Home Search Collections Journals About Contact us My IOPscience

Cementitious Composites Engineered with Embedded Carbon Nanotube Thin Films for

Enhanced Sensing Performance

This content has been downloaded from IOPscience. Please scroll down to see the full text.

2015 J. Phys.: Conf. Ser. 628 012042

([Link]

View the table of contents for this issue, or go to the journal homepage for more

Download details:

IP Address: [Link]
This content was downloaded on 30/03/2017 at 21:53

Please note that terms and conditions apply.

You may also be interested in:

Validity and reliability of estimated modulus of elasticity of cementitious materials

N U Kockal

Influence of electrified surface of cementitious materials on structure formation of hardened

cement paste
A Alekseev and A Gusakov

Monitoring of cement hydration using embedded piezoelectric transducers

Lei Qin and Zongjin Li

Reinforced cementitous composite with in-situ shrinking microfibers

Eric S Kim, Jason K Lee, Patrick C Lee et al.

Embedded white light interferometer fibre optic strain sensor for monitoring crack-tip opening in
concrete beams
Libo Yuan and Fahard Ansari

Quantitative characterization of the microstructure of fresh cement paste via random packing of
polydispersed Platonic cement particles
W X Xu and H S Chen

A probabilistic method for determining the volume fraction of pre-embedded capsules in self-healing
materials
Zhong Lv and Huisu Chen

Sealing of cracks in cement using microencapsulated sodium silicate

P Giannaros, A Kanellopoulos and A Al-Tabbaa
11th International Conference on Damage Assessment of Structures (DAMAS 2015) IOP Publishing
Journal of Physics: Conference Series 628 (2015) 012042 doi:10.1088/1742-6596/628/1/012042

Cementitious Composites Engineered with Embedded Carbon

Nanotube Thin Films for Enhanced Sensing Performance

Kenneth J. Loh* and Jesus Gonzalez

Department of Civil & Environmental Engineering, University of California, Davis
One Shields Avenue, Davis, CA 95616-5294, USA

E-mail: kjloh@[Link]

Abstract. Cementitious composites such as concrete pavements are susceptible to different

damage modes, which are primarily caused by repeated loading and long-term deterioration.
There is even greater concern that damage could worsen and occur more frequently with the
use of heavier vehicles or new aircraft carrying greater payloads. Thus, the objective of this
research is to engineer cementitious composites with capabilities of self-sensing or detecting
damage. The approach was to enhance the damage sensitivity of cementitious composites by
incorporating multi-walled carbon nanotubes (MWNT) as part of the mix design and during
casting. However, as opposed to directly dispersing MWNTs in the cement matrix, which is the
current state-of-art, MWNT-based thin films were airbrushed and coated onto sand particles.
The film-coated sand was then used as part of the mix design for casting mortar specimens.
Mortar specimens were subjected to compressive cyclic loading tests while their electrical
properties were recorded simultaneously. The results showed that the electrical properties of
these cementitious composites designed with film-coated sand exhibited extremely high strain
sensitivities. The electrical response was also stable and consistent between specimens.

1. Introduction
Multifunctional materials have gained considerable attention over the last few decades. In particular, a
material, whether it is homogeneous or a composite, is multifunctional when it has been specifically
encoded with unique properties that allow it to perform multiple engineering functionalities [1]. Such
multifunctional materials are particularly desirable for structural applications. For instance, while
infrastructure materials (e.g., steel, concrete, and fiber-reinforced polymer (FRP) composites) are
designed to carry loads, these structures can also experience damage during their operational service
lifetime. As an example, FRPs are known to suffer from matrix cracking, debonding, and delamination
that are not only invisible to the naked eye but can also propagate to cause catastrophic structural
failure. Thus, multifunctional FRP composites that can bear loads and sense damage are currently
being developed [2, 3].

