Construction and Building Materials 173 (2018) 1–9

Contents lists available at ScienceDirect

Construction and Building Materials

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

Assessment of self-sensing capability of Carbon Black Engineered

Cementitious Composites
Hanwen Deng a,⇑, Hongliang Li b
a
School of Transportation, Southeast University, Sipailou, Nanjing 210096, PR China
b
Department of Civil Engineering, Tianjin University, Tianjin 300072, PR China

h i g h l i g h t s

Incorporating CB particles can decrease the bulk resistivity of ECC significantly.

Incorporating CB particles and AEA can enhance the tensile strain capacity of ECC significantly.

CB-ECC can be potentially utilized to detect micro cracking at the early inelastic stage.

A new method of nondestructive testing was proposed in this investigation.

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

Article history: The protection and health monitoring of degraded concrete infrastructure required a multifunctional
Received 21 January 2018 material which possessing good damage tolerance, whilst providing a self-sensing capacity that designed
Received in revised form 26 March 2018 to specifically diagnose cracking. Engineered Cementitious Composites (ECC) presents superb tensile duc-
Accepted 4 April 2018
tility and pseudo strain-hardening property. In this paper, a type of multifunctional ECC incorporating
Available online 6 April 2018
Carbon Black (CB) nano-particles and Air Entraining Agent (AEA) to decrease the bulk resistivity while
increasing the tensile strain capacity is studied. The effect of CB particles on the electrical response of
Keywords:
ECC, HFA-ECC (high fly ash ECC) and CB-ECC under direct tension was investigated experimentally by
Engineered Cementitious Composites (ECC)
Self-sensing capacity
four-point probe test. The experimental results showed that the bulk resistivity of all specimens
Gauge Factor (GF) increased with the crack propagation, and the increase ratio in inelastic phase was much higher than that
Carbon Black (CB) in elastic phase. The Gauge Factor (GF) in the strain-hardening stage was calculated by using the change
Resistivity in bulk resistivity and tensile strain, and the relationship between GF and crack width was also investi-
gated. The combination of CB and AEA obtained the highest sensing ability from elastic stage to first
cracking stage, which can determine whether the internal microcracks are generated by the change of
the bulk resistivity.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction Engineered Cementitious Composites (ECC), a special type of

High Performance Fiber Reinforced Cementitious Composites
Piezoresistive property of cement-based materials was already (HPFRCC), has been researched widely since it was developed by
discovered and applied in monitoring traffic in highway and struc- Li and his teams [2] in the 1990s. ECC possesses an extreme tensile
tural health in many years ago. The change in resistivity of con- ductility, in the range of 3–5% (about 300–500 times that of con-
crete itself can be used to sense the strain/stress and damage, crete), and obtains the multiple cracking behavior and keep the
which is much more important for engineers and researchers to crack width constant at 60 lm while the number of cracks
monitor concrete structure timely and thereby enhance the service increase. Therefore, it is feasible to investigate the self-sensing
life [1]. However, the research of self-sensing concrete is mainly properties of ECC due to its excellent mechanical performance
restricted in compressive strain and stress, this is because that brit- especially in tension. ECC, as a sensor, has many advantages com-
tleness of concrete limits the investigation into self-sensing prop- pared with embedded/attached strain/stress sensors. For example,
erties under tension loading. (1) it could detect a wider strain range, including the whole tension
stage; (2) it requires no extra sensors to be installed; (3) it could
⇑ Corresponding author. monitor the integrated structure, not just its specific parts over
E-mail address: [email protected] (H. Deng).
the whole service life [3].

https://doi.org/10.1016/j.conbuildmat.2018.04.031
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
2 H. Deng, H. Li / Construction and Building Materials 173 (2018) 1–9

