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

net/publication/358262646

Self-sensing properties and piezoresistive effect of high ductility

cementitious composite

Article in Construction and Building Materials · March 2022

DOI: 10.1016/j.conbuildmat.2022.126390

CITATIONS READS

0 32

3 authors, including:

Jingming Cai
Southeast University (China)
46 PUBLICATIONS 456 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

The development of green geopolymers with construction waste View project

The fire-resistant performance of ECC-encased CFST columns View project

All content following this page was uploaded by Jingming Cai on 04 February 2022.

The user has requested enhancement of the downloaded file.

 Construction and Building Materials 323 (2022) 126390

Contents lists available at ScienceDirect

Construction and Building Materials

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

Self-sensing properties and piezoresistive effect of high ductility

cementitious composite
Jinsheng Han , Jinlong Pan *, Jingming Cai
Key Laboratory of Concrete and Prestressed Concrete Structures of Ministry of Education, School of Civil Engineering, Southeast University, Nanjing 211189, PR China

A R T I C L E I N F O A B S T R A C T

Keywords: Self-sensing high ductility cementitious composite (HDCC) has unique advantages in structural health moni­
High ductility cementitious composite toring. In this paper, the carbon fiber (CF) reinforced HDCC with the addition of nano-carbon black (NCB) was
Carbon fiber designed and tested under compression and tension conditions. The sensing properties and piezoresistive effect of
Nano-carbon black
the proposed HDCC were analyzed based on the experimental results. The optimal mix design of HDCC was
Self-sensing
Piezoresistive effect
proposed with the consideration of both mechanical and electrical characteristics, i.e., tensile ductility, strain
sensitivity, signal-to-noise ratio, and signal linearity. It was concluded that the ultimate tensile strain of the
proposed HDCC can be as high as 3.5% with the water-binder ratio of 0.3, indicating that it is feasible to develop
conductive HDCC with the addition of hybrid fibers, i.e., polyvinyl alcohol fiber and CF. It was also found that
the addition of CF is conducive to the electrical sensitivity and signal linearity of HDCC, while the addition of
NCB is effective to improve the signal-to-noise ratio. The piezoresistive effect model of HDCC was also proposed
based on the tunnel effect theory, in which the relationship between mechanical and electrical characteristics
was established, and the validity of the proposed model was verified experimentally.

1. Introduction electrical signals, the stress and deformation can be accurately deter­
mined. Compared with conventional health monitoring methods with
High ductility cementitious composites (HDCCs), also known as en­ additional sensors, self-sensing cementitious material has the advan­
gineering cementitious composites (ECC), have pseudo strain-hardening tages of high sensitivity, simple construction, high durability and good
and multiple cracking behaviors under tensile load [1]. These properties compatibility with structures, which has a great application prospect in
are obtained by tailoring the properties of fiber, matrix and their structure health monitoring [8]. However, ordinary self-sensing
interface properties [2]. In general, the addition of 1.5–2.0 vol% fibers cementitious material can only monitor compressive loads and cannot
to the matrix enables the material to have pseudo strain-hardening reflect the deformation under tension, which seriously limits its further
property, and the ultimate tensile strain exceeds 3% stably. The steady applications [9]. This is because ordinary self-sensing cementitious
state crack opening behavior of HDCC could effectively control the composites are commonly quasi-brittle materials, which are easy to
width of the cracks [3], which is conducive to preventing the invasion of crack and fail under tension load. Considering the inherent defect, the
external harmful substances and improving the durability. The tensile self-sensing HDCC can not only monitor the stress and deformation of
strain-hardening property of HDCC also increased the safety of engi­ materials, but also have extraordinary tensile ductility, which can
neering structures when compared with conventional concrete, as con­ monitor the structure health in a wider range [10].
crete is inherently brittle and easy-cracking [4,5]. These excellent Generally, the conductivity of self-sensing materials is determined by
properties provide the possibility for it to become a self-sensing building the conductive fillers incorporated into the materials. The conductive
material with high ductility, wide detection range and strong durability fillers mainly include carbon-based and metal-based fillers [11–14],
[6]. such as carbon fiber (CF), carbon black (CB), carbon nanotubes,
Usually, self-sensing cementitious composite has a piezoresistive graphite, graphene, steel slag, metal oxide, steel fiber (SF) and nickel
effect, that is, the behavior that the resistivity of material changes when powder. CF has become a widely used fibrous conductive filler with its
it is subjected to an external force [7]. By analyzing the changes in excellent electrical conductivity, corrosion resistance, high temperature

* Corresponding author.
E-mail address: [email protected] (J. Pan).

https://doi.org/10.1016/j.conbuildmat.2022.126390
Received 26 November 2021; Received in revised form 20 December 2021; Accepted 8 January 2022
Available online 31 January 2022
0950-0618/© 2022 Elsevier Ltd. All rights reserved.
J. Han et al. Construction and Building Materials 323 (2022) 126390

resistance and low cost. Due to its larger aspect ratio, CF can signifi­ 0.4 vol%, 0.6 vol% and 0.8 vol%, respectively. In order to compare the
cantly improve the material conductivity and reduce the percolation effects of NCB on the sensing and mechanical properties, the dosages of
threshold. Nano-carbon black (NCB) is also a common conductive filler NCB were introduced into material at cementitious material weight
because of its excellent conductivity, high specific surface area and low percentage of 0 wt%, 0.5 wt%, 1.0 wt% and 1.5 wt%, respectively.
cost [15]. NCB, with the particle filling effect and excellent electrical According to the demand of liquidity, the W/Bs were 0.25, 0.3 and 0.35,
conductivity, not only can improve the mechanical properties, but also respectively. The mix proportion and the raw materials properties were
can provide active contacts to improve the sensitivity and signal-to- shown in Table 1 and Table 2. For tensile ductility and sensing proper­
noise ratio (SNR) of the sensing property [16,17]. Mixing the two ties, PVA fiber, CF and NCB are dominating materials, and their
conductive fillers into the cementitious material can improve the macroscopic and microscopic morphologies are shown in Fig. 1.
comprehensive performances. For example, Ou et al. [18] mixed NCB
and CF into the cementitious material to improve the repeatability and
linearity of sensing property. In addition, due to the low cost of NCB, 2.2. Specimen preparation
mixing CF and NCB can reduce the construction cost [19].
The electrical conductivity of self-sensing HDCC is the basis of The laboratory temperature is 20 ± 2℃ and the relative humidity is
sensing properties and it can be obtained by adding an appropriate 50 ± 5%. In the preparation process, the dispersion of CF is critical issue
amount of conductive filler into matrix. Shi et al. [20,21] mixed 0.58 vol as it is easy to form agglomerations [26]. If agglomeration occurs, it will
% SF and 1 wt% CB into HDCC with 1.75 vol% polyvinyl alcohol (PVA) not only affect the mechanical properties seriously, but also affect the
fibers, which improved the conductivity of HDCC and made it have sensing performance [27]. In this paper, the dispersion methods were
piezoresistive effect. Hou and Lynch [22] mixed 0.1 vol% SF and 0.4 vol adjusting the mixing sequence and adding dispersant [16,28]. The
% CF into HDCC with 2 vol% PVA, respectively, and self-sensing HDCC method is shown in Fig. 2, the CF was added to the liquid first, and
with obvious piezoresistive effect was obtained. Al-Dahawi et al. [23,24] mixed for 2 min. Then the solid raw materials were added and mixed
mixed CF into HDCC with 2 vol% PVA and the HDCC with excellent well. During this process, the dispersant was added to the water in
sensing performance was obtained. It should be noted that since the advance to increase the viscosity of the water and promote the disper­
tensile ductility of HDCC is achieved by optimizing fiber, matrix and sion of the CF. The effectiveness of the dispersion methods were verified
their interface properties, the incorporation of conductive fillers of by analyzing the coefficients of variation of different specimen imped­
different shapes and sizes into HDCC is bound to affect its mechanical ances in the previous work [25]. The mixing process of the material is
and sensing properties. shown in Fig. 2. The fresh paste was molded and then covered with
In this paper, the effects of different water-binder ratio (W/B), dos­ plastic film. The specimens were demolded after 1 day and then cured in
ages of CF and NCB on the basic mechanical and sensing properties of a standard curing room (temperature: 20 ± 2℃, relative humidity:
HDCC were studied. For mechanical properties, the effects of W/B and
the dosage of conductive filler on the tensile ductility and compressive Table 2
strength were analyzed through uniaxial tensile and compression tests. The raw materials properties.
For sensing properties, the specific impedance under alternating current Raw materials Properties
(AC) at the frequency of 1 KHz was measured by using the Kelvin Four-
Cement P.O. 42.5, CaO wt%=61.23%, SiO2 wt%=21.45%, Al2O3 wt%=
terminal method, and the sensing properties under uniaxial tension and 5.52%, fineness = 350 m2/kg, density = 3.10 g/cm3, bulk density
compression tests were studied. The effects of dosages of CF and NCB on = 1.44 g/cm3
sensitivity, linearity and SNR of sensing properties were analyzed. Fly ash CaO wt%=3.71%, SiO2 wt%=54.86%, Al2O3 wt%=29.16%,
density = 2.60 g/cm3, bulk density = 1.00 g/cm3
Finally, based on tunnel effect theory and the equivalent circuit model,
Quartz sand Size is between 0.20 mm and 0.30 mm, bulk density = 1.70 g/cm3
the relationship between the spacing of conductive fillers and the spe­ Water Ordinary tap water
cific impedance change of material was analyzed and the piezoresistive Superplasticizer Polycarboxylic acid superplasticizer
effect model was established. Finally, the validity of the proposed model Dispersant Surfactant, hydroxymethyl cellulose, viscosity greater than 4 ×
was verified by experimental results. 104 mpa⋅s
Defoamer Organosilicon powder
CF Pan-based, length = 9 mm, diameter = 7–10 μm, density =
2. Experimental design 1.60–1.70 g/cm3, tensile strength = 4 GPa, tensile modulus = 220
GPa, volume resistivity = 1.5 × 10-3 Ω⋅cm
NCB Density = 1.80–2.10 g/cm3, average size = 100 nm, average
2.1. Mix proportion and raw materials
specific surface area = 125 m2/g
PVA fiber Length = 12 mm, diameter = 39–40 μm, density = 1.30 g/cm3,
Conductive filler is the most important factor affecting the sensing modulus of elasticity for tension = 43 GPa, Tensile strength =
properties of material. According to the previous test [25], the perco­ 1620Mpa.
lation threshold of CF is 0.6 vol%. In this paper, the dosages of CF were

Table 1
Mix proportion.
W/ Water/(kg/ Cement/(kg/ Fly ash/(kg/ Sand/(kg/ Superplasticizer/(kg/ Surfactant/(kg/ Defoamer/(kg/ CF/vol CB/wt PVA /vol
C m3) m3) m3) m3) m3) m3) m3) % % %

0.25 345 230 1150 230 22.0 1.38 1.38 0.4 10 2

0.6
0.8
0.35 441 210 1050 210 12.6 1.26 1.26 0.4
0.6
0.8
0.30 396 220 1100 220 18.5 1.32 1.32 0.4
0.8
0.6 0
5
10
15

2
J. Han et al. Construction and Building Materials 323 (2022) 126390

(a) PVA fiber (b) CF (c) NCB

Fig. 1. The macroscopic and microscopic morphologies of PVA fiber, CF and NCB.

Fig. 2. Mixing process of the material.

above 95%) for 28 days. In sensing properties test, the electrodes need to 2.3. Experimental method for mechanical and sensing properties
be installed on the specimen in advance. When the fresh paste was
molded, two copper electrodes were inserted into the matrix (as shown As shown in Fig. 4(a) and Fig. 4(b), the universal testing machine
in Fig. 3). And the gap between electrodes and matrix was blocked with with a capacity of 100KN and the automatic pressure testing machine
conductive silver glue to reduce the contact resistance. The copper with a capacity of 3000KN were used for tension and compression tests,
electrode is shown in Fig. 3 (c), and the shape was used to reduce the respectively. In the tension test, the loading rate is 0.5 mm/min and the
poor contact between electrode and matrix. Before the test, in order to test was stopped when the tensile stress dropped to 80% of its maximum
reduce the influence of moisture changes on impedance and mechanical value. In the compression test, the loading rate is 0.3 MPa/s and the test
properties, the specimens were placed in a ventilated place in the lab­ was also stopped when the compressive stress dropped to 80% of the
oratory for 7 days [29]. The specimen dimensions are shown in Fig. 3, maximum of its maximum value.
and totally three identical specimens were prepared for each mix design. The AC two-electrode method was used in impedance test. Compared
with the four-electrode method, the two-electrode method has the
advantage of easy operation and less impact on mechanical properties
[30]. Although it can bring some contact resistance, it does not affect the
accuracy of resistance change. As shown in Fig. 4(c), the test equipment
was LCR digital bridge, and the specimen impedance was measured by
Kelvin Four-terminal method. Since the impedance changes with the AC
frequency, the AC frequency of 1 KHz was used in order to reduce the
polarization effect and improve the accuracy [31]. Insulating tape and
insulating PC board were used in the specimen and instrument clamping
area to isolate the current.

2.4. The evaluation index of sensing properties

In order to quantitatively evaluate the sensing properties, several

evaluation indices were applied in this paper, i.e., specific impedance,
fraction change in specific impedance (FCI), strain sensitivity, linearity
and SNR. Specific impedance refers to the impedance of the material per
unit length, which indicates the material conductivity and the calcula­
tion formula is shown in Eq. (1).
|Z|A
(a) Tension (b) Compression (c) Copper electrode ρ= (1)
L
Fig. 3. Specimen sizes and electrode style. where ρ is the specific impedance, |Z| is the material impedance, A is

3
J. Han et al. Construction and Building Materials 323 (2022) 126390

Fig. 4. Test set-up (a) Tension (b) Compression (c) Impedance.

the area of the specimen cross section, L is the test length of the 3. Results and discussions
specimen.
In order to show the specific impedance change clearly, FCI was used 3.1. Uniaxial compression and tension results
to represent the change of specific impedance of specimens. It refers to
the ratio of the change value of specific impedance at some time to the The compression test was carried out on the cube specimens, and the
initial specific impedance, the formula is shown in Eq. (2). results are shown in Fig. 5. As shown in Fig. 5, W/B and the dosage of CF
have significant effects on compressive strength. The strength of the
FCI = (ρt − ρ0 )/ρ0 × 100 % (2)
material increased with the decrease of the W/B, but the discreteness of
where ρt and ρ0 are the specific impedances at time t and the initial the strength also increased because of the lack of fluidity. In specimen
time. preparation, the addition of dispersant can improve the viscosity of
The strain sensitivity expressed by gauge factor (GF) is usually used liquid and promote the dispersion of CF, but it will lead to poor fluidity
to describe the sensing properties. GF represents the change of specific and workability of fresh paste, which makes a greater discreteness of the
impedance under unit strain. The larger the value, the greater the strength [32]. The compressive strength decreased with the increase of
sensitivity and the better the sensing properties. The calculation formula the dosage of CF, because the addition of CF introduces additional de­
is as following: fects and weak interfaces. The more CFs, the worse the dispersion, and
the agglomeration of CFs forms new defects, which reduce the density of
FCI
GF = (3) the matrix, leading to the reduction of the strength [26].
ε
For self-sensing HDCC, high ductility is the critical mechanical
where ε is specimen strain along the loading direction.
The linearity represented by δ refers to the ratio of the maximum
difference between the impedance output data curve and the line fitting 40
curve to the full range output value. The lower the linearity, the smaller
the error and the better the sensing properties. The formula is shown in 0.8vol%
Eq. (4). 36 0.6vol%
Compressive strength/MPa

0.4vol%
ΔYmax
δ= (4)
Y 32
where ΔYmax is the maximum difference and Y is the maximum range
output value.
28
SNR refers to the ratio between the signal and the noise, its unit is dB.
The larger the SNR, the better the sensing properties. The formula is
shown in Eq. (5). 24
( ) ( 2)
Ps A
SNR = 10log10 = 10log10 s2 (5)
Pn An 20
0.25 0.3 0.35
where Ps and Pn are the power of signal and noise, respectively. As W/B
and An are the amplitude of signal and noise, respectively.
Fig. 5. Compressive strength results of specimens with different W/B and the
dosages of CF.

4
J. Han et al. Construction and Building Materials 323 (2022) 126390

property for the health monitoring. In this paper, the high ductility was continue to open and the fibers pull out or rupture, which make the
achieved by steady state cracking of the fine cracks, that is, the material material have quasi-brittle behavior and low ductility. As shown in
exhibits pseudo strain-hardening property under tensile test. The tensile Fig. 6(g) to Fig. 6(i), when the W/B is 0.35, the first cracking stress of the
stress–strain curves of HDCC with different W/Bs and the dosages of CF matrix is lower. However, the residual energy of the bridge between
are shown in Fig. 6. fiber and matrix is low, that is, the bond between the fiber and the
As shown in Fig. 6, W/B has an important influence on the tensile matrix is too weak because of the low W/B. As a result, the fiber keeps
properties of HDCC. With the increase of W/B, the first cracking stress pulling out after first cracking and the stress–strain curves shows a long
and peak stress decreased gradually, which is determined by the decline section. As shown in Fig. 6(d) to Fig. 6(f), when the W/B is 0.3,
strength of the matrix. While the tensile ductility is affected by many the matrix stress and the bond between the fiber and the matrix are
factors. Studies show that the steady state multiple cracking behavior of appropriate. After first cracking, the load on the cracked section can still
HDCC needs to meet the First Crack Stress Criterion and Steady State be transferred to the matrix by fiber bridging, and new cracks are
Cracking Criterion [33]. The First Crack Stress Criterion requires that constantly formed on the matrix and steady state cracking occurs. The
the first cracking stress must not exceed the maximum bridging stress of bearing capacity can still rise slowly until the tensile stress is higher than
the fiber at the crack section. The Steady State Cracking Criterion re­ the fiber bridging stress. In the tension process, the material shows
quires that the residual energy of the bridge between fiber and matrix obvious strain-hardening characteristics. Finally, when the tensile stress
should remain in a reasonable range after the first cracking to meet the is higher than the fiber maximum bridging stress, the specimens begin to
requirement of steady state multiple cracking. fail and the stress–strain curve becomes the descending section. There­
As shown in Fig. 6(a) to Fig. 6(c), when the W/B is 0.25, the first fore, the specimen with a W/B of 0.3 meets the requirements of high
cracking stress of the matrix is higher. After cracking, the crack will ductility and steady state multiple cracking.

5 5 5
(a) W/B=0.25 CF=0.4vol% (b) W/B=0.25 CF=0.6vol% (c) W/B=0.25 CF=0.8vol%
4 4 4
Tensile stress (MPa)

Tensile stress (MPa)

Tensile stress (MPa)

3 3 3

2 2 2

1 1 1

0 0 0
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
Pseudo tensile strain (%) Pseudo tensile strain (%) Pseudo tensile strain(%)

5 5
5
(e) W/B=0.3 CF=0.6vol% (f) W/B=0.3 CF=0.8vol%
(d) W/B=0.3 CF=0.4vol%
4 4
4
Tensile stress (MPa)

Tensile stress (MPa)

Tensile stress (MPa)

3 3
3

2 2
2

1 1 1

0 0 0
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
Pseudo tensile strain(%) Pseudo tensile strain (%) Pseudo tensile strain (%)

5 5
5
(g) (h) W/B=0.35 CF=0.6vol%
W/B=0.35 CF=0.4vol% (i) W/B=0.35 CF=0.8vol%
4 4 4
Tensile stress (MPa)
Tensile stress (MPa)

Tensile stress (MPa)

3 3 3

2 2 2

1 1 1

0 0 0
1 2 3 4 5 1 2 3 4 5 1 2 3 4 5
Pseudo tensile strain (%) Pseudo tensile strain (%) Pseudo tensile strain (%)

Fig. 6. The tensile stress–strain curves of materials with different W/Bs and the dosages of CF.

5
J. Han et al. Construction and Building Materials 323 (2022) 126390

As shown in Fig. 6(d) to Fig. 6(f), the dosage of CF also has a certain
effect on the tensile properties of the HDCC. With the increase of the
dosage of CF, the first cracking stress and peak stress decreased slightly,
while the tensile ductility increased first and then decreased. When the
dosage of CF is lower, CF can assist PVA fibers to bridge micro-cracks
and improve the tensile ductility. When the dosage of CF is higher, the
dispersion of CF becomes worse. It is easy to agglomerate and lead to
poor fluidity and workability of fresh paste, which may change the
toughness and cracking stress of matrix [32]. Therefore, high the dosage
of CF is not good for tensile ductility.
The tensile failure modes of specimens with three W/Bs were
analyzed. As shown in Fig. 7 (a), when the W/B is 0.25, the first cracking
stress is higher, and only a few cracks occur after the initial crack, fol­
lowed by the main crack. In this failure mode, there are a small number
of cracks on the specimen, and the main crack occurs in the weak po­
sition of the specimen. As shown in Fig. 7 (b), when the W/B is 0.3, the
matrix stress and the bond between the fiber and the matrix are
appropriate. In this failure mode, a large number of fine cracks appear
on the specimen and are evenly distributed, which is a typical fine
cracking mode. As shown in Fig. 7 (c), when the W/B is 0.35, the first
cracking strength of the matrix is low. As the bond between the fiber and
the matrix is too weak, the fiber keeps pulling out after the first cracking,
and then the crack forms the main crack. In this failure mode, a rela­ Fig. 8. The microscopic morphologies of PVA fiber and CF on the specimen
tively wide main crack appears on the specimen, and there is a small fracture surface.
amount of slag on both sides of the crack. These indicate that the matrix
strength is low and the bond with the fiber is weak. Therefore, the failure the material is the result of joint bridging action of PVA fiber and CF.
mode of the specimen with a W/B of 0.3 is determined to meet the re­ Due to their different diameters, they can play a synergistic role at
quirements of multiple cracking. different scales, which are beneficial to the tensile ductility of materials.
To future analyze the effects of CF and PVA fiber on the material Based on the results of compression and tension test, the W/B of 0.3
properties, the scanning electron microscope test on the microstructure and the dosage of CF with 0.6 vol% were selected as the basic mix ratio
of the specimen with the W/B of 0.3 after tensile test was carried out. to explore the influence of NCB. The change in mechanical properties
The results are shown in Fig. 8. was analyzed by mixing NCB with different dosages and the mix pro­
As shown in Fig. 8, the PVA fiber showed obvious tensile ruptured, portion is shown in Table 1. The results of compressive strength and
and the surface was damaged due to the friction in the process of pulling tensile stress–strain curves are shown in Fig. 9.
out, forming “barbs” and scratches. These indicate that PVA fiber had Fig. 9 illustrates that the compressive strength of the material
played a bridging role in the tensile process. The CFs only pull out on the increased first and then decreased with the increase of the dosage of
fracture surface, but does not rupture. This is because the tensile NCB. NCB, as a nanoscale material, can fill the pores and improve the
strength and modulus of CF are much higher than that of PVA fiber and density of the matrix, which improves the strength. However, when
matrix. There were some shallow scratches and matrix slag attached on there is too much NCB, it is easy to produce agglomerations. The
the CF surface, which indicates that CFs also play a certain bridging role excessive NCB is equivalent to defects in the matrix, leading to the
in the tensile process. Therefore, the tensile strain-hardening behavior of

(a) (b) (c)

Fig.7. Three failure modes (a) W/B = 0.25, (b) W/B = 0.3, (c) W/B = 0.35.

6
J. Han et al. Construction and Building Materials 323 (2022) 126390

30 6
28 W/B=0.3
5

Compressive strength (MPa)

26

Tensile stress (MPa)

24 4
22
20 3
18 NCB=0 wt%
2 NCB=0.5 wt%
16
NCB=1.0 wt%
14 1 NCB=1.5 wt%
12
10 0
0.0 0.5 1.0 1.5 1 2 3 4 5
NCB (wt%) Pseudo tensile strain (%)

(a) Compressive strength (b) Tensile stress-strain curves

Fig. 9. The compressive strength and tensile ductility of specimens with different dosage of NCB.

reduction of the strength [34]. In the tension test, the dosage of NCB had stage. In the elastic stage, elastic deformation occurs in the specimens
no obvious effect on the first cracking stress and tensile ductility of the and the impedance decreases. With the increase of strain, the average
material. Studies shown that the increase of defects is conducive to the spacing of conductive fillers inside the specimen decrease along the
multiple cracking of materials [35]. In this paper, NCB can be equivalent direction of stress, and conductive tunnel effect is more likely to occur
to some defects and slightly reduce matrix toughness, yet it is a nano­ [36], which make the impedance and FCI decrease. In non-elastic stage,
scale material and the size is much smaller than the pores generated in with the increase of stress, the micro-cracks begin to generate and
the preparation process. NCB has no obvious effect on the multiple expand inside the specimen, which will break some conductive paths
cracking behavior of material, while it can fill some pores and affect the and lead to the increase of impedance. However, the impedance
compressive strength. reduction caused by compressive strain still plays a dominant role and
the specimen shows a reduction of impedance. In descent stage, the
3.2. Sensing properties specimen has been crushed, but the specimen can still maintain the
intact shape because of fiber bridging effect. There are a lot of cracks in
As the mix design of HDCC was optimized, the sensing properties of the specimen, and the impedance increases rapidly. In unstressed stage,
HDCC under uniaxial compression and uniaxial tension test were further the compression test is terminated, the impedance stops changing, some
studied. The specimens with W/B of 0.3 and CF of 0.6 vol% and NCB of damages occur inside the specimen, and the impedance higher than the
1.0 wt% were taken as examples. The curves are the mean value curves initial impedance.
of three specimens, and the results are shown in Fig. 10 and Fig. 11. As shown in Fig. 11 (a), according to the change of stress–strain
As shown in Fig. 10 (a), according to the change of curve trend, the curve trend, the curve can be divided into three stages, namely, elastic
compressive stress–strain curve can be divided into three stages, namely, stage, strain-hardening stage and descent stage, respectively. As shown
elastic ascending stage, non-elastic ascending stage and descent stage, in Fig. 11 (b), there is a strong consistency between the impedance
respectively. As shown in Fig. 10 (b), the change trend of FCI is change and the strain change. According to the change law of FCI, the
consistent with that of the stress–strain curve. According to the change FCI-time curve also can be divided into three different stages, namely,
law of FCI, the FCI-time curve can be divided into four different stages, elastic stage, strain-hardening stage and descent stage. In the elastic
namely, elastic stage, non-elastic stage, descent stage and unstressed stage, elastic deformation occurs in the specimens under the tensile

25 0.7 35

Strain 30
Fraction change in impedance (%)

Stress 0.6
Compressive strength (MPa)

FCI-1
20 Unstressed 25
0.5 FCI-2 stage 20
FCI-3
15 15
Strain (%)

Non-elastic 0.4 Descent

ascending stage Descent stage 10
stage 0.3
10 5

0.2 0
Non-elastic
Elastic -5
5 stage
0.1 stage
Elastic -10
ascending stage
0 0.0 -15
0.001 0.002 0.003 0.004 0.005 0.006 10 20 30 40 50 60 70 80 90
Strain Time (s)
(a) stress-strain curve (b) Strain-time and FCI-time curves
Fig. 10. The results under compression test.

7
J. Han et al. Construction and Building Materials 323 (2022) 126390

4.0 8 160
Tensile strain FCI-1

Fraction change in impedance (%)

3.5 Stress 7 0.10
FCI-2 140

Fraction change in impedance/%

0.08

3.0 6 Descent 120

Tensile strain (%)

Tensile strain/%
Tensile stress (MPa)
0.06

0.04 stage
2.5 Descent 5 0.02
100
Strain hardening stage stage
4 80
0.00

2.0 500 1000 1500 2000 2500 3000

Time /0.125s

1.5 3 60
Elastic
1.0 stage 2 40
Elastic
1 stage Strain hardening stage 20
0.5

0.0 0 0
1 2 3 4 5 6 7 2500 5000 7500 10000 12500 15000 17500
Pseudo tensile strain (%) Time (0.125s)
(a) Stress-strain curve (b) Strain-time and FCI-time curves
Fig. 11. The results under tension test.

stress and the strain is between 0 and 0.1%. Meanwhile, the change FCI represents the elastic stage and “S” represents the strain-hardening
is about between 0 and 2%. Strain-hardening stage contributes more stage.
than 95% of the deformation during the tension process. In the stage, the It can be seen from Table 3 that the dosages of CF and NCB had
specimen has steady state cracking characteristics, and the cracks are important effects on the sensing properties of HDCC. In compression
bridged by fibers. As conductive fiber filler, CF can not only bridge the ascending stage, the sensitivity and linearity decreased with the increase
cracks, but also transmit current, so that the specimen still has con­ of the dosage of CF, while the SNR increased. This is because the more
ductivity under cracking. However, cracking inevitably destroys some CFs, the higher density of conductive paths formed by CFs, resulting in a
the conductive paths, resulting in a rapid increase in impedance. Steady lower impedance [37]. Especially, when the dosage of CF is higher than
state cracking makes the impedance change almost linear, which es­ the percolation threshold, the change of impedance becomes small and
tablishes a foundation for the piezoresistive effect. In descent stage, the the sensitivity decreases under the compression. The denser the
tensile stress is greater than the bridging force of the fiber, and main conductive network, the less the data error, so the linearity decreases.
crack appears and expands. At this moment, CF is pulled out constantly Meanwhile, the probability of impedance response to the strain in­
and the conductive paths are destroyed, resulting in a sharp increase in creases, which reduces noise and improves SNR. With the increase of the
impedance until the specimen is destroyed. dosage of NCB, the sensitivity decreases and the SNR increases, while
In both the compression and tension processes, the change of FCI is the linearity has no change significantly. The addition of NCB provides
surprisingly consistent with the strain. Even slight changes in strain will active contacts for impedance change during the compression process,
lead to obvious changes in impedance, which reflects the excellent which shortens the gap between CFs, and results in lower impedance
sensitivity of the material sensing properties. change and lower sensitivity. The addition of active contacts also re­
In order to further optimize the sensing properties, the effects of duces noise and improves the SNR. As an auxiliary granular conductive
different dosages of CF and NCB on the sensitivity, linearity and SNR filler, a small amount of NCB does not play a major conductive role and
were analyzed in this paper. The results are shown in Table 3. For has little effect on the accuracy of data, thus it has no significant effect
structural health monitoring, it is valuable to monitor the deformation of on the linearity.
materials before failure. Therefore, the sensing properties in the In tension elastic stage, with the increase of the dosage of CF, the
compression ascending stage and the tension elastic and strain- sensitivity and linearity gradually decreased, which is the same as the
hardening stages were analyzed in this paper. In Table 3, “E” principle of the compression. The increase of CF leads to the decrease of

Table 3
The effects of dosages of CF and NCB on the sensing properties.
Loading stage CF/ vol% CB/ wt% Sensitivity Linearity SNR/dB Strain range

Compressive ascending stage 0.4 1.0 37.5–40.0 0.12–0.14 15–16 0–0.004

0.6 32.5–35.0 0.06–0.08 21–25
0.8 23.5–25.0 0.04–0.06 26–30
0.6 0.0 100–120 0.04–0.06 14–16
0.5 35–40 0.03–0.05 20–22
1.5 30–35 0.05–0.07 22–25
Tension E 0.4 1.0 6000–7000 0.250–0.300 16–18 0–0.050
S 1750–2000 0.020–0.030
E 0.6 5000–6000 0.080–0.120 30–32
S 2000–2100 0.003–0.004
E 0.8 2000–2500 0.050–0.100 35–37
S 2250–2500 0.002–0.003
E 0.6 0.0 5000–6000 0.050–0.080 26–28
S 1600–1700 0.006–0.007
E 0.5 5000–6000 0.050–0.080 32–34
S 2000–2100 0.002–0.003
E 1.5 5000–6000 0.040–0.090 34–36
S 1800–2000 0.002–0.003

8
J. Han et al. Construction and Building Materials 323 (2022) 126390

the initial impedance, sensitivity and linearity. In strain-hardening under tensile and compressive tests were analyzed according to the test
stage, the increase of the dosage of CF results in increased sensitivity results.
and decreased linearity. In the case of cracking, CF plays a bridging role
and also builds conductive networks across the cracks, so that the ma­
terial still has sensing properties. The more the dosage of CF, the denser 4.1. Derivation of the model
the conductive network across the cracks and the higher the sensitivity.
The linearity depends on the accuracy of the signal, and the increase of In order to better explain the change of the spacing of conductive
dosage of CF improves the accuracy and reduces the linearity. In the fillers, the distribution of conductive fillers inside the matrix was
whole tension process, NCB was separated by tensile micro-cracks, simplified and analyzed. As shown in Fig. 12, θ is the included angle
hindering its conductivity, so the increase of NCB has no obvious ef­ between CF and the deformation direction, ΔS is the spacing of
fect on sensitivity and linearity. For SNR, the values in elastic and strain- conductive fillers and Sc is the critical spacing for tunnel effect theory.
hardening stages were the same, and it increased with the increase of the In order to better analyze the piezoresistive effect, the equivalent
dosages of CF and NCB. This is because the increase of the dosage of circuit model was established. The distribution of conductive fillers was
conductive fillers improves the conductive networks, reduces the noise used as the basis for establishing the equivalent circuit model in this
and improves the SNR. paper. As shown in Fig. 11, the distributions of conductive filler can be
The changes in the dosages of CF and NCB have important effects on divided into three types, namely connected path, semi-connected path
the sensing properties. The increase in the dosage of CF has opposite and insulated path. In the connected path, the CFs are overlapped with
effects on the sensitivity of elastic and strain-hardened stages. This is each other or the distance is less than the critical distance, so it can be
because that CF only plays a conductive role in the elastic stage, but equivalent to resistance (R1). In semi-connected path, the distance be­
plays a bridging role in the strain-hardening stage to build conductive tween CFs is larger than the critical distance and they are wrapped by
networks. SNR depends on the perfection of the conductive networks, insulating matrix, which can be represented by a series of resistor (R2)
the more the conductive fillers, the greater the SNR. The linearity de­ and capacitor (C2). In insulated path, the dosage of CF is very small, and
creases with the increase of the dosage of CF, but it is not affected by the the insulation matrix can be represented by the capacitor (C1). The
dosage of NCB. A reasonable combination of the dosages of CF and NCB model was also analyzed by Nyquist plots under the AC frequency was
can meet the requirements of sensitivity, linearity and SNR, and they can from 50 Hz to 100 KHz to verify the validity of the model [25]. And this
play a positive synergistic effect to improve the sensing properties of model is also consistent with the results in Refs [34] and [42].
material [38]. Based on the results, CF of 0.6 vol% and NCB of 1.0 wt% In order to establish the relationship between the impedance and
are determined as an optimal ratio. Based on the ratio, the piezoresistive spacing of conductive fillers. A universal tunneling effect formula pro­
effect of the material was analyzed in this paper. posed by Simmons [43] was applied in this paper, as shown in Eq. (6).
[ ] [ ]
4. Piezoresistive effect model J = 3(2mφ)1/2 /2S (e/h)2 Vexp − (4πS/h)(2mφ)1/2 (6)

The establishment of a quantitative relationship between material where J is the current density, Am− 2, m and e are the mass and charge
deformation and resistivity, namely, piezoresistive effect, is the basis for of an electron respectively, kg and C, h is Planck constant, φ is the barrier
its application in health monitoring [6]. The piezoresistive effect should height, eV, S is the width of filler spacing, nm, and V is the voltage, V.
be established from the conduction theory. At present, there are two According to the Eq. (6), the relationship between impedance (|Z|, Ω)
complete theories, namely percolation theory and tunnel effect theory. and spacing (S) is determined, as shown in Eq. (7).
The percolation theory mainly explains the relationship between the V 2Sh2
dosage of filler and resistivity at a macro scale, but fails to explain the |Z| = = [ ] (7)
J 3(2mφ)1/2 e2 exp − (4πS/h)(2mφ)1/2
nature of the conductivity [39]. While tunnel effect theory explains the
relationship between filler spacing and resistivity at the micro scale by
Then, Eqs. (1) and (2) are substituted into Eq. (7), and the rela­
quantum mechanics. According to the tunnel effect theory, the con­
tionship between FCI and S is obtained, as shown in Eq. (8).
ductivity of material is not due to the direct contact of conductive fillers, [ ]
⎧ ⎫
but when the spacing between conductive fillers is within a certain ⎨St exp − (4π S0 /h)(2mφ)1/2 ⎬
range, the electrons can cross the barrier and form an electric current FCI = [ ] − 1 × 100% (8)
⎩S exp − (4πS /h)(2mφ)1/2 ⎭
under the external electric field [40]. In practice, as long as the material 0 t

is deformed, the spacing of the conductive filler will be changed,

where
resulting in changes of the resistivity [41]. Therefore, in this paper,
based on the tunnel effect theory, the relationship between the change of St = S0 + ΔS
conductive filler spacing and strain was analyzed, and the piezoresistive
St and S0 are the filler spacing at time t and the initial time. The
effect model of material was established, and the piezoresistive effects
relationship between FCI and S is shown in Eq. (9).

Fig. 12. The distribution of CF and the equivalent circuit model.

9
J. Han et al. Construction and Building Materials 323 (2022) 126390

⎧ ⎫
⎨ ⎬ |Z1| and |Z2| are greater than R1 in the equivalent circuit model.
1 ΔS
FCI = [ ]+ [ ]− 1 Therefore, |Ztotal| can be approximately equal to R1 numerically in this
⎩exp − (4πΔS/h)(2mφ)1/2 S exp − (4π ΔS/h)(2mφ)1/2 ⎭
0 paper.
×100% In the tension process, the distance between the fillers in the branch
“R1” increases and the resistance increases, so the total impedance in­
(9)
creases. In the compression process, branch “R1” has complete conduc­
According to the distribution of CF in Fig. 10, the relationship be­ tive paths, and compression has no effect on the resistance. But in the
tween ΔS and strain is obtained, as shown in Eq. (10). semi-connected circuit, the spacing of some fillers decreases and
become new connected circuits “R1”. An increase in the number of
εlcosθ
ΔS = (10) parallel resistors results in a decrease in the total resistivity. Thus, the
L
total impedance decreases under compression process.
where ε is the strain along the direction of the force, l is the length of In order to further determine the values of A and B in Eq. (12), the
CF and L is the effective length of the specimen. Finally, the relationship test results of specimen with W/B of 0.3, CF of 0.6 vol% and NCB of 1.0
between FCI and ε is obtained, as shown in Eq. (11). wt% under tension and compression tests were analyzed, and the results
{
1 Bε
} are obtained as shown in Fig. 13.
FCI = + − 1 × 100% (11) It can be seen from Fig. 13 that the test results are in good agreement
exp[Aε] exp[Aε]
with the formula results and the R2 are 0.993 and 0.958, respectively. It
where shows that the model can describe the relationship between strain and
(
lcosθ
) impedance change well. The impedance of specimens increases with the
A = − (4π/h)(2mφ)1/2 increase of tensile strain and decreases with the increase of compressive
L
strain. The parameters A is equal to 15.578 ± 0.099, B is equal to
lcosθ − 0.015 ± 0.001 in tension process, and A is equal to − 59.062 ± 4.136, B
B=
S0 L is equal to 1.182 ± 0.232 in compression process.
In order to further verify the accuracy of the model, the new test
In the matrix, m, h, φ, l, L and S0 are the constants of the properties of results of another group of specimens with the same mix ratio were
the material. Since the elastic modulus and strength of CF are much compared, and the results are obtained as shown in Fig. 14. It can be
higher than that of cement matrix, it can be considered that CF does not seen from Fig. 14 that the model is in good agreement with the test
rotate and slip during the test, and θ is also considered to be a constant. It results and can accurately describe the relationship between strain and
should be noted that the formula is obtained through theoretical anal­ impedance change of materials in the process of tension and
ysis, and specific values of material properties need to be determined compression.
through experiments. The research work in this paper was carried out in a laboratory
environment, and the sensitivity, linearity and SNR of sensing properties
should be redetermined before practical application. Although the
4.2. Analysis and validation of piezoresistive effect model model is obtained by simplifying the distribution of conductive fillers,
the experimental results demonstrate that the model has good validity
According to the equivalent circuit model in Fig. 12, the calculation and reliability, and the model can provide some help for the application
formula of the total impedance is shown in Eq. (12). of self-sensing materials.
1 1 1 1
= + + (12) 5. Conclusions
|Ztotal | R1 |Z2 | |Z1 |
where |Ztotal| is the total impedance, |Z1| is the capacitive reactance In this paper, the effects of W/B, the dosages of CF and NCB on the
of the insulated path, |Z2| is the impedance of the semi-connected path compressive strength, tensile ductility and sensing properties of HDCC
and R1 is the resistance of the connected path. were analyzed by experimental study. The strength and tensile ductility
In a parallel circuit, the total resistance is always less than that of the under different W/B were analyzed, and the sensitivity, SNR and
branch. Since the resistivity of the matrix is much higher than that of CF,

140 2
Fraction change in impedance (%)
Fraction change in impedance (%)

120 FCI in experiment 0

Formula curve FCI in experiment
-2 Formula curve
100
-4
80 FCI=1/exp(15.578ε)
FCI=1/exp(1.820ε)
-0.015ε/exp(15.578ε)-1 -6
-59.062ε/exp(1.820ε)-1
60 R2=0.993
-8 R2=0.958
40
-10
20 -12

0 -14
1 2 3 4 5 6 7 0.1 0.2 0.3 0.4 0.5
Strain (%) Strain (%)

(a) Tension (b) Compression

Fig. 13. Test and formula results of the relationship between strain and FCI.

10
J. Han et al. Construction and Building Materials 323 (2022) 126390

140 2

Fraction change in impedance (%) FCI in specimen 1 0

120 FCI in specimen 1

Fraction change in impedance (%)

FCI in specimen 2
FCI in specimen 3 FCI in specimen 2
-2
100 FCI in specimen 3
Formula curve
-4 Formula curve
80
-6
60
-8

40 -10

20 -12

-14
0 0.0 0.1 0.2 0.3 0.4 0.5
1 2 3 4 5 6 7
Strain (%) Strain (%)

(a) Tension (b) Compression

Fig. 14. Test and formula checking results of the relationship between strain and FCI.

linearity of sensing properties with different conductive filler ratio were Declaration of Competing Interest
compared. Finally, based on the tunnel effect theory, the piezoresistive
effect model was established by theoretical derivation, and the validity The authors declare that they have no known competing financial
of the model was verified by experimental results. The main conclusions interests or personal relationships that could have appeared to influence
are as follows: the work reported in this paper.

(1) The W/B, dosages of CF and NCB have important effects on the Acknowledgements
compressive strength. The higher the W/B, the lower the
strength, and the high dosage of CF leads to a decrease in This study was funded by National Natural Science Foundation of
strength. With the increase of the dosage of NCB, the material China (No. 52130210 and No. 51778131).
strength increases first and then decreases, and the appropriate
dosage can play a filling role and improve the compressive References
strength.
(2) The high ductility of the material is mainly reflected in its high [1] V.C. Li, S.X. Wang, C. Wu, Tensile strain-hardening behavior of polyvinyl alcohol
engineered cementitious composite (PVA-ECC), Aci Mater J 98 (6) (2001)
tensile ductility. The ultimate tensile strain of the proposed HDCC 483–492.
can be as high as 3.5% with the W/B of 0.3, indicating that it is [2] V.C. Li, On Engineered Cementitious Composites (ECC) A Review of the Material
feasible to develop self-sensing HDCC with the addition of hybrid and Its Applications, J. Adv. Concr. Technol. 1 (3) (2003) 215–230.
[3] J. Zhang, C.K.Y. Leung, Y. GaoGao, Simulation of crack propagation of fiber
fibers, i.e., PVA fiber and carbon fiber. reinforced cementitious composite under direct tension, Eng. Fract. Mech. 78 (12)
(3) The CF and NCB have important effects on the sensing properties. (2011) 2439–2454.
It was found that CF is the main body of the conductive networks [4] G. Fischer, V.C. Li, Effect of matrix ductility on deformation behavior of steel-
reinforced ECC flexural members under reversed cyclic loading conditions, ACI
due to its high aspect ratio, while NCB plays an auxiliary role as Struct. J. 99 (6) (2002) 781–790.
active contact. Notably, the conductive network maintained its [5] B.A. Canbolat, G.J. Parra-Montesinos, J.K. Wight, Experimental study on seismic
continuality due to the bridge effect of CF in tension process. For behavior of high-performance fiber-reinforced cement composite coupling beams,
ACI Struct. J. 102 (1) (2005) 159–166.
the sensing index, CF can affect the sensitivity and linearity,
[6] R. Ranade, J. Zhang, J.P. Lynch, V.C. Li, Influence of micro-cracking on the
while the addition of NCB can improve the SNR. composite resistivity of Engineered Cementitious Composites, Cem. Concr. Res. 58
(4) Based on the simplified distribution of conductive fillers, the (2014) 1–12.
equivalent circuit model of the material was established. Based [7] C. Pu-Woei, D.D.L. Chung, Carbon-Fiber-Reinforced Concrete as an Intrinsically
Smart Concrete for Damage Assessment during Dynamic Loading, J. Am. Ceram.
on the tunnel effect theory, the relationship between the spacing Soc. 78 (3) (1995) 816–818.
of conductive fillers and the change fraction of specific imped­ [8] D.D.L. Chung, Cement reinforced with short carbon fibers: a multifunctional
ance (FCI) was established. The piezoresistive effect model was material, Compos Part B-Eng 31 (6–7) (2000) 511–526.
[9] B. Han, X. Yu, E. Kwon, A self-sensing carbon nanotube/cement composite for
also established by combining the filler distribution and the traffic monitoring, Nanotechnology 20 (44) (2009) 445–501.
equivalent circuit model. [10] H.W. Deng, H.L. Li, Assessment of self-sensing capability of Carbon Black
(5) By analyzing the experimental results of the relationship between Engineered Cementitious Composites, Constr Build Mater 173 (2018) 1–9.
[11] N. Banthia, S. Djeridane, M. Pigeon, Electrical-Resistivity of Carbon and Steel
the FCI and strain, the parameters of the piezoresistive effect Micro-Fiber Reinforced Cements, Cem. Concr. Res. 22 (5) (1992) 804–814.
model during compression and tension were proposed. The val­ [12] S. Yehia, C.Y. Tuan, Conductive concrete overlay for bridge deck deicing, Aci Mater
idity of the model was also verified by experimental results. J 96 (3) (1999) 382–390.
[13] B.G. Han, B.Z. Han, X. Yu, Effects of the content level and particle size of nickel
powder on the piezoresistivity of cement-based composites/sensors, Smart Mater
Struct 19 (6) (2010).
[14] B.G. Han, B.Z. Han, J.P. Ou, Experimental study on use of nickel powder-filled
Portland cement-based composite for fabrication of piezoresistive sensors with
CRediT authorship contribution statement
high sensitivity, Sensor Actuat a-Phys 149 (1) (2009) 51–55.
[15] F.J. Baeza, O. Galao, E. Zornoza, P. Garcés, Multifunctional cement composites
Jinsheng Han: Conceptualization, Methodology. Jinlong Pan: strain and damage sensors applied on reinforced concrete (RC) structural elements,
Project administration, Writing – review & editing. Jingming Cai: Materials (3) (2013) 841–855.

Visualization, Investigation.

11
J. Han et al. Construction and Building Materials 323 (2022) 126390

[16] G. Yıldırım, O. Öztürk, A. Al-Dahawi, A. Afşın Ulu, M. Şahmaran, Self-sensing [29] Y. Hong, Z. Li, G. Qiao, J. Ou, W. Cheng, Pressure sensitivity of multiscale carbon-
capability of Engineered Cementitious Composites: Effects of aging and loading admixtures–enhanced cement-based composites, Nanomater. Nanotechnol. 8
conditions, Constr Build Mater 231 (2020), 117132. (2018) 1–8.
[17] Y. Ding, Z. Chen, Z. Han, Y. Zhang, F. Pacheco-Torgal, Nano-carbon black and [30] S. Ding, Y.-W. Wang, Y.-Q. Ni, B. Han, Structural modal identification and health
carbon fiber as conductive materials for the diagnosing of the damage of concrete monitoring of building structures using self-sensing cementitious composites,
beam, Constr Build Mater 43 (6) (2013) 233–241. Smart Mater Struct 29 (5) (2020).
[18] J.P. Ou, B.G. Han, Piezoresistive Cement-based Strain Sensors and Self-sensing [31] B. Han, K. Zhang, Y. Xun, E. Kwon, J. Ou, Electrical characteristics and pressure-
Concrete Components, J. Intell. Mater. Syst. Struct. 20 (3) (2009) 329–336. sensitive response measurements of carboxyl MWNT/cement composites, Cem.
[19] S. Wen, D.D.L. Chung, Partial replacement of carbon fiber by carbon black in Concr. Compos. 34 (6) (2012) 794–800.
multifunctional cement–matrix composites, Carbon 45 (3) (2007) 505–513. [32] D.K. Hardy, M.F. Fadden, M.J. Khattak, A. Khattab, Development and
[20] L.F. Shi, Y. Lu, R.Q. Guan, Detection of crack development in steel fibre engineered characterization of self-sensing CNF HPFRCC, Mater. Struct. 49 (12) (2016)
cementitious composite using electrical resistivity tomography, Smart Mater. 5327–5342.
Struct. 28 (12) (2019) 14. [33] T. Kanda, V.C. Li, Multiple Cracking Sequence and Saturation in Fiber Reinforced
[21] L.F. Shi, Y. Lu, Y. Bai, Mechanical and electrical characterisation of steel fiber and Cementitious Composites, Concrete Research and Technology 9 (2) (1998) 19–33.
carbon black engineered cementitious composites, in: W.K. Chiu, S. Galea, A. Mita, [34] X. Li, M. Li, Multifunctional self-sensing and ductile cementitious materials, Cem.
N. Takeda (Eds.), Structural Health Monitoring - from Sensing To Diagnosis and Concr. Res. 123 (2019), 105714.
Prognosis, Elsevier Science Bv, Amsterdam, 2017, pp. 325–332. [35] K. Tosun Felekoglu, E. Godek, Role of matrix structure and flaw size distribution
[22] T.C. Hou, J.P. Lynch, Conductivity-based strain monitoring and damage modification on deflection hardening behavior of polyvinyl alcohol fiber reinforced
characterization of fiber reinforced cementitious structural components, in: M. engineered cementitious composites (PVA-ECC), J. Cent. South Univ. 26 (12)
Tomizuka (Ed.), Smart Structures and Materials 2005: Sensors and Smart (2019) 3279–3294.
Structures Technologies for Civil, Mechanical, and Aerospace, Pts 1 and 22005, pp. [36] X.W. Zhang, Y. Pan, Q. Zheng, X.S. Yi, Piezoresistance of conductor filled insulator
419-429. composites, Polym. Int. 50 (2) (2001) 229–236.
[23] A. Al-Dahawi, M.H. Sarwary, O. Ozturk, G. Yildirim, A. Akin, M. Sahmaran, [37] P.W. Chen, D.D.L. Chung, Improving the electrical conductivity of composites
M. Lachemi, Electrical percolation threshold of cementitious composites possessing comprised of short conducting fibers in a nonconducting matrix: The addition of a
self-sensing functionality incorporating different carbon-based materials, Smart nonconducting particulate filler, J. Electron. Mater. 24 (1) (1995) 47–51.
Mater. Struct. 25 (10) (2016) 15. [38] A. Hussain, Y. Ding, G. Liu, A. Naqi, Study on self-monitoring of multiple cracked
[24] A. Al-Dahawi, G. Yildirim, O. Ozturk, M. Sahmaran, Assessment of self-sensing concrete beams with multiphase conductive materials subjected to bending, Smart
capability of Engineered Cementitious Composites within the elastic and plastic Mater Struct 28 (9) (2019) 10.
ranges of cyclic flexural loading, Constr Build Mater 145 (2017) 1–10. [39] C.W. Nan, Y. Shen, J. Ma, Physical Properties of Composites Near Percolation,
[25] J. Han, J. Cai, J. Pan, Y. Sun, Study on the conductivity of carbon fiber self-sensing Annu. Rev. Mater. Sci. 40 (1) (2010) 131–151.
high ductility cementitious composite, Journal of Building Engineering 43 (2021), [40] A.B. Kaiser, Y.W. Park, Current-voltage characteristics of conducting polymers and
103125. carbon nanotubes, Synth. Met. 152 (1–3) (2005) 181–184.
[26] L. Hui, H.G. Xiao, Y. Jie, J. Ou, Microstructure of Cement Mortar with Nano- [41] X.-W. Zhang, Y. Pan, Q. Zheng, X.-S. Yi, Time dependence of piezoresistance for the
Particles, Compos. B Eng. 35 (2) (2004) 185–189. conductor-filled polymer composites, J. Polym. Sci., Part B: Polym. Phys. 38 (21)
[27] A. Al-Dahawi, O. Ozturk, F. Emami, G. Yildirim, M. Sahmaran, Effect of mixing (2000) 2739–2749.
methods on the electrical properties of cementitious composites incorporating [42] M. Cabeza, P. Merino, A. Miranda, X.R. Nóvoa, I. Sanchez, Impedance
different carbon-based materials, Constr Build Mater 104 (2016) 160–168. Spectroscopy Study of Hardened Portland Cement Paste, Cem. Concr. Res. 32 (6)
[28] B. Han, J. Ou, Embedded piezoresistive cement-based stress/strain sensor, Sensor (2002) 881–891.
Actuat a-Phys 138 (2) (2006) 294–298. [43] J.G. Simmons, Generalized Formula for the Electric Tunnel Effect between Similar
Electrodes Separated by a Thin Insulating Film, J. Appl. Phys. 34 (6) (1963)
1793–1803.

12

View publication stats