Another even more widely used infrastructure material is cementitious composites, in which they are
often employed for civil infrastructure applications such as for bridges, buildings, foundations, and
pavements. While the use of concrete as load-bearing materials can be traced back several centuries
ago, engineering them to be multifunctional occurred more recently. In fact, in the early 1990s, Chen
and Chung [4] showed that the electrical properties of cementitious composites could be engineered to
be sensitive to damage by adding electrically conductive, short, carbon fibers to mortar during casting.

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
11th International Conference on Damage Assessment of Structures (DAMAS 2015) IOP Publishing
Journal of Physics: Conference Series 628 (2015) 012042 doi:10.1088/1742-6596/628/1/012042

Because the electrical resistivity (or conductivity) of the composite was calibrated and varied in
response to applied stress/strain or damage, one could use the cementitious composite as both a load-
bearing structure and a sensor.

A variety of research aimed at developing self-sensing or smart concrete was investigated over the
past few decades. The conductive additives embedded in the cementitious matrix ranged from using
carbon fibers (as mentioned earlier) [5], steel fibers [6], carbon black [7], to, more recently, carbon
nanotubes (CNT) [8-10]. While different advantages and unique attributes were reported for each of
these cement composites enhanced with different types of conductive additives, their general approach
to designing and casting remained the same. The procedure was to disperse the conductive fibers
during mixing (i.e., either by stirring the fibers with other components in a dry state or by pre-
dispersing the additives in a solution that was then added to the mix later). For both scenarios, fiber
agglomeration, material uniformity, and consistency could not be easily controlled or guaranteed. In
addition, even if one were able to disperse 0.1% (by weight of cement) of multi-walled carbon
nanotubes (MWNT) in cement (as was used by Han et al. [8]), the increase in raw material costs
would be significant (i.e., an order of magnitude or greater), as compared to the plain concrete.

Thus, the goal of this research is to demonstrate a new paradigm of nanomaterial-enhanced smart
concrete. Instead of undertaking the challenging task of dispersing low concentrations of CNTs in the
cement matrix, nanotubes were successfully dispersed in polymer thin films in previous studies [11,
3]. These nanocomposites exhibit high electrical conductivity, are ultra-thin (i.e., on the order of 1
m), and efficiently uses the smallest quantities of CNTs possible. In that regard, MWNT-films were
deposited and coated onto sand particles, prior to using them for casting mortar specimens. The
rationale was to modify the cement-particle interface to create a more effective electrically conductive
network within the cement matrix. With this technique, MWNTs would remain well dispersed within
the thin film, and yet, the films coated on sand could still make contact with other films/particles to
enhance the bulk cementitious composites electromechanical (or sensing) response. To the authors
best knowledge, this approach of designing self-sensing concrete using embedded particles pre-coated
with nanocomposites has not been reported before. This paper begins with a presentation of the mix
design, casting procedures, and testing protocol. Then, the nanocomposite-enhanced smart concretes
sensing results are presented, discussed, and compared to other results reported in the literature.

2. Experimental Details
The aim of this work was to establish a new, effective, and low-cost method for incorporating nano-
materials (i.e., MWNTs) in mortar specimens so that they could be used as multifunctional materials
for load bearing and self-sensing. First, section 2.1 begins with a discussion of the mortar mix design
and casting procedure. Second, since the mortar mix employed sand pre-coated with an MWNT-based
thin film, section 2.2 describes the film preparation procedure. Finally, a description of the cyclic
compressive tests and the data acquisition strategy are discussed in section 2.3.

2.1. Mix Design and Casting Procedure

Mortar specimens were cast following U.S. Federal Aviation Administration (FAA) rigid concrete
pavement specifications [12], as well as those provided by ASTM C109 for strength testing [13]. Here,
Type I/II Portland cement mixed with 25 wt.% (of total cementitious material) Type F ground
granulated blast furnace slag (GGBFS) was used. A 2.75:1 sand-to-cement ratio was employed, as
stated in ASTM C109 [13]. The sand was crushed granite with a density of 2.71 g/cm3 and a minimum
particle size of 0.149 mm. The water-to-cement (w/c) ratio was 0.52, and 2.9 mL/kg of cement of
Advacast superplasticizer was also incorporated as a water-reducing agent.