The mechanism of electric conduction in cement-based materi- ECC in this paper. When the mixture is stirred, in order to obtain good flow ability
of fresh matrix, more water and super plasticizer are added since CB has large speci-
als is that the movement of positive and negative ions was con-
fic surface area, 254 m2/kg [12]. So the water to cement ratio of CB-ECC in Table 1 is
nected to micro-pores and promote the electrical conduction of 0.31 not 0.27. Air Entraining Agent (AEA), main component is Sodium Lauryl Sulfate,
cementitious materials [4,5]. The bulk resistivity of cementitious was added into matrix in order to form more air bubbles. Portland Ordinary
materials is effected by many factors, including different cement, Cement, Type I, was partially replaced by fly ash (Class I) to develop high-tensile
ratio of water to binder, hydration degree and porosity. Li et al. ductility of ECC. The size of silica sand used in this paper ranged from 70 to 140
meshes (109–212 mm). PVA fiber, with a surface oil coating of 1.2% by weight,
[6] demonstrated that the resistivity of wet concrete ranged from
was produced by Kuraray Company in Japan. The physical and mechanical proper-
103 to 104 O-cm while that of dry concrete could increase to ties of PVA fiber is listed in Table 2.
108–109 O-cm, this was because that the resistivity of hydration
product was much higher than that of pore solution. The large dif- 2.2. Specimen preparation and experimental tests
ference in resistivity makes more difficult to use electrically con-
ductive performance of cementitious material in structural health Firstly, all solid materials, including cement, fly ash, silica sand, AEA and other
solid additives, were mixed according to the proportions shown in Table 1 and stir-
monitoring.
red for 3 min at low speed to ensure that mixtures were well stirred. Then, slowly
Some conductive materials, such as steel fiber, graphite, and adding water and super plasticizer at firstly low-speed stirring for 3 min and then at
carbon nanotube, was extra added into cementitious material to high-speed stirring for 5 min. When the slurry was in a fluid state, the PVA fibers
enhance its conductive ability and simultaneously reduce depen- were slowly added, stirred for 5 min at low speed and then stirred at high speed
dence on ions inside pore solution. Carbon Black (CB), character- for 6 min. when CB was added, the stirring time would be prolonged. Finally, when
the fibers were dispersed evenly and without agglomeration, and the slurry had
ized by desirable electrical conductivity and readily dispersible
good workability, the slurry was poured into the prepared molds, slightly shaken
ability in cement paste, was used to improve electrical conductiv- and smoothed in a timely manner. The specimens were demolded after 24 h, and
ity of cementitious materials. Therefore, it is feasible to produce then cured under standard condition at 95 ± 5% RH and temperature of 20 ± 2 °C
ECC material modified with CB, while simultaneously retaining until tested.
The copper tape, produced by 3J Company in China, was used as the electrode.
higher electrical conductivity.
The surface of specimens should be smooth and clean in order to paste the elec-
For self-sensing properties of ECC, the relationship of change trodes. The copper tape was cut into the electrode strip with a width of 4 mm
between resistivity and strain during loading test has been suc- and a suitable length. The silver conductive epoxy, produced by Xinwei New Mate-
cessfully developed by a large number of researchers [7–11]. Han rials Co. Ltd in China, was smeared in the electrode position around shown in Fig. 2,
and Ou [10] analyzed a number of parameters to evaluate the sens- where will paste the electrode strip. The silver conductive epoxy can enhance the
conductivity between electrode and matrix. The pasted specimen was placed at
ing properties of Piezoresistive Cement-based Stress/Strain (PCSS),
room temperature for 24 h.
including the range of input/output, linearity, repeatability, hys- The size of specimen and the location of electrode tape are shown in Fig. 1. Sec-
teresis, Signal to Noise Ratio (SNR) and zero drift. For tensile strain tion 1 is located between two electrode strips connected to the voltmeter with a
monitoring, Ranade et al. [7] defined Gauge Factor (GF) as the rel- distance of 30 mm. Section 2 is located between one electrode connected voltmeter
and one electrode connected ammeter with a distance of 17 mm. Section 3 is out-
ative change in resistivity per unit strain, and characterized it for
side the external electrode strip position. The distance between two electrodes con-
the piezoelectric resistivity of the ECC material. Due to the excel- nected to the power source is 64 mm. The width and thickness of Sections 1 and 2 is
lent strain-hardening behavior and deformation ability of ECC, 30 mm and 13 mm respectively. In addition, the distance between two LVDT (Linear
the uniaxial tensile test and bending test were both used to evalu- Variable Differential Transformer) is 80 mm.
ate the self-sensing properties of ECC [1,12,13]. Moreover, GF can During the uniaxial tensile test, four-point probe method was used to measure
the resistivity of specimen, the polarization and contact impedance could be mini-
be used to effectively monitor the progress of strain and resistivity
mized by the change in electrical potential [9,14,15]. The resistivity measurements
changes at different loading stages with specific values [12,13]. were carried out using NI-cDAQ-9174 with NI 9203 and NI 9229 separately con-
Ranade et al. [7] studied the effect of crack pattern on self- nected ammeters and voltmeter. An arbitrary/Waveform Generator was added into
sensing properties of ECC and predicted the relationship between the circuit in series to offer a constant frequency alternating electrical potential. The
frequency range from 1 Hz to 10 kHz can minimize the effect on capacitance of ECC
single crack behavior and crack width by using two-point probe
[14]. A constant AC voltage with amplitude of 5 V and frequency of 1 kHz was
method with alternating current (AC). However, adding CB into applied across the outer two copper tape electrodes, and the current produced
ECC material can increase the sensing ability but reduce its ductil- across the same electrodes was measured by NI 9203 every 1.6 s. The resistance
ity capacity. Li et al. [8] found that increasing the CB content could and bulk resistivity of ECC could be calculated based on ohm’s law, shown in for-
reduce the bulk resistivity and tensile strain capacity of ECC mainly mula (1) and (2).

since the addition of CB leaded to an uneven distribution of fibers. V

R¼ ð1Þ
Therefore, in order to achieve a good self-sensing performance, the I
nano Carbon Black and Air Entraining Agent (AEA) were used to where V and I are the voltage and current measured by voltmeter and ammeter
add into ECC mixture to decrease the bulk resistivity while increas- respectively.
ing tensile strain capacity in this study. The systematic research on
RS
means of measurement and error was done, and four-point probe q¼ ð2Þ
L
method under above 1 kHz frequency AC supply was recom-
where q and R are the bulk resistivity and resistance; S and L represent area of cross
mended to minimize the error produced by material polarization
section and the distance between electrodes respectively.
and contact impedance [9]. The mechanical property of ECC mixtures under uniaxial tension was investi-
In this study, four-point probe method with high frequency AC gated in this study. All specimens were tested at the age of 28 days. Four specimens
was used to evaluate the self-sensing properties of CB-ECC during for ECC, HFA-ECC and CB-ECC were prepared for tension test. The geometry of spec-
tensile loading. The mechanical properties of ECC, including first imen and the tensile setup are shown in Figs. 1 and 2.

cracking strength, ultimate tensile strength, ultimate tensile strain,

crack width, crack number and final fracture location, were 3. Results and discussions
investigated.
3.1. Mechanical behavior
2. Experimental program
The tensile properties of ECC, HFA-ECC and CB-ECC are shown in
2.1. Materials and mixture proportions Fig. 3. As shown in Fig. 4, all specimens showed pseudo strain-
The mix proportions are presented in Table 1, where water, silica sand and CB
hardening behavior and multiple cracking under tension test.
are the ratio to binder (cement and fly ash), while fly ash is the ratio to cement by Table 3 summarized the tensile properties obtained from the
weight. Carbon Black-VXC72R from Cabot Corporation was chosen to develop CB- experimental tests. The value in Table 3 was the average of three
 H. Deng, H. Li / Construction and Building Materials 173 (2018) 1–9 3

Table 1
Mix design of self-sensing ECC (weight ratios).

Mix No Cement Fly ash Silica sand Water Super Plasticizer (g/L) AEA (g/L) Carbon Black PVA fiber (vol.%)
ECC 1.0 1.5 0.36 0.27 3 0.1 0 2
HFA-ECC 1.0 3.0 0.36 0.27 3 0.1 0 2
CB-ECC 1.0 1.5 0.36 0.31 6 0.1 0.01 2

Table 2
Physical and mechanical properties of PVA fiber.

Diameter (lm) Length (mm) Elongation (%) Density (g/cm3) Elastic modulus (GPa) Tenacity (MPa)
REC 15 39 12 7 1.3 42.8 1620

Fig. 1. Schematic figure of dogbone-shaped specimen pasted with electrodes and

the position fixed with LVDT.

tested specimens. In the tensile stress–strain curves, the first

cracking strength is defined as the tensile stress corresponding
wo the end point of the linear stage, while maximum tensile stress Fig. 2. Uniaxial tension test setup.
is defined as the tensile strength, and the corresponding strain is
defined as the tensile strain capacity [16].
The average ultimate tensile stress/strain of ECC was 5.69 it was noted that the spherical shape of fly ash causes the unifica-
MPa/3.84% with COV (Coefficient of Variation, COV = Standard tion and escape of entrained air bubbles. This is known to reduce
deviation/Average value) of 6%/23%. Meanwhile, the tensile strain the stability of entrained air content in SCC (self-compacting con-
capacity of HFA-ECC and CB-ECC was 4.51% and 4.02% respectively crete) [17,18]. Meanwhile, adding some certain content AEA into
with COV of 2% and 9%. However, the ultimate tensile strength of HFA-ECC leads to large stability of fine air bubbles. Puthipad
CB-ECC was lower than that of ECC, achieving only 4.26 MPa with et al. [19] analyze that the ball-bearing of entrained air bubbles
COV of 13%. From Fig. 3 and Table 3, it can be seen that the tensile significantly enhanced the self-compactability of fresh concrete
strain capacity of ECC increase obviously as fly ash content (FA/C) with high volume fly ash when using a certain type of AEA. There-
increase from 1.5 to 3.0, while the first cracking strength decrease fore, the addition of AEA can promote the self-compactability of
with fly ash content. Table 3 shows the cracking characteristics of HFA-ECC, which maybe enhance the effect of chemical bond and
three ECC mixtures. It indicates that when fly ash to cement ratio increased frictional bond strength in fiber/matrix, thereby result-
increase from 1.5 to 3.0, the crack number increases along the ing in the increase of the fiber bridging capacity in HFA-ECC.
length of specimen and the crack width becomes tighter, which When adding 1% of CB content into mixtures, it can be seen
is beneficial to the durability and self-sensing behavior. from Fig. 3 and Table 3 that the first cracking strength and ultimate
However, the ultimate tensile strength of HFA-ECC was bigger tensile strength of CB-ECC has a significantly decrease due to the
than ordinary ECC. In the previous paper [17], fly ash was found higher water to cement ratio, while the tensile strain capacity is
to increase the amount of large entrained air bubbles. Furthermore, enhanced compare with ECC. Based on test results, the addition
4 H. Deng, H. Li / Construction and Building Materials 173 (2018) 1–9

6
Tensile stress (MPa)

5 Final crack
4
location

2 ECC
HFA-ECC
CB-ECC (a)
1

0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5
Tensile strain (%)
Final crack
Fig. 3. Tensile behavior of ECC, HFA-ECC and CB-ECC.
location

of CB together with AEA into mixture can promote to achieve the

high tensile strain capacity and multiple-cracking behavior. Air
bubbles inside fresh mixture could be generated by hydrophobic
ends of AEA orientating towards gaseous phase and hydrophilic
ends towards aqueous phase [20]. When adding a certain dosage (b)
of AEA, a number of small air voids coalesced to form larger size
air bubbles [21,22]. The presence of Carbon Black, which is
hydrophobic nano-particles, improved the water contact angle
[23], so the stabilized air bubbles inside matrix were formed.
And some specific size of air bubbles could be considered as artifi-
cial flaw to initiate cracking [24]. Therefore, the combination of
Final crack
Carbon Black and AEA, coupled with a suitable water to cement
ratio, leads a uniform state in ECC paste during mixing and is a location
new method to enhance the strain-hardening and multiple crack-
ing ability of CB-ECC.

3.2. Crack pattern (c)

Fig. 4. Photographs of specimens showing crack patterns in: (a) ECC, (b) HFA-ECC
Due to the limited size of the specimen, only the resistivity and (c) CB-ECC.
change of Section 1 can be observed. Therefore, the crack width
and number of cracks in Section 1 can be observed in this study.
The typical broken specimen under uniaxial loadings is shown in
Fig. 4. The final crack location occurs in Section 1 or Section 3. As one obtained from 0.5%CB-ECC to 1%CB-ECC blackened. The fiber
shown in Table 3 and Fig. 4, the ultimate tensile strain of CB-ECC obtained from 1%CB-ECC was the darkest since more CB particles
is up to 4% and specimen finally cracks within the Section 1. The were absorbed by the PVA fiber surface. The narrowed crack open-
CB-ECC specimen both processes strain-hardening behavior and ing after incorporating CB maybe contribute to the fact that much
multiple tight cracks, overcoming the lower tensile ductility of CB particles actively attached to the surface of fiber, which further
CB-ECC in previous paper [8]. Simultaneously, specimens with 1% enhance the frictional bond strength between fiber and matrix, and
CB content (CB to cement ratio is 0.01) are darker colored, as would eventually narrowed the crack width [8]. So the number of cracks
be expected. A large number of fine pores was found in the surface of CB-ECC is higher than ordinary ECC and crack width is lower.
of CB-ECC specimen. It can be explained that the addition of AEA Meanwhile, the crack width of HFA-ECC is the least in all mix-
leads to the occurrence of increased air bubbles in matrix. The tures. In previous paper, Wang & Li [25] and Yang et al.’s research
residual crack width distribution and lognormal best-fit curves [26] on micro-scale investigation of ECC shows that increasing fly
are shown in Fig. 5. The crack width of ECC, HFA-ECC and CB-ECC ash content can decrease the chemical bonding effect, while
is mainly distributed in 45 lm, 15 lm and 23 lm respectively. enhance the frictional force between fiber and matrix. The influ-
The achievement of strain-hardening and multiple-cracking ence on fiber/matrix interface is beneficial to obtain good strain-
behavior of CB-ECC can be contributed to the combination of Car- hardening behavior of HFA-ECC, thus enhances the tensile strain
bon Black and AEA, which makes ECC matrix denser than before. capacity and crack number, while decreases the crack width. It
Fig. 6 shows microscopic observation of PVA fibers from ECC, indicates that HFA-ECC has better crack self-controlled capability.
0.5%CB-ECC and 1.0%CB-ECC mixtures. After mixing, a single fiber
was taken out of the fresh mixtures and cleaned by running water 3.3. Electrical behavior
until no matrix on the fiber surface. Then the fibers obtained from
ECC, 0.5%CB-ECC and 1%CB-ECC were observed by optical micro- According to previous paper [7–9], Gauge Factor (GF) can be
scope at a magnification of 10 (seen in Fig. 6). It can be seen from used to evaluate the self-sensing properties of concrete materials,
Fig. 6 that the fiber taken from ECC was nearly transparent while the formula is as follows:
 H. Deng, H. Li / Construction and Building Materials 173 (2018) 1–9 5

Table 3
The tensile properties of ECC, HFA-ECC and CB-ECC.

Mix No. First cracking Elastic strain Ultimate tensile Ultimate tensile Final crack Crack Average crack
strength (MPa) (%) strength (MPa) strain (%) location number* width* (lm)
ECC 1 4.62 0.012 5.47 3.15 Sec I 12 47
2 4.41 0.012 5.52 3.52 Sec I 14 50
3 4.80 0.006 6.08 4.85 Sec III 16 38
mean 4.61 ± 0.20 0.010 5.69 ± 0.34 3.84 ± 0.89 – 14 44
HFA-ECC 1 3.40 0.002 5.82 4.60 Sec III 38 13
2 5.15 0.008 6.58 4.41 Sec III 34 15
3 3.47 0.005 4.95 4.53 Sec III 42 14
mean 4.00 ± 0.99 0.005 5.78 ± 0.82 4.51 ± 0.10 – 38 14
CB-ECC 1 3.20 0.014 3.96 3.58 Sec I 26 26
2 2.98 0.016 3.92 4.23 Sec I 31 22
3 3.65 0.013 4.90 4.26 Sec III 29 19
mean 3.28 ± 0.34 0.014 4.26 ± 0.55 4.02 ± 0.38 – 29 22
*
Cracks occurred in Section 1, excluding Sections 2 and 3.

3.5 25
ECC HFA-ECC
3.0
lognormal best-fit lognormal best-fit
20
2.5
Crack number

2.0 Crack number 15

1.5
10

1.0
5
0.5

0.0 0
0 10 20 30 40 50 60 70 80 90 100 110 0 5 10 15 20 25 30 35 40 45 50 55 60
Crack width/µm Crack width/µm
(a) (b)

18

16 CB-ECC
lognormal best-fit
14

12
Crack number

10

0
0 10 20 30 40 50 60 70 80 90 100
Crack width/µm
(c)
Fig. 5. Residual crack width distribution and lognormal best-fit curves: (a) ECC; (b) HFA-ECC; (c) CB-ECC.

Dq=q0 Due to the different microstructures and compositions of three

GF ¼ ð3Þ
De mixtures, the initial bulk resistivity of ECC, HFA-ECC and CB-ECC at
the initial strain (e ¼ 0) was greatly different. Meanwhile, environ-
mental factors, e.g., temperature and humidity, will also effect the
where Dq is the change in bulk resistivity during the corresponding
change of q0 . Thereby, the self-sensing test were tested at room
strain range; q0 is the initial bulk resistivity corresponding strain is
temperature (60 ± 5% RH and temperature of 20 ± 2 °C), q0 can be
zero; De is the change in tensile strain.
6 H. Deng, H. Li / Construction and Building Materials 173 (2018) 1–9

7 300
ECC

Relative change in resistivity (%)

6

Tensile stress (MPa)

5
200
4
0.5%CB-ECC
3
100
2
1%CB-ECC ECC-1
ECC-2
1 ECC-3

0 0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
Fig. 6. Microscopic observation of PVA fibers from ECC, 0.5%CB-ECC and 1.0%CB- Tensile strain (%)
ECC mixtures.
(a)

maintained at a constant value. The initial bulk resistivity of ECC, 7 100

HFA-ECC and CB-ECC at the initial strain (e ¼ 0) was shown in

Relative change in resistivity (%)

6
Fig. 7. The initial bulk resistivity of HFA-ECC was the largest than 80
others, reached to 34,542 O-cm, with COV of 5%, about 1.7 times

Tensile stress (MPa)

5
of ECC. This is because of the high porosity of HFA-ECC. Although
the use of a small number of fly ash instead of cement can signif- 60
4
icantly reduce the porosity of the concrete, but the use of large
number of fly ash (e.g., ECC) can also increases the porosity of 3
40
the concrete [27]. At the same time, the use of higher fly ash con- HFA-ECC-1
tent can also reduce the pore size in concrete [27], resulting in a 2 HFA-ECC-2
significant decrease in the propagation rate of conductive ions, so HFA-ECC-3 20
1
the initial bulk resistivity of HFA-ECC is much higher than that of
ECC. The initial bulk resistivity of CB-ECC is the smallest, only
0 0
9734 O-cm, with COV of 4%, about one half of ECC. This indicates 0 1 2 3 4 5 6
that the addition of nanoscale CB particles significantly improves
Tensile strain (%)
the conductivity of ECC, including ECC matrix and fiber surface,
which is beneficial to the research of self-sensing properties. (b)
Fig. 8 shows the mechanical and electrical behaviors of ECC,
7 400
HFA-ECC and CB-ECC. When the ultimate tensile strain capacity

Relative change in resistivity (%)

is reached, the relative change in resistivity of ECC, HFA-ECC and 6
CB-ECC is about 250%, 90% and 370% respectively, exhibited an
300
Tensile stress (MPa)

ascending tendency resulted from tensile loading. It indicated that 5

the electrical conduction of three mixtures was weaken during
loading until specimen broken, thereby resulting in a suddenly 4
increase of resistivity, especially in CB-ECC. The relative change 200
rate in resistivity of CB-ECC in the inelastic phase is much higher 3
than that in elastic phase. As shown in Table 3, the average number
2 CB-ECC-1
of cracks increased from 14 for ECC to 29 for CB-ECC, the average 100
CB-ECC-2
crack width was also reduced from 44 lm for ECC to 22 lm for CB-ECC-3
1

0 0
40000
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0
The initial bulk resistivity (ohm-cm)

35000 Tensile strain (%)

(c)
30000
Fig. 8. Mechanical and electrical behaviors of: (a) ECC, (b) HFA-ECC and (c) CB-ECC.
25000

20000 CB-ECC. This indicates that the generation of microcracks partially

blocks the conductive matrix of ECC, resulting in the increase of
15000
relative change in resistivity. During the inelastic phase, with the
10000 continuous occurrence of multiple cracks, the relative change rate
in resistivity is basically maintained in a high range. This phe-
5000 nomenon provides a possibility for ECC materials to sense the
occurrence of cracks.
0
The tested ECC specimens show a significant piezoresistive
ECC HFA-ECC CB-ECC
behavior with a strong correlation between resistivity and the
Fig. 7. The comparison of the initial bulk resistivity of ECC, HFA-ECC and CB-ECC. increasing tensile strain. It can be seen from Fig. 8 that the resistiv-
 H. Deng, H. Li / Construction and Building Materials 173 (2018) 1–9 7

ity of all specimens is growing slowly during the elastic phase due 350
to the presence of piezoresistive effect of ECC matrix. However, the ECC-1
increase in resistivity is significantly accelerated by the generation 300 ECC-2
ECC-3
of multiple cracks, resulting in the formation of physical barrier of
the ion conductive path. GF is used to the sensing properties of 250
tested specimens. There are three specimens in every mixture.

GF value
During strain-hardening phase, the fractional GF values at every 200
0.5% of tensile strain are calculated and presented in Fig. 9. The
GF values from elastic phase to 0.5% of tensile strain changes from 150
97.9 to 53.0. GF of ECC-1 reached to 97.9 due to the smallest initial
resistivity, only 19,672 O-cm. Because the initial resistivity of HFA- 100
ECC is the biggest than others, so the corresponding GF value (ee <
e < 0.5%) varies from 15.8 to 51.7, lower than that of the others. The 50
GF values of two CB-ECC specimens (ee < e < 0.5%) is more than 130
and last one is 103.4, because the initial resistivity of CB-ECC-3 is 0
the biggest at the initial strain (e ¼ 0). 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
The GF value of ECC specimen fluctuates in the whole strain- Tensile strain (%)
hardening stage, and the fluctuation is the largest in the stain range (a)
of 0.5% < e < 1.0% and 3.5% < e < 4.0%. The GF value of ECC2 speci-
men is only single digit during strain range of 0.5% < e < 1.5%. But 350
the GF in strain range of 3.5% < e < 3.52% reached to 324.2, this is
HFA-ECC-1
because the final strain capacity is reached and specimen has bro- 300 HFA-ECC-2
ken, so it has the highest sensitivity. The GF values of HFA-ECC-1 HFA-ECC-3
specimen is below 10 during strain range of 2.0% < e < 4.5%, and 250
the sensitivity coefficient is the lowest, which is not suitable for GF value
self-sensing material. Other HFA-ECC specimens also have single- 200
digit GF values in different strain intervals, and the fluctuation
range is very small. Overall, the results of HFA-ECC specimens in 150
the self-sensing test were the worst. The number of GF values of
CB-ECC over 80 is much larger than that of ECC in strain- 100
hardening stage, and fluctuates greatly in strain range of 0.5% < e
< 4.0%, unlike the partial fluctuation of ECC specimens. 50
The apparent increase in bulk resistivity is the fact that, after
the first cracking, the new cracks will continue to appear and the 0
crack width will gradually increase with the increase of loading 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
force. In the crack pattern mentioned previous, it was observed Tensile strain (%)
that there are more cracks in HFA-ECC and CB-ECC. The average (b)
number of cracks for ECC is 0.37 times of HFA-ECC, 0.48 times of
CB-ECC (seen in Table 3). It can be seen from Fig. 9 that the GF val- 350
ues of HFA-ECC is the smallest among all mixtures. The lognormal CB-ECC-1
best-fit curves of residual crack width in ECC, HFA-ECC and CB-ECC 300 CB-ECC-2
are shown in Fig. 10. As shown in Fig. 10, the residual crack width CB-ECC-3
of ECC, HFA-ECC and CB-ECC is mainly distributed in the range of 250
20–100 lm, 20–40 lm and 20–60 lm respectively.
GF value

The main reason why the GF value of ECC fluctuates greatly in 200
the whole strain-hardening stage and smaller fluctuation of HFA-
ECC is that the new cracks in HFA-ECC make its resistivity maintain 150
a specific rate of growth, so the GF value can be maintained in a
small range of fluctuation. When the tensile strain of HFA-ECC is 100
in the range of 1.0% < e < 3.0% and 3.5% < e < 4.0%, the GF value
increases with low amplitude, while the crack width of ECC contin- 50
ues to develop at the greatest extent, so the GF value of HFA-ECC
with the steady growth of tensile strain substantial growth. It is 0
found that the development of crack width is the main reason 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
for the increase of GF value with the increase of tensile strain; com- Tensile strain (%)
pared to the number of cracks, GF value is more susceptible to (c)
crack width. Therefore, when the tensile strain reaches the maxi-
mum value, the GF value of ECC exceeds CB-ECC, although the ini- Fig. 9. GF values for each 0.5% strain interval during strain-hardening phase: (a)
tial GF value of ECC is smaller than CB-ECC. ECC, (b) HFA-ECC and (c) CB-ECC.

Fig. 11 shows the relationship between the GF value from first

cracking strain to 0.5% strain and initial resistivity of ECC, HFA-
degree of non-linear positive correlation between the GF value
ECC and CB-ECC. As shown in Fig. 11, all points are displayed in
from the first cracking strain to 0.5% strain and the initial resistiv-
the same scatter plot, and a fitting curve is obtained, that is, the
ity. Thus, it can be concluded from Fig. 11 that the lower bulk resis-
exponential function is y ¼ 233:63e0:057x ; the nonlinear correla-
tivity is advantageous for obtaining more sensing capability after
tion coefficient is R2 ¼ 0:7894, it is indicated that there is a high cracking.
8 H. Deng, H. Li / Construction and Building Materials 173 (2018) 1–9

45 and the addition of CB were beneficial to the formation of

ECC
more cracks and the effective self-controlled capacity of
40
HFA-ECC crack width.
35 CB-ECC (4) From first cracking strain to 0.5% of train, GF values were
inversely proportional to bulk resistivity (e = 0%). CB-ECC
30
Crack number

specimens showed the most sensitive response to first

25 cracking, which could be potentially utilized to detect micro
cracking by means of variation in resistivity.
20 (5) When the tensile strain of HFA-ECC is in the range of 1.0% <
15
e < 3.0% and 3.5% < e < 4.0%, the GF value increases with low
amplitude, and the crack width of ECC continues to develop
10 at the greatest extent, so the GF value of ECC with the steady
growth of tensile strain substantial growth. It is found that
5
the development of crack width is the main reason for the
0 increase of GF value with the increase of tensile strain; com-
0 10 20 30 40 50 60 70 80 90 100 110 120 pared to the number of cracks, GF value is more susceptible
Crack width/µm to crack width.
(6) Continuous cracking helps maintain a stable GF value, and
Fig. 10. Lognormal best-fit curves of residual crack width in ECC, HFA-ECC and CB- the gradual development of the crack width contributes to
ECC.
the growth of GF values. This relationship can be used to
assess the tensile strain capacity of concrete materials, pro-
viding a new method for nondestructive testing.
160
(from first cracking strain to 0.5%

140 ECC Conflict of interest

120
The author(s) declare(s) that there is no conflict of interest
100
of strain)
GF value

regarding the publication of this paper.

80 y = 233.63e -0.057x
R2 = 0.7894
60 Acknowledgement

40
The authors would like to thank the National Natural Science
20 Foundation of China (No. 51278097).

0
0 10 20 30 40 References
Resistivity (ε=0%, kohm-cm)
[1] B. Han, S. Ding, X. Yu, Intrinsic self-sensing concrete and structures: a review,
Fig. 11. The relationship between GF value and initial resistivity of ECC, HFA-ECC Measurement 59 (2015) 110–128.
[2] V.C. Li, S. Wang, C. Wu, Tensile strain-hardening behavior of polyvinyl alcohol
and CB-ECC.
engineered cementitious composite (PVA-ECC), ACI Mater. J. 98 (2001) 483–
492.
[3] J.M. Lopez-Higuera, L.R. Cobo, A.Q. Incera, A. Cobo, Fiber optic sensors in
4. Conclusions structural health monitoring, J. Lightwave Technol. 29 (4) (2011) 587–608.
[4] Guanling Song, Equivalent circuit model for AC electrochemical impedance
spectroscopy of concrete, Cem. Concr. Res. 30 (11) (2000) 1723–1730.
This study aims to develop self-sensing properties of ECC, HFA- [5] D.A. Koleva, O. Copuroglu, K. van Breugel, G. Ye, J.H.W. de Wit, Electrical
ECC and CB-ECC based on four-point probe method. The use of high resistivity and microstructural properties of concrete materials in conditions of
frequency AC with equal to or greater than 1 kHz can greatly elim- current flow, Cem. Concr. Compos. 30 (2008) 731–744.
[6] Hui Li, Hui-gang Xiao, Ou. Jin-ping, Effect of compressive strain on electrical
inate the phenomenon of material polarization. In addition, the use resistivity of carbon black-filled cement-based composites, Cem. Concr.
of four-point probe test can effectively eliminate the contact impe- Compos. 28 (2006) 824–828.
dance between electrode and cement. The specific conclusions are [7] Ravi Ranade, Jie Zhang, Jerome P. Lynch, Victor Li, Influence of micro-cracking
on the composite resistivity of Engineered Cementitious Composites, Cem.
as follows: Concr. Res. 58 (2014) 1–12.
[8] Mo Li, Vincent W.J. Lin, Jerome P. Lynch, Victor Li, Carbon black engineered
(1) The addition of CB effectively reduces the resistivity of spec- cementitious composites-mechanical and electrical characterization, Am.
Concr. Inst. Spec. Publ. 292 (2011) 59–73.
imens, almost 0.46 times of ECC, and 0.28 times of HFA-ECC. [9] Tsung-Chin Hou, Wireless and electromechanical approaches for strain sensing
The bulk resistivity of CB-ECC is proved to be the least differ- and crack detection in fiber reinforced cementitious material PhD thesis,
ential, favoring the self-sensing properties test. University of Michigan, USA, Ann Arbor, 2008 [in American].
[10] B. Han, J. Ou, Embedded piezoresistive cement-based stress/strain sensor,
(2) The ultimate tensile strain of ECC, HFA-ECC and CB-ECC were Sens. Actuators, A 138 (2) (2007) 294–298.
3.84%, 4.51% and 4.02% respectively, and corresponding ulti- [11] D.L. Nguyen et al., Comparative electromechanical damage-sensing behaviors
mate tensile strength were 5.69 MPa, 5.78 MPa and 4.26 of six strain-hardening steel fiber-reinforced cementitious composites under
direct tension, Composites Part B 69 (2015) 159–168.
MPa respectively, and corresponding first cracking strength
[12] T.C. Hou, J.P. Lynch, Monitoring strain in engineered cementitious composites
is 4.61 MPa, 4.00 MPa and 3.28 MPa respectively. The combi- using wireless sensors, in: Proceedings of the International Conference on
nation of CB and AEA effectively improves the tensile prop- Fracture, Turin, Italy, 2005, pp. 1488–1493.
erties of ECC materials. [13] T.C. Hou, J.P. Lynch, Conductivity-based strain monitoring and damage
characterization of fiber reinforced cementitious structural components, in:
(3) The average number of residual cracks of ECC, HFA-ECC and Proceedings of Smart Structures and Materials, San Diego, CA, American, 2005,
CB-ECC in Section 1 were 14, 38 and 29 respectively, and pp. 419–429.
corresponding to the average crack width were 44 lm, 14 [14] W.J. Mccarter, H.M. Taha, B. Suryanto, G. Starrs, Two-point concrete resistivity
measurements: interfacial phenomena at the electrode–concrete contact zone,
lm and 22 lm respectively. The increase of fly ash content Meas. Sci. Technol. 26 (8) (2015) 085007.
 H. Deng, H. Li / Construction and Building Materials 173 (2018) 1–9 9

[15] W. McCarter, R. Brousseau, The AC response of hardened cement paste, Cem. [22] S. Wang, VC. Li, Lightweight Engineered Cementitious Composites (ECC), in:
Concr. Res. 20 (6) (1990) 891–900. Proceedings of High Performance Fiber Reinforced Cement Composites
[16] V.C. Li, C.K.Y. Leung, Steady-state and multiple cracking of short random fiber (HPFRCC4) Workshop, Ann Arbor, American, 2003, pp. 379–390.
composites, J. Eng. Mech. 118 (11) (1992) 2246–2264. [23] S. Muzenski, I. Flores-Vivian, K. Sobolev, Hydrophobic engineered
[17] N. Puthipad, M. Ouchi, S. Rath, A. Attachaiyawuth, Enhancement in self- cementitious composites for highway applications, Cem. Concr. Compos. 57
compactability and stability in volume of entrained air in self-compacting (2015) 68–74.
concrete with high volume of fly ash, Constr. Build. Mater. 128 (2016) 349–360. [24] K. Tosun-Felekoglu, B. Felekoglu, R. Ranada, V.C. Li, The role of flaw size and
[18] S. Rath, N. Puthipad, A. Attachaiyawuth, M. Ouchi, Critical size of entrained air fiber distribution on tensile ductility of PVA-ECC, Composites Part B 56 (1)
to stability of air volume in mortar of self-compacting concrete at fresh state, J. (2014) 536–545.
Adv. Concr. Technol. 15 (2017) 29–37. [25] E.H. Yang, Y.Z. Yang, V.C. Li, Use of high volumes of fly ash to improve ECC
[19] Nipat Puthipad, Massahiro Ouchi, Sovannsathya Rath, Anuwat mechanical properties and material greenness, ACI Mater. J. 104 (6) (2007)
Attachaiyawuth, Enhancement in self-compactability and stability in volume 620–628.
of entrained air in self-compacting concrete with high volume fly ash, Constr. [26] S.X. Wang, V.C. Li, Engineered cementitious composites with high-volume fly
Build. Mater. 128 (2016) 349–360. ash, ACI Mater. J. 104 (3) (2007) 233–241.
[20] Y.M. Gao, H.S. Shim, R.H.H. And, E.M. Suuberg, N.Y.C. Yang, Effects of carbon on [27] C.S. Poon, L. Lam, Y.L. Wong, A study on high strength concrete prepared with
air entrainment in fly ash concrete: the role of soot and carbon black, Energy large volumes of low calcium fly ash, Cem. Concr. Res. 30 (2000) 447–455.
Fuels 11 (2) (1997) 457–462.
[21] H.K. Kim, J.H. Jeon, H.K. Lee, Workability and Mechanical, acoustic and thermal
properties of lightweight aggregate concrete with a high volume of entrained
air, Constr. Build. Mater. 29 (2012) 193–200.