P. Tian, P. Zhang, K.Y. Zhao, Z.X. Du, T.J. Zhao, Application of Ag/AgCl Sensor
This work is supported by National key research and develop- for Chloride Monitoring of Mortar under Dry-Wet Cycles, Sensors 20 (5) (2020)
ment plan (2018YFC0705400). This work is also supported by the 1394.
Special Fund for Basic Research on Scientific Instruments of the [17] G. Martinola, M.F. Baeuml, A. Walliser, Numerical modeling of the long-term
behavior of passive fire protection mortars applied on concrete tunnels, J.
National Natural Science Foundation of China (4182780021), Mater. Civil Eng. 19 (6) (2007) 484–491.
National Natural Science Foundation of China (51878496, [18] T.J. Liu, H.N. Wei, A. Zhou, D.J. Zou, H.S. Jian, Multiscale investigation on tensile
51878479, 51878480, 51878481), the Funds of Fundamental properties of ultra-high performance concrete with silane coupling agent
modified steel fibers, Cem. Concr. Compos. 111 (2020).
Research Plan for the Central Universities. The authors declare that
[19] C. Xu, H.X. Li, X.J. Yang, B.Q. Dong, W.S. Wang, Action of the combined presence
they have no conflict of interest. of C-S-Hs-PCE and triethanolamine on the performances of cement
paste/mortar, Constr. Build. Mater. 121345 (2020).
[20] H.X. Li, Z. Xue, G.W. Liang, K. Wu, B.Q. Dong, W.S. Wang, Effect of C-S-Hs-PCE
References and sodium sulfate on the hydration kinetics and mechanical properties of
cement paste, Constr. Build. Mater. 266 (2021).
[1] K. Gu, B. Chen, Research on the incorporation of untreated flue gas [21] A.K. Pandey, Factors affecting bond between new and old concrete, ACI Mater.
desulfurization gypsum into magnesium oxysulfate cement, J. Clean. Prod. J. 109 (3) (2012) 389–390.
271 (2020). [22] L. Courard, T. Piotrowski, A. Garbacz, Near-to-surface properties affecting bond
[2] Y.T. Liu, B. Chen, Z.H. Qin, D. Pen, M.A. Haque, Experimental research on strength in concrete repair, Cem. Concr. Compos. 46 (2014) 73–80.
properties and microstructures of magnesium-iron phosphate cement, Constr. [23] Y. He, X. Zhang, R.D. Hooton, X.W. Zhang, Effects of interface roughness and
Build. Mater. 257 (2020) 119570. interface adhesion on new-to-old concrete bonding, Constr. Build. Mater. 151
[3] Z.Y. Mo, X.J. Gao, A.S. An, Mechanical performances and microstructures of (2017) 582–590.
metakaolin contained UHPC matrix under steam curing conditions, Constr. [24] Q. Chen, H. Wang, Z.W. Jiang, H.H. Zhu, J.W. Ju, Z.G. Yan, Reaction-degree-
Build. Mater. 268 (2021) 121112, https://doi.org/10.1016/j.conbuildmat.2020. based multi-scale predictions for the effective properties of ultra-high-
121112. performance concrete, Mag. Concr. Res. (2020), https://doi.org/10.1680/
[4] Y.X. Zhang, H.H. Zhu, Q.H. Guo, R. Carvel, Z.G. Yan, The effect of technical jmacr.19.00415.
installations on evacuation performance in urban road tunnel fires, Tunn. [25] H.X. Li, C. Xu, B.Q. Dong, Q. Chen, L.N. Gu, X.J. Yang, W.S. Wang, Differences
Undergr. Space Technol. 107 (2021). between their influences of TEA and TEAHCl on the properties of cement
[5] W. Meng, M. Valipour, K.H. Khayat, Optimization and performance of cost- paste, Constr. Build. Mater. 239 (2020).
effective ultra-high performance concrete, Mater. Struct. 50 (1) (2017) 29. [26] Q.J. Dai, D.H. He, F.C. Meng, P. Liu, X.K. Liu, Dielectric constant, dielectric loss
[6] D.Y. Yoo, S. Kim, M.J. Kim, Comparative shrinkage behavior of ultra-high- and thermal conductivity of Si3N4 ceramics by hot pressing with CeO2–MgO
performance fiber-reinforced concrete under ambient and heat curing as sintering aid, Mater. Sci. Semicon. Proce. 121 (2021).
conditions, Constr. Build. Mater. 162 (2018) 406–419. [27] E. Bonaldo, J.A.O. Barros, P.B. Lourenco, Bond characterization between
[7] D.Y. Yoo, M.J. Kim, High energy absorbent ultra-high-performance concrete concrete substrate and repairing SFRC using pull-off testing, Int. J. Adhes.
with hybrid steel and polyethylene fibers, Constr. Build. Mater. 209 (2019) Adhes. 25 (6) (2005) 463–474.
354–363. [28] A. Momayez, M.R. Ehsani, A.A. Ramezanianpour, H. Rajaie, Comparison of
[8] Q. Chen, H. Wang, H.X. Li, Z.W. Jiang, H.H. Zhu, J.W. Ju, Z.G. Yan, Multiscale methods for evaluating bond strength between concrete substrate and repair
modelling for the ultra-high performance concrete: From hydration kinetics to materials, Cem. Concr. Res. 35 (4) (2005) 748–757.
macroscopic elastic moduli, Constr. Build. Mater. 247 (2020).

9
Q. Chen, R. Ma, H. Li et al. Construction and Building Materials 272 (2021) 121971

[29] M.A.C. Munoz, D.K. Harris, T.M. Ahlborn, D.C. Froster, Bond Performance [46] B.C. Chen, T. Ji, Q.W. Huang, H.Z. Wu, Q.J. Ding, Y.W. Zhan, Review of Research
between Ultrahigh-Performance Concrete and Normal-Strength Concrete, J. on Ultra-high Performance Concrete, J. Archite. Civil Eng. 31 (3) (2014) 1–24.
Mater. Civil Eng. 26 (8) (2014) 04014031. [47] M.C.G. Juenger, P.J.M. Monteiro, E.M. Gartner, G.P. Denbeaux, A soft X-ray
[30] M.D. Cohen, A. Goldman, W.F. Chen, The role of silica fume in mortar: microscope investigation into the effects of calcium chloride on tricalcium
transition zone versus bulk paste modification, Cem. Concr. Res. 24 (1) (1994) silicate hydration, Cem. Concr. Res. 35 (1) (2005) 19–25.
95–98. [48] B.A. Tayeh, B.A. Bakar, M.M. Johari, Y.L. Voo, Evaluation of bond strength
[31] M. Kuroda, T. Watanabe, N. Terashi, Increase of bond strength at interfacial between normal concrete substrate and ultra-high-performance fiber concrete
transition zone by the use of fly ash, Cem. Concr. Res. 30 (2) (2000) 253–258. as a repair material, Proc. Eng. 54 (2013) 554–563.
[32] M. Farzad, M. Shafieifar, A. Azizinamini, Experimental and numerical study on [49] A.M. Rashad, Effect of steel fibers on geopolymer properties – The best
bond strength between conventional concrete and Ultra High-Performance synopsis for civil engineer, Constr. Build. Mater. 246 (2020) 118534.
Concrete (UHPC), Eng. Struct. 186 (2019) 297–305. [50] A. Khabaz, Performance evaluation of corrugated steel fiber in cementitious
[33] L. Hussein, L. Amleh, H. Siad, M. Lachemi, Effect of very severe sulfate matrix, Constr. Build. Mater. 128 (2016) 373–383.
environment on bonded composite concrete system, Constr. Build. Mater. 191 [51] Y. Zhang, J.W. Ju, H.H. Zhu, Z.G. Yan, A novel multi-scale model for predicting
(2018) 752–763. the thermal damage of hybrid fiber-reinforced concrete, Int. J. Dam. Mech. 29
[34] M.J. Chajes, T.A. Thomson, C.A. Farschman, Durability of concrete beams (1) (2020) 19–44.
externally reinforced with composite fabrics, Constr. Build. Mater. 9 (3) (1995) [52] T. Zhang, Y. Zhang, H.H. Zhu, Z.G. Yan, Experimental investigation and multi-
141–148. level modeling of the effective thermal conductivity of hybrid micro-fiber
[35] H.A. Toutanji, W. Gomez, Durability characteristics of concrete beams reinforced cementitious composites at elevated temperatures, Compos. Struct.
externally bonded with FRP composite sheets, Cem. Concr. Compos. 19 (4) 256 (2021).
(1997) 351–358. [53] A.S. El-Dieb, Mechanical, durability and microstructural characteristics of
[36] A. Katz, N. Berman, Modeling the effect of high temperature on the bond of FRP ultra-high-strength self-compacting concrete incorporating steel fibers, Mater.
reinforcing bars to concrete, Cem. Concr. Compos. 22 (2000) 433–443. Des. 30 (2009) 4286–4292.
[37] H.V.S. GangaRao, J. Barger, Aging of bond between FRP and concrete cubes, Int. [54] D. Wang, C. Shi, Z. Wu, J. Xiao, Z. Huang, Z. Fang, A review on ultra high
J. Mater. Prod. Tec. 19 (1–2) (2003) 83–95. performance concrete: Part II. Hydration, microstructure and properties,
[38] J. Wang, B.G. Ma, H.B. Tan, C. Du, Z.G. Chu, Z.T. Luo, P.G. Wang, Hydration and Constr. Build. Mater. 96 (2015) 368–377.
mechanical properties of cement-marble powder system incorporating [55] C.N. Cao, Principle of corrosion electrochemistry, Chemical Industry Press,
triisopropanolamine, Constr. Build. Mater. 266 (2021). Beijing, China, 2008.
[39] J.J. Zhang, H.B. Tan, M. Bao, X.H. Liu, Z.T. Luo, P.G. Wang, Low carbon [56] ASTM C876-09. Standard test method for corrosion potentials of uncoated
cementitious materials: Sodium sulfate activated ultra-fine slag/fly ash blends reinforcing steel in concrete. US: ASTM, 2009.
at ambient temperature, J. Clean. Prod. 280 (2021). [57] J.S. Shen, J.Z. Liu, J.H. Mao, W.L. Jin, Y.D. Xu, Electrochemical Characteristics of
[40] L. Qin, X.J. Gao, A.S. Su, Q.Y. Li, Effect of carbonation curing on sulfate resistance Steel Reinforced Concrete at Different Deterioration Stages, J. Build. Mater. 23
of cement-coal gangue paste, J. Clean. Prod. 278 (2021). (4) (2020) 963–968.
[41] Q. Wu, H. Xie, Y. Xun, Mechanical property degradation of TRC composite [58] G. Li, H. Xie, G. Xiong, Transition zone studies of new-to-old concrete with
beam under coupling of stress and chloride, J. Jiangsu Univ. Sci. Technol. (Nat. different binders, Cem. Concr. Compos. 23 (4–5) (2001) 381–387.
Sci. Ed.) 32 (2) (2018) 291–297. [59] C. Zanotti, N. Banthia, G. Plizzari, A study of some factors affecting bond in
[42] L. Hussein, L. Amleh, H. Siad, M. Lachemi, Effect of severe chloride environment cementitious fiber reinforced repairs, Cem. Concr. Res. 63 (2014) 117–126.
on the flexural behaviour of hybrid concrete systems, Mag. Concr. Res. 72 (15) [60] Y. Shen, The effect of chloride corrosion on interfacial bonding properties of
(2020) 757–767. GFRP cement-based materials (Master’s thesis), Hunan University of Science
[43] H.G. Midgley, J.M. Illston, The penetration of chlorides into hardened cement and Technology, Xiangtan, China, 2018.
pastes, Cem. Concr. Res. 14 (4) (1984) 546–558. [61] S.P. Yin, L. Zhao, D.F. Qiang, P.H. Li, Interfacial properties between textile
[44] Y.F. Fan, Y.Z. Zhang, Y.D. Hu, J.L. Liu, Experimental study on property of reinforced concrete (TRC) and old concrete under chloride wet-dry
concrete corroded by NaCl, J. Dalian Maritime Univ. 34 (1) (2008) 125–128. environment, Mater. Rep. 31 (2017) 346–350.
[45] Y.B. Yin, Study on the durability of CFRP reinforced concrete structure under
chloride environment (Master’s thesis), Qingdao Technological University,
Qingdao, China, 2014.

10