2
11th International Conference on Damage Assessment of Structures (DAMAS 2015) IOP Publishing
Journal of Physics: Conference Series 628 (2015) 012042 doi:10.1088/1742-6596/628/1/012042

(a) (b)
Figure 1. Pictures of the (a) pristine and (b) MWNT-based nanocomposite-coated sand are shown.

The procedure for casting the mortar specimens began with mixing all the various constituents using a
Hobart N-50 mixer. First, the water, superplasticizer, and cementitious materials were placed in a
stainless steel mixing bowl, and the contents were mixed at low speed (60 rpm) for 30 s. Second, the
nanocomposite-coated sand (as will be discussed later in section 2.2) was gradually poured into the
bowl over the course of 30 s and while the mixer was operating at 60 rpm. Upon doing so, the contents
were mixed at medium speed (124 rpm) for 1 min to obtain an even mixture. Third, after mixing was
complete, any materials adhered to the walls of the bowl were scrapped off, and the entire mixture was
set aside for ~15 s prior to pouring them into the molds for casting.

The molds that were used in this study were 555 cm3 (222 in3) cubes. The molds were lubricated
using WD-40, and the seams were sealed with petrolatum. Since the objective of this study was to
measure how the electrical properties of these specimens varied in response to applied strains and
stresses, electrodes were also incorporated during specimen casting. Here, 4.56.5 cm2 (1.752.25 in2)
copper mesh electrodes were cut and placed close to two opposite ends of the mold. They were
inserted such that the mesh electrodes covered the entire specimens cross-section with a portion of it
protruding out for electrical measurement purposes during load testing. Then, with the electrodes in
place and the molds ready, the mortar mixture was scooped into each mold until it was half-full. The
mold was gently vibrated for 10 s before it was filled completely and then vibrated for another 10 s.
Next, the filled molds were sealed with plastic wrap and left for 24 h in a sealed container at room
temperature. Finally, the specimens were de-molded, submerged in a limewater solution, and cured for
28 days prior to testing. A total of five mortar specimens were casted as part of this study.

2.2. Nanocomposite Fabrication and Sand Coating

As mentioned earlier, the sand used for mortar casting was pre-coated with an MWNT-based thin film.
Film deposition onto the sand was accomplished via airbrushing and by following procedures
developed in a previous study reported in Mortensen et al. [11]. In general, MWNTs (SouthWest
NanoTechnologies) were dispersed via sonication in a 2 wt.% poly(sodium 4-styrenesulfonate) (PSS)
solution. N-methyl-2-pyrrolidinone (NMP) was added to the MWNT-PSS solution and then mixed
with a latex solution. This procedure yielded MWNT-based inks that could then be spray coated onto
virtually any surface using a Paasche airbrush [3].

Prior to spray coating the sand (figure 1a), sand was dried in a vacuum oven for 24 h. The gradation of
sand used was reported in Wong et al. [9]. Upon drying, sand was poured onto a flat surface and
placed in a fume hood. It should be mentioned that the sand was spread to form a thin layer no thicker

3
11th International Conference on Damage Assessment of Structures (DAMAS 2015) IOP Publishing
Journal of Physics: Conference Series 628 (2015) 012042 doi:10.1088/1742-6596/628/1/012042

Figure 2. A dried mortar specimen that incorporated nanocomposite-coated sand is shown. The
specimen also had two copper mesh electrodes embedded on its two opposite ends so that its
electrical resistance could be measured during load testing.

than ~6 mm (~0.25 in). Then, using the MWNT-based ink solution and the airbrush, the thin film was
sprayed onto the surface of the sand by using a sweeping motion while holding the spray nozzle ~30
cm away from the target. Once one coat was complete, the sand was then mixed in place using a
plastic spatula so as to reveal any underlying sand that had not been coated with the MWNT-based
ink. This procedure was repeated for three times. Upon doing so, the sand was left to air-dry for 1 h,
followed by additional drying in a vacuum oven set to 50 C for 30 min. A picture of the coated sand
is shown in figure 1b; note that the color of the sand became dark-gray after film deposition.

2.3. Electromechanical Characterization

The 555 cm3 (222 in3) mortar cubes were subjected to compressive cyclic load tests so that their
electromechanical (sensing) response could be characterized. An MTS load frame equipped with a 407
controller was employed for load testing. Each load test was conducted using the following procedure.
Upon mounting the specimen in the load frame, a 5,000 N preload was applied. Then, the load frame
was commanded to execute a 0.1 Hz cyclic-compressive load pattern (i.e., five to 10 cycles) to a peak
strain of ~0.45% during the loading ramp, followed by unloading until the preload was released. The
tests were displacement-controlled. Applied load and displacement were sampled at 55 Hz and were
recorded using a customized LabVIEW program. In addition to measuring the specimens load-
displacement response, an Agilent 34401A digital multimeter (DMM) also recorded the electrical
resistance of each specimen (sampling rate: 2.8 Hz). Resistance data was also recorded and time-
synchronized using the same LabVIEW client. It should be mentioned that the Agilent DMM was
controlled to record ~20 s of baseline resistance prior to the start of each load test. This was conducted
to minimize any electrical polarization (i.e., resistance drift) inherent to these cementitious composites.

However, immediately prior to test, each mortar specimen was removed from the limewater solution
after 28 days of curing. The saturated specimen was pat dried and then dried in a vacuum oven
operating at 50 C for 1 h. Figure 2 shows a picture of a pristine mortar specimen. Then, the specimen
was subjected to three consecutive multi-cycle compressive tests. The time for testing was limited to
be between 2 to 3 min for minimizing specimen drying during each test. Then, the specimen was
placed back in the vacuum oven for an additional 5 h of drying so that the total drying time was 6 h.
Upon drying, the specimen was subjected to three more consecutive cyclic compressive tests. This test
procedure was repeated for all of the five different mortar specimens casted for this study.

4
11th International Conference on Damage Assessment of Structures (DAMAS 2015) IOP Publishing
Journal of Physics: Conference Series 628 (2015) 012042 doi:10.1088/1742-6596/628/1/012042

Figure 3. A mortar specimen subjected to 1 h of vacuum oven drying was subjected to compressive
cyclic loading. Its resistance time history was overlaid with the applied strain pattern.

3. Results and Discussions

As mentioned in section 2.3, each of the five mortar specimens was subjected to a total of six multi-
cycle compressive load tests. The first three test results corresponded to the case when the mortar was
dried for 1 h. These tests considered the response of the specimens in a more saturated or damp state.
In contrast, the latter three tests investigated the electromechanical response of mortar specimens dried
for a total time of 6 h, in which the specimens were presumed to be in a much drier state.

Figure 3 shows a representative result of a nanocomposite-enhanced mortar specimen subjected to

seven cycles of compressive loads to ~0.45%. The plot in figure 3 includes the measured resistance
time history overlaid on top of the applied cyclic compressive strain pattern. From this result, it is
clear that the mortar specimens electrical resistance decreased in tandem with increasing compressive
strains. Its response was linear, as can be observed from the overlay in figure 3. In addition, its
nominal resistance did not appear to drift or vary with time significantly. Overall, the response was
repeatable and stable.

On the other hand, figure 4 shows two representative results for a mortar specimen that was dried for 6
h in the vacuum oven. Similar trends can be observed when comparing figures 3 and 4, such that in
both cases, resistance decreased with greater compressive strains applied. The main difference for
these specimens is that specimens dried for 6 h exhibited more prominent nominal resistance drifts,
that fortunately and typically only occurred during the first of three consecutive load tests (figure 4a).
Afterward, its electromechanical response varied closely with that of the applied strain (figure 4b). It
should be mentioned that such nominal resistance drifts are expected, since electrical polarization in
cementitious composites had been documented in the literature [14].

In order to quantify the nanocomposite-enhanced mortar specimens sensitivity to strains or,

equivalently, to calculate its gage factor, the raw data from figure 4b is used to plot figure 5. Figure 5
shows the normalized change in resistance of the specimen plotted as a function of the applied strains;
note that the x-axis shows negative strains because the applied loading was in compression (similar to
figures 3 and 4). In general, strain sensitivity (S) or gage factor can be calculated using (1):

5
11th International Conference on Damage Assessment of Structures (DAMAS 2015) IOP Publishing
Journal of Physics: Conference Series 628 (2015) 012042 doi:10.1088/1742-6596/628/1/012042

(a) (b)
Figure 4. Nanocomposite-enhanced mortar specimens dried for 6 h in a vacuum oven were
subjected to cyclic compressive tests. (a) While the specimens resistance exhibited sensitivity to
applied strains, nominal resistance drift was observed. (b) However, the nominal resistance drift
plateaued after a short period of time to yield stable electromechanical response.

Figure 5. The normalized change in resistance of the raw data plotted in figure 4b is plotted with
respect to the applied strain.

R
R0
S= (1)

where is strain, R0 is the mortars nominal or unstrained resistance, and R is the instantaneous
change in resistance of the mortar relative to its nominal resistance. Using figure 5, the specimens
strain sensitivity can be obtained from the slope of the linear least-squares regression line fitted to the
data. Upon doing so, the average strain sensitivity computed from all the tests and specimens was
found to be ~118. For reference, the typical gage factor for a strain gage is 2. As compared to another
previous study in which 0.025 wt.% of MWNTs were directly dispersed in the cement matrix of
mortar specimens, the gage factor obtained ranged between 12.2 and 13.5 [9].

6
11th International Conference on Damage Assessment of Structures (DAMAS 2015) IOP Publishing
Journal of Physics: Conference Series 628 (2015) 012042 doi:10.1088/1742-6596/628/1/012042

Nevertheless, while an average gage factor was obtained and reported here, closer examination of
figure 5 shows that these specimens exhibited bilinear sensitivity to applied compressive strains. At
low strains (i.e., up to -0.2%), the sensitivity was extremely high; S is 248 for the data shown in figure
5. When compressive strain exceeded -0.2%, another linear least-squares best-fit line was fitted to the
high strain region. The corresponding gage factor obtained for the dataset shown in figure 5 was 60.

Given that the mortar specimens were subjected to a peak strain of -0.45%, it was unlikely that these
results were due to non-linear response or damage occurring in the specimens, although additional
tests need to be conducted to verify this assumption. The current working hypothesis is that bilinear
strain sensitivity occurred due to a change in the percolated network formed by the MWNT-based thin
film-coated sand particles. It is possible that compression up to -0.25% strains allowed thin film-
coated sand to begin to come into contact with one another, thereby facilitating significantly better
electrical current transport with each additional direct contact formed. After this point, additional
compression would yield less significant changes in particle contacts, thereby reducing the relative
change in resistance to result in lower sensitivities. Previous studies by Mortensen et al. [11] and
Loyola et al. [3] revealed that these MWNT-latex films exhibited piezoresistivity. Another possibility
is that the inherent piezoresistivity of these thin films contributed to the bilinear strain sensitivity
observed in figure 5. Future studies will focus on an in-depth characterization of how these thin films
affected the cement-sand interface from not only an electrical perspective but also on the mechanical
performance of the bulk cementitious composite.

4. Conclusions
In this study, a novel technique for incorporating carbon nanotube fillers for designing multifunctional
cementitious composites was presented. Instead of directly dispersing nanotubes in the cement matrix,
the approach was to pre-coat sand with conductive, multi-walled carbon nanotube-latex thin films. The
nanocomposite-coated sand was then incorporated during mixing and casting of mortar specimens.
Upon casting 555 cm3 (222 in3) nanocomposite-enhanced mortar cube specimens, each specimen
was subjected to multiple cycles of compressive load tests while their electrical resistance was
measured. Specimens were tested after two different drying times, namely after 1 h and 6 h of total
drying at 50 C in a vacuum oven.

Both sets of results showed that the electrical properties of these cementitious composites exhibited
sensitivity to applied compressive strains. Specifically, its electrical resistance decreased in tandem
with increasingly applied compressive strains. For specimens that were dried for 6 h, a slight nominal
resistance drift was observed, but this effect quickly plateaued to produce stable electromechanical or
sensing response. While the average gage factor or strain sensitivity was found to be 118, closer
examination of the results showed that bilinear strain sensitivity was observed. At low strains, the
sensitivity was higher; one specimen yielded a gage factor of 248 up to -0.2% strains. Subsequent
application of compressive loading caused an immediate decrease in gage factor to 60. Future work
will focus on characterizing the effects of the films on the cement-sand interface. This casting
technique will also be scaled up for fabricating larger specimens, as well as for casting concrete
specimens that incorporate large aggregates.

5. Acknowledgements
The U.S. Federal Aviation Administration (FAA) supported this research under cooperative agreement
no. 13-G-017. The authors also thank Prof. John Bolander, Mr. Brandon Dashwood, and Ms. Jennifer
Yasui for their assistance with equipment training/use, specimen casting, and testing. Dr. Rongzong
Wu and Dr. Navneet Garg are also acknowledged for their collaboration and support for this project.

7
11th International Conference on Damage Assessment of Structures (DAMAS 2015) IOP Publishing
Journal of Physics: Conference Series 628 (2015) 012042 doi:10.1088/1742-6596/628/1/012042

6. References
[1] Momoda L A 2004 The future of engineering materials: Multifunction for performance-tailored
structures The Bridge 34 18-21
[2] Boger L, Wichmann M H G, Meyer L O and Schulte K 2008 Load and health monitoring in
glass fibre reinforced composites with an electrically conductive nanocomposite epoxy
matrix Composites Science and Technology 68 1886-1894
[3] Loyola B R, Briggs T M, Arronche L, Loh K J, Saponara V L, O'Bryan G and Skinner J L 2013
Detection of spatially distributed damage in fiber-reinforced polymer composites Structural
Health Monitoring 12 225-240
[4] Chen P-W and Chung D D L 1993 Carbon fiber reinforced concrete for smart structures capable
of non-destructive flaw detection Smart Materials and Structures 2 22-30
[5] Fu X and Chung D D L 1996 Self-monitoring of fatigue damage in carbon fiber reinforced
cement Cement and Concrete Research 26 15-20
[6] Teomete E and Kocyigit O I 2013 Tensile strain sensitivity of steel fiber reinforced cement
matrix composites tested by split tensile test Construction and Building Materials 47 962-
968
[7] Ou J and Han B 2009 Piezoresistive cement-based strain sensors and self-sensing concrete
components Journal of Intelligent Material Systems and Structures 20 329-336
[8] Han B, Yu X and Kwon E 2009 A self-sensing carbon nanotube/cement composite for traffic
monitoring Nanotechnology 20 1-5
[9] Wong I, Loh K J, Wu R and Garg N 2014 Sensing properties of mortar incorporating ultra-low
concentrations of carbon-based conductive additives. In: Proceedings of the 5th Asia-Pacific
Workshop on Structural Health Monitoring, (Shenzhen, China pp 1-10
[10] Wille K and Loh K J 2010 Nano-engineering ultra-high performance concrete with multi-walled
carbon nanotubes Journal of the Transportation Research Board 2142 119-126
[11] Mortensen L P, Ryu D, Zhao Y and Loh K J 2013 Rapid assembly of multifunctional thin film
sensors for wind turbine blade monitoring Key Engineering Materials 569-570 515-522
[12] O'Donnell M J 2011 Standards for specifying construction of airports. In: Advisory Circular:
U.S. Department of Transportation, Federal Aviation Administration)
[13] ASTM standard c 109/c 109m - 02. In: Standard Test Method for Compressive Strength of
Hydraulic Concrete Mortars (Using 2-in. (or 50-mm) Cube Specimens),
[14] Wen S and Chung D D L 2001 Electric polarization in carbon fiber-reinforced cement Cement
and Concrete Research 31 141-147