Construction and Building Materials 168 (2018) 459–470

Contents lists available at ScienceDirect

Construction and Building Materials

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

On the mechanical and shrinkage behavior of cement mortars reinforced

with carbon nanotubes
A. Hawreen a,⇑, J.A. Bogas a, A.P.S. Dias b
a
CERIS-ICIST, DECivil, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1049-001 Lisbon, Portugal
b
LAETA, IDMEC, Instituto Superior Técnico, Universidade de Lisboa, Av. Rovisco Pais, 1049-001 Lisbon, Portugal

h i g h l i g h t s

CNTs aspect ratio and dispersion process highly affect the CNT-mortar performance.

CNTs are more effective in mortars with low w/c and early testing ages.

Shrinkage and fracture toughness of mortars highly improve with CNT incorporation.

Increment in compression, flexion and elastic modulus of CNT-mortars.

CNTs provide efficient crack bridging and enhance quality of aggregate-paste ITZ.

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

Article history: This paper investigates the reinforcement efficiency of mortars produced with different types of carbon
Received 8 November 2017 nanotubes (CNTs) in terms of their mechanical strength, ultrasonic pulse velocity, elastic modulus, frac-
Received in revised form 28 January 2018 ture toughness, shrinkage and microstructure. The incorporation of 0.05–0.1% of CNTs was able to reduce
Accepted 21 February 2018
the early shrinkage up to 62% and increase the flexural strength and fracture energy up to 33% and 65%,
Available online 27 February 2018
respectively. Well-dispersed CNTs could improve the post-cracking resistance of mortars, being able to
effectively bridge microcracks up to around 1 lm wide. Microscopic analysis shows that CNTs can poten-
Keywords:
tially improve the quality of the cement matrix in the aggregate-paste interfacial transition zone.
Carbon nanotube
Mortar
Ó 2018 Elsevier Ltd. All rights reserved.
Mechanical strength
Toughness
Shrinkage
Microstructure

1. Introduction research concerning cementitious composites reinforced with

The first polymer nanocomposite using CNTs as fillers was
reported in 1994 by Ajayan [1]. Since then, CNT/polymer [2,3]
and CNT/metal [4,5] composites have been widely investigated
for Young’s modulus, tensile strength, fracture toughness and
strain/stress sensing. The positive effect of the incorporation of
CNTs in the mechanical properties of ceramic matrices has also
been reported [6–8]. These findings suggest that cementitious
materials can also benefit from the incorporation of CNTs due to
their significant mechanical properties [9]. In fact, some studies
have shown the contribution of CNTs towards improving the initial
shrinkage [6,10], mechanical properties [6,11], durability [12] and
microstructure [6,7,13] of cementitious matrices. Nonetheless,
⇑ Corresponding author.
E-mail addresses: hawreen.a@gmail.com (A. Hawreen), abogas@civil.ist.utl.pt
(J.A. Bogas), apsoares@tecnico.ulisboa.pt (A.P.S. Dias).
CNTs is still recent.
Some research investigations on mechanical characterization
have been recently conducted, especially concerning the incorpo-
ration of up to 1% of CNTs by weight of binder in cement pastes
[6,7,10,11,13,14]. Li et al. [7] investigated the mechanical beha-
viour of cement pastes reinforced with 0.4–0.5% of surface modi-
fied CNTs and reported 19–25% improvements in the flexural and
compressive strength. Incorporation of pristine CNTs led to lower
compressive strengths as compared to plain mixes. The low effec-
tiveness of CNTs in mechanical strength reinforcement of cement
pastes [11,15] or mortars [16] was also reported by other research-
ers. Nevertheless, some authors could attain strength improve-
ments as high as 35% when CNTs were well dispersed in the
cement matrix [6,17,18]. Konsta et al. [10] found up to 35% lower
shrinkage in cement pastes with 0.028–0.048% of CNTs as com-
pared to plain pastes. A greater contribution of CNTs was found
in the toughness of cement pastes and mortars. Zou et al. [19], Li

https://doi.org/10.1016/j.conbuildmat.2018.02.146
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
460 A. Hawreen et al. / Construction and Building Materials 168 (2018) 459–470
et al. [20] and Stynoski et al. [21] found that the addition of 0.033–
0.125% of CNTs resulted in fracture load, fracture toughness index
and crack mouth opening displacement up to 51%, 19% and 52%
higher than those of specimens without CNTs, respectively.
To sum up, depending on the type, amount and dispersion pro-
cedure of CNTs, as well as the studied property, different efficiency
of CNTs-reinforced cement based materials has been reported in
literature. Further research is thus necessary to better understand
the CNTs potential in reinforcing cementitious matrices. In addi-
tion, studies have been essentially focused on the reinforcement
of cement pastes, with less knowledge about the efficiency of CNTs
in mortars and concretes, where aggregates are also present. That
said, this paper aims the mechanical behaviour of cement mortars
with different types of CNTs and a wide range of distinct composi-
tions. Mortars were characterized in terms of mechanical and
shrinkage behaviour. The dispersion efficiency and participation
of CNTs in the reinforcement of mortars were analysed by means
of scanning electron microscopy (SEM).
2. Experimental programme
2.1. Materials
Five types of CNTs were supplied by Timesnano Company
(Table 1). CNTSL and CNTSS, with distinct aspect ratios, were sup-
plied in suspension form at CNT concentrations of 5% and 9%,
respectively. They were previously dispersed in water with the
aid of a polyethylene glycol aromatic imidazole surfactant-
TNWDIS. Pristine CNTPL had the same aspect ratio of CNTSL, but
it was supplied in powder form. CNTCOOH and CNTOH, also sup-
plied in powder form, were initially acid treated to render –COOH
and –OH sidewalls functionalization, respectively. CNTOH pre-
sented the highest aspect ratio. The CNTs were selected in order
to cover and study the influence of different factors, namely the
aspect ratio, the surface functionalisation and the deliver condition
(powder form or previously dispersed in aqueous suspension).
Aspect ratios considered in this study are in the range of those usu-
ally considered in literature [17,22].
To disperse the powder form CNTs (CNTPL, CNTCOOH and
CNTOH), the anionic commercial surfactant Dolapix PC67
(Zschimmer-Schwarz) was used. This surfactant was selected
based on a previous study [23]. Portland cement type I 42.5R
according to EN197-1 [24] and a polycarboxylate based superplas-
ticizer, SP (MasterGlenium SKY 548) were used in mortar produc-
tion. Aggregates consisted of 0.25/3.35 coarse sand and 0.125/1.0
fine sand.
2.2. Carbon nanotubes dispersion
The dispersion of powder CNTs in water was performed accord-
ing to Guedes et al. [23]. The required CNT content was mixed with
Table 1
Characteristics of CNTs (provided by supplier).
Notation CNTSS
Commercial name TNIM8
Form as supplied Suspension
Purity (%) >90
Outer diameter (nm) >50
Inner diameter (nm) 5–15
Length (mm) 10–20
Aspect ratio 300
True density (g/cm3) 2.1
COOH (%)
OH (%)
70% of the mixing water and Dolapix PC67 (CNT:Dolapix of 1:1 for
CNTPL, 1:0.5 for CNTCOOH and 1:0.25 for CNTOH), by means of a
mechanical procedure consisting of 4 h of magnetic stirring, fol-
lowed by about 30 min of sonication. The chemical method was
not considered for CNTSS and CNTSL, since these products were
previously supplied in aqueous suspensions, but they were also
subjected to stirring and sonication.
2.3. Mix compositions and sample preparations
Mortars with different w/c (0.35, 0.45, 0.5 and 0.55) were pro-
duced to cover a wide spectrum of cementitious compositions.
From a previous study [6] concerning the mechanical characteriza-
tion of cement pastes with 0.015–1%, by weight of cement, of the
same CNTs used in this study, the optimal amount of CNTs was
achieved for 0.1% of CNTSS of lower aspect ratio and 0.05% of the
other types of CNTs with higher aspect ratios (CNTPL, CNTSL,
CNTCOOH and CNTOH). Based on this previous study, mortars
were produced with the same optimum CNTs content. For shrink-
age tests, the amount of CNTs varied between 0.1 and 1% in order
to analyse the effect of higher CNT contents. Mortars were pro-
duced with 1:3 cement to sand ratio, in which siliceous natural
sand consisted of 70% coarse and 30% fine sand, for better granular
packing. Mortars with w/c of 0.35 and 0.45 were produced with
1.2% and 0.15% of SP (by weight of cement), respectively. In paral-
lel, reference cement mortars (RCM) without CNTs were also
produced.
Mortars were produced in a traditional multi-speed planetary
mortar mixer. The dispersed CNT suspension was gradually added
to the cement and mixed for 3 min. For mortars with w/c of 0.35
and 0.45, SP was added to the cement paste after mixing 1 min.
In another mixer, fine and coarse sands were mixed and saturated
with the remaining 30% of mixing water. Then, the cement paste
was added to the saturated aggregates and mixed for more 2
min. For each composition, test and curing age (see 2.4), three pris-
matic specimens of 40  40  160 mm were produced. For shrink-
age test, steel pins were incorporated at the centre of both ending
surfaces of each prism, to allow the shrinkage measurements.
The specimens were compacted in a common mortar moulding
compaction table and then covered with a plastic film until 24 h.
After demolding, all the specimens, except those for shrinkage test,
were water-cured in a chamber with relative humidity over 95%
and temperature of 20 ± 2 °C until testing. The specimens used
for shrinkage tests were placed in a controlled chamber with a
temperature of 22 ± 2 °C and a relative humidity of 50 ± 5%, based
on LNEC E398 [25].
2.4. Testing procedures
For compressive and flexural tests, samples were tested at 7, 28
and 90 d according to EN 1015-11 [26]. The samples for compres-
sive strength test were obtained from the two remaining parts of
CNTSL CNTPL CNTCOOH CNTOH
TNIM6 TNIM6 TNIMC6 TNIMH4
Suspension Powder Powder Powder
>90 >90 >90 >90
20–40 20–40 20–40 10–30
5–10 5–10 5–10 5–10
10–30 10–30 10–30 10–30
667 667 667 1000
2.1 2.1 2.1 2.1
1.36–1.5
2.48
 A. Hawreen et al. / Construction and Building Materials 168 (2018) 459–470 461

the beams tested for flexural strength. The ultrasonic pulse velocity
(UPV) and dynamic modulus of elasticity were determined at the
same ages, following the recommendations of EN 12504-4 [27]
and ASTM E1876 [28], respectively. The UPV was measured by
direct transmission in specimens, by means of the portable ultra-
sonic non-destructive digital indicating tester (PUNDIT). The time
taken by an ultrasonic pulse to travel through the sample was mea-
sured with accuracy up to 0.1 ms, using 54 kHz transducers posi-
tioned in the middle of each opposing specimen’s face. The
dynamic modulus of elasticity was determined by impulse excita-
tion of vibration, using the GrindoSonic system equipment. An
elastic impulse was transmitted to the test specimen and then
converted by the detector to an electrical signal, which is used to
compute the dynamic modulus. According to LNEC E398 [25],
shrinkage was monitored by a length comparator with a gauge
length of 5 mm and a precision of 0.01 mm. Shrinkage was first
measured at 24 h and then daily up to 7 d, every two days up to
30 d, every 15 d up to 90 d and monthly up to one year.
Toughness test was performed in notched beam specimens
under three-point loading, according to RILEM TCS [29]. Prior to
testing, specimens were notched with about 13 mm crack at half
span. A deflectometer with 1 mm precision and 5 mm course was
installed to measure the mid-span deflection. In addition, a
dynamic clip-on extensometer was also installed to measure the
Crack Mouth Opening Displacement (CMOD) [30]. The samples
were then loaded until failure, and deflection, CMOD and the max-
imum failure load were recorded. The fracture strength, rf, and sta-
tic Young’s modulus, Es, were calculated based on the maximum
load values [22] and the load–CMOD curves’ initial linear elastic
segment [30], respectively. The fracture energy was calculated
from the area under the load–deflection curve based on the RILEM
[29], according to Eq. (1).
(m); d - beam depth (m); b is the beam width (m); a0 - notch depth
(m).
The dispersion and interaction of CNTs with the cement matrix
were analysed by means of SEM, using samples of about 10  10 
10 mm, collected from broken specimens after the mechanical
tests.
3. Results and discussions
The average values of hardened density, q, flexural strength,
fctm, compressive strength, fcm, dynamic modulus of elasticity, Ed,
and ultrasonic pulse velocity, UPV, at 7, 28 and 90 d are presented
in Table 2. The coefficient of variation, CV, for each test is also indi-
cated. In general, CV was similar in reinforced and non-reinforced
mortars. Minor differences were found between the hardened den-
sities of RCM and CNT-reinforced mortars, being lower than 4%.
This suggests that the surfactants used to disperse the low amount
of CNTs adopted in this study didn’t introduce significant extra air
in mortar mixes. Nevertheless, slightly lower densities were found
in mortars where the manufacturer’s surfactant was introduced
(CNTSS, CNTSL).
3.1. Flexural strength
The incorporation of CNTs was effective to increase the flexural
strength of mortars (Table 2) due to the pore filling with CNTs in
the spaces between the hydration products [6,7,31], the generation
of extra nucleation sites on CNTs for the faster and more uniform
growth of C–S–H [6,12] and the retention of microcrack propaga-
tion provided by CNTs bridging them ([6,12,32], see 3.7).

Ao þ mgd0 3.1.1. Influence of type of carbon nanotubes

GF ¼ ð1Þ
ðd  a0 Þb The type of CNTs highly affected the flexural strength of mor-
tars. This is due to the fact that CNT dispersion and reinforcement
where: Ao - total area under the load–deflection curve (N/m); mg - efficiency are mainly controlled by the characteristics of CNTs
beam self-weight (kg); d0 - deflection at final failure of the beam (structure defects and aspect ratio) and dispersion technique

Table 2
Density, q, flexural strength, fctm, compressive strength, fcm, dynamic elasticity modulus, Ed, and ultrasonic pulse velocity, UPV, of mortars.

Mix w/c CNT (%) q fctm fcm Ed UPV

kg/m3 MPa CV MPa CV GPa CV km/s 7/28/90 d CV
28 d 7/28/90 d (%) 7/28/90 d (%) 7/28/90 d (%) (%)
28 d 28 d 28d 28 d
RCM_0.35 0.35 0 2347 9.6/10.3/10.4 2.5 68.1/72.1/73.6 0.7 43.5/45.0/45.0 0.1 4.4/4.4/4.5 0.6
CNTSS_0.35 0.35 0.1 2335 11.8/13.1/13.3 2.5 76.8/84.3/86.8 1.0 45.0/47.5/47.7 3.0 4.5/4.5/4.6 0.2
CNTSL_0.35 0.35 0.05 2319 10.9/12.7/12.0 2.2 72.7/75.8/77.9 1.5 42.1/43.4/44.8 2.1 4.3/4.4/4.4 1.0
CNTCOOH_0.35 0.35 0.05 2349 11.4/12.1/13.0 2.4 72.4/83.1/86.2 1.5 44.7/47.1/47.1 0.7 4.5/4.5/4.6 0.9
CNTPL_0.35 0.35 0.05 2406 12.1/13.3/13.6 1.0 80.6/88.8/90.4 1.3 45.8/49.7/49.7 0.9 4.5/4.7/4.7 0.1
CNTOH_0.35 0.35 0.05 2440 11.1/11.6/12.0 0.6 76.8/85.3/88.1 1.6 46.7/48.7/48.3 0.5 4.5/4.6/4.6 0.5
RCM_0.45 0.45 0 2338 7.1/8.9/9.6 2.6 45.3/56.4/57.9 1.0 36.9/37.4/38.7 1.3 4.2/4.3/4.4 –
CNTSS_0.45 0.45 0.1 2310 9.3/10.6/11.1 0.8 53.2/64.7/65.8 1.3 39.6/40.8/40.8 0.3 4.3/4.4/4.4 0.6
CNTSL_0.45 0.45 0.05 2295 7.9/9.9/10.2 1.6 46.7/56.5/57.5 1.8 36.3/38.7/36.8 3.3 4.1/4.2/4.2 0.3
CNTCOOH_0.45 0.45 0.05 2393 9.3/10.4/10.5 2.0 58.4/63.0/66.7 1.6 39.6/41.5/41.9 0.6 4.3/4.4/4.4 1.1
CNTPL_0.45 0.45 0.05 2368 9.3/10.5/10.6 2.0 55.4/65.2/65.7 1.2 36.9/41.2/43.6 0.2 4.2/4.4/4.3 0.6
CNTOH_0.35 0.45 0.05 2392 8.3/10.0/10.5 0.5 52.6/63.6/67.6 0.5 39.2/41.8/43.5 0.7 4.2/4.3/4.4 0.1
RCM_0.5 0.5 0 2311 6.8/8.3/9.2 3.4 38.1/51.3/52.4 2.6 35.0/36.2/37.6 1.7 4.1/4.2/4.3 –
CNTSS_0.5 0.5 0.1 2316 9.0/9.9/10.2 0.5 46.3/58.8/61.5 1.0 38.0/38.3/39.0 0.3 4.1/4.2/4.3 0.1
CNTSL_0.5 0.5 0.05 2277 7.4/8.8/9.2 1.9 43.0/52.9/56.5 1.2 34.5/35.5/37.5 2.4 4.0/4.1/4.2 0.1
CNTCOOH_0.5 0.5 0.05 2356 8.5/9.5/10.1 3.9 45.1/55.6/57.0 0.9 37.0/37.2/40.4 1.3 4.0/4.3/4.3 0.5
CNTPL_0.5 0.5 0.05 2396 8.2/9.8/10.1 0.8 45.6/58.8/59.5 1.6 36.7/40.2/43.1 1.4 4.1/4.4/4.4 0.3
CNTOH_0.5 0.5 0.05 2369 8.4/9.5/9.7 3.0 46.6/59.6/63.5 2.4 36.8/39.5/40.8 0.8 4.0/4.2/4.3 0.2
RCM_0.55 0.55 0 2300 6.5/7.6/8.4 2.7 36.9/48.8/50.6 2.2 34.5/35.2/35.7 1.0 4.0/4.1/4.2 –
CNTSS_0.55 0.55 0.1 2262 8.3/9.2/9.3 2.2 38.9/51.0/52.0 1.6 37.1/37.8/38.8 1.0 4.0/4.1/4.2 0.2
CNTSL_0.55 0.55 0.05 2282 7.2/7.8/8.5 2.6 34.7/44.3/44.7 1.7 30.0/34.1/35.1 1.6 3.7/3.9/4.0 0.2
CNTCOOH_0.55 0.55 0.05 2386 7.9/8.7/9.0 2.0 43.7/55.4/57.7 2.1 33.6/38.9/37.6 1.8 4.0/4.2/4.2 0.8
CNTPL_0.55 0.55 0.05 2345 7.8/8.8/9.3 1.9 42.1/54.5/58.1 0.8 34.4/36.6/38.8 0.3 3.9/4.3/4.3 –
CNTOH_0.55 0.55 0.05 2351 7.7/8.6/8.9 4.9 41.4/51.8/54.5 1.4 33.4/36.5/37.1 1.2 4.0/4.2/4.2 0.5
462 A. Hawreen et al. / Construction and Building Materials 168 (2018) 459–470

1.4

1.29
fctm, CNT/ fctm, RCM

1.27

1.21
11.19
19
1.3

1.19

1.18
1.18
1.17
1.17
w/c 0.35

1.16
1.14
1.14

1.14
1.14

1.13
1.13
1.12
1.11
1.2 w/c 0.45

1.06
w/c 0.50

1.03
1.1 w/c 0.55
1.0
CNTSS CNTSL CNTCOOH CNTPL CNTOH
Fig. 1. Flexural strength of CNT-reinforced mortars (0.1% of CNTSS or 0.05% of other types of CNT) related to the flexural strength of RCM, as a function of w/c and type of CNTs
at 28 d.

[6,12]. Fig. 1 presents the flexural strength of each CNT-reinforced
mix relative to that of RCM, at 28 d.
Except for mortars with w/c of 0.35, in which CNTPL was
slightly more effective, the remaining mixes with CNTSS were able
to attain the maximum flexural strengths, being about 19–27%
higher than those of RCM at 28 d. Short CNTs are less likely to
agglomerate, being possible to effectively disperse greater
amounts of CNTs. Nevertheless, mortars with only 0.05% of CNTPL
of higher aspect ratio could attain similar mechanical behaviour,
since it led to about the same free spacing between CNTs as found
in mortars with the double amount of short CNTSS [12]. In fact,
more fibers bridging microcracks with greater bonding strength
may be attained with well dispersed CNTs of higher aspect ratio.
The lowest strength improvement was found in mortars with
CNTSL. This confirms the great relevance of CNT dispersion proce-
dure, because mortars with CNTSL and CNTPL showed very differ-
ent results despite involving the same types of CNTs with identical
aspect ratio. For example, in mortars with w/c of 0.35 the strength
increment with 0.05% CNTPL was about twice as that of identical
amount of CNTSL. In this case, the surfactant and sonication proce-
dure used to disperse CNTPL (Dolapix PC67) were more efficient
than those adopted by the manufacturer for the same type of
CNT (CNTSL). Also considering the results obtained for CNTSS, it
is concluded that the adopted commercial surfactant was less
effective for CNTs of higher aspect ratio (CNTSL).
Slightly lower increments were found in functionalised
CNTCOOH-mortars when compared to CNTPL-mixes with CNTs of
identical aspect ratio. For mortars with CNTOH of greater aspect
ratio, the strength increment was even lower. Some reasons can
be attributed to the lower performance of functionalised CNT-
mortars: the CNT functionalisation process can introduce CNT
structural defects, leading to lower reinforcement efficiency; the
simultaneous presence of surfactants may counteract the interac-
tion between treated CNTs and C–S–H; acid treated CNTs can lead
to the degradation of the mechanical properties due to formation
of expansive ettringite (3.7), which weaken the formation of C–
S–H [6,32,33]. The slightly lower efficiency of CNTOH than
CNTCOOH can be related with its higher aspect ratio and the sub-
sequent worse dispersibility.
3.1.2. Influence of w/c
In general, the same type and concentration of CNTs con-
tributed more to the increase of flexural strength in mortars with
w/c of 0.35 than in those with higher w/c (Fig. 1). For example,
comparing to RCM, the strength increments at 28 d in mortars with
CNTSS and w/c of 0.35, 0.45, 0.5 and 0.55 were 27%, 19%, 20% and
21%, respectively. These results may be explained by incorporating
SP in mortars with low w/c, which assisted the cement and CNT
dispersion. In fact, the use of SP allows the more effective break-
down of cement agglomerates, which eventually contributes to
the better dispersion of CNTs. In addition, the greater amount of
CNTs incorporated in mortars with lower w/c and higher cement
content also contributes to their higher performance, since the
amount of CNTs is defined by weight of cement. For example, mor-
tars with w/c of 0.35 had more 15% of CNTs per volume of paste
than mixes with w/c of 0.45.
3.1.3. Influence of curing age
In general, the maximum increase of flexural strength in CNT-
reinforced mortars was attained at 7 d, regardless of the w/c
(Fig. 2). This phenomenon is related with the influence of CNTs
on the morphology of the cement hydration products, being able
to accelerate the hydration process at early ages due to the CNTs
nucleation effect [6,12]. The nucleation effect of the CNTs used in
this study was also shown by thermogravimetric analysis [12].
Only for w/c of 0.35, the relative strength of CNTSS and CNTPL mor-
tars were similar to slightly higher at later ages. In these composi-
tions with SP, the good dispersion of cement particles is also
attained in RCP mixes and the nucleation and acceleration beha-
viour promoted by CNTs is less prominent.
Except for w/c of 0.35, the relative flexural strength tended to
decrease with the testing age, being lower at 90 d. No particular
reason was found for this downward trend, but the progressive
strength recovering of RCM, in which the hydration process was

1.4
w/c 0.35 w/c 0.45 w/c 0.50 w/c 0.55

1.3
fctm,CNT/fctm,RCP

CNTSS

1.2 CNTSL
CNTCOOH
1.1 CNTPL
CNTOH
1.0
7 28 90 7 28 90 7 28 90 7 28 90
Age (day)

Fig. 2. Flexural strength at 7, 28 and 90 d of CNT-reinforced mortars (0.1% of CNTSS or 0.05% of other types of CNT) related to the flexural strength of RCM with different w/c.
 A. Hawreen et al. / Construction and Building Materials 168 (2018) 459–470 463

slower, may explain this over time reduction. Other hypothesis is
that for later ages the influence of the quality of the aggregate-
paste transition zone (ITZ) may become more relevant (3.7).
3.1.4. Influence of aggregate
Contrary to cement pastes, mortars reinforced with CNTs are
also affected by the presence of fine aggregates. For common
mortars produced with normal weight aggregates the stiffness
of aggregates should be higher than that of surrounding matrix.
Therefore, the strength is mainly affected by the cement paste
and the quality of the aggregate-paste ITZ. Comparing to other
results for cement pastes reported in literature, the contribution
of CNTs tends to decrease in reinforced mortars [6,14,19]. Taking
into account pastes with the same w/c and type and amount of
CNTs, Hawreen et al. [6] found >20% higher strength increase
than that obtained in this study for mortars. On the one hand,
the mixing procedure for mortars is more complex and difficult
to be implemented than that used for cement pastes, which
may result in poorer dispersion. On the other hand, the presence
of fine aggregates disturbed the uniform dispersion of CNTs due
to their much greater size that affect the CNT spacing and compli-
cated their efficient reinforcement of the matrix [12]. However,
SEM analysis showed reasonable dispersion of CNTs, with no
occurrence of big agglomerates (3.7). In addition, since the failure
path usually travelled through the paste and contour around
aggregates, the mechanical strength was affected by the
aggregate-paste ITZ. In these terms, if CNTs were not able to also
improve the strength of these regions, namely the aggregate-
paste bonding strength, their reinforcement efficiency would be
lower in mortars.
1.29
3.2. Compressive strength
Basically, for the same reasons pointed out for flexural strength
results, it is shown that CNTs could also contribute for the com-
pressive strength enhancement of cementitious materials (Table 2,
Fig. 3). As previously reported by other authors [6,7,14,34], it was
confirmed that the addition of CNTs led to a lower increase in com-
pressive strength (up to 23%) than in flexural strength (up to 29%),
at 28 d.
Except for CNTSL, the compressive strength improvement
obtained at 7 and 28 d varied between 5 and 29%, depending on
the w/c, age and type of CNTs (Fig. 3a and b). Mortars with
CNTCOOH, CNTPL and CNTOH shown similar to slightly higher
compressive strength improvements than those with CNTSS of
lower aspect ratio (Fig. 3). This is not in line with the flexural
strength results. In this situation, the eventual better dispersion
and more efficient force transfer provided by CNTSS did not prob-
ably compensate the slightly lower density and higher void con-
tent of these mortars. In fact, the compressive strength may be
4–5% reduced for each 1% of extra air voids in the cement matrix
[35]. Since the role provided by CNTs was shown to be less effec-
tive in compressive strength than in flexural strength, namely on
the bridging effect and retention of crack propagation, the influ-
ence of air entrainment becomes more relevant. For example, the
maximum strength increment (29%) at 7 d was attained in the
mortar with w/c of 0.45 and 0.05% CNTCOOH, which had the high-
est density among all the mortars with the same w/c (Table 2).
Also based on flexural results, we may conclude that mortars
with longer CNTs (CNTPL) could attain similar to slightly higher
mechanical performance at lower contents as mortars with higher
concentration of shorter CNTs (CNTSS). Once more, the lowest

(a)
1.22

1.22
1.4
1.22

1.20
1.18
1.18

1.18
1.17

1.16
1.14
1.13
1.13

1.13

1.12
1.07
fcm,CNT/ fcm,RCM

1.06
1.05

1.03

1.2
w/c 0.35
0.94

1.0 w/c 0.45

w/c 0.50
0.8 w/c 0.55
CNTSS CNTSL CNTCOOH CNTPL CNTOH

1.4
1.23

(b)
1.18
1.17

1.16
1.16
1.15
1.15
1.15

1.15
1.14

1.13
fcm,CNT/ fcm,RCM

1.12

1.12
1.08

1.06

1.2
1.05
1.05

w/c 0.35
1.03
1.00

w/c 0.45
0.91

1.0 w/c 0.50

w/c 0.55
0.8
CNTSS CNTSL CNTCOOH CNTPL CNTOH

(c)
1.23

1.4
1.21
1.20
1.18

1.17

1.17

1.17
1.15

1.15
1.14
1.14

1.14
1.13
fcm,CNT/ fcm,RCM

1.09
1.08

1.08
1.06

w/c 0.35
1.03

1.2
0.99

w/c 0.45
0.88

1.0 w/c 0.50

w/c 0.55
0.8
CNTSS CNTSL CNTCOOH CNTPL CNTOH
Fig. 3. Compressive strength of CNT-reinforced mortars (0.1% of CNTSS or 0.05% of other types of CNT) related to the compressive strength of RCM, as a function of w/c and
types of CNTs at (a) 7 and (b) 28 and (c) 90 d.
464 A. Hawreen et al. / Construction and Building Materials 168 (2018) 459–470

strength improvement was found in mortars with CNTSL. As for 60

flexural strength, similar increments were observed in treated 50 R² = 0.92
CNTCOOH-mortars when compared to the most effective CNTPL-

Ed (GPa)
mixes with CNTs of identical aspect ratio. Contrary to what was 40
found in flexural strength, mortars with CNTOH did not show 30 RCM
worse behaviour. Again, in general, it is confirmed that the addition CNT
20
of the same type of CNTs had higher influence on the strength 20 30 40 50 60 70 80 90 100
increase of mortars with low w/c and early age (7 d). fcm (MPaa)

3.3. Dynamic elastic modulus Fig. 4. Compressive strength vs dynamic modulus of elasticity of RCM and CNT-
mortars.
The incorporation of CNTs was effective to improve the dynamic
modulus of elasticity of mortars, although the maximum incre-
4.8
ment was in general lower than 10% (Table 2). This modest
4.6 R² = 0.89
enhancement of elastic modulus, when compared to other

/)
(Km/s)
4.4
mechanical properties, was already expected since the mortar stiff-

ness is also controlled by the type and volume of aggregates, which
do not vary between different mixes of equal w/c. In fact, unlike
compressive and flexural strength, the modulus of elasticity is less
affected by the paste quality. As indicated in 2.3, the volume of
sand is about 50–60%. Therefore, the reinforcement provided by
CNTs can only improve part of the mortar (cement paste), which
leads to lower global variation of the modulus of elasticity. In addi-
tion, these dynamic tests were performed in unloaded and
uncracked specimens, not allowing the full contribution of CNTs,
namely through their bridging effect.
In general, the effect of CNTs in the dynamic modulus of elastic-
ity improvement followed the same trends found in flexural and
compressive strength. Mortars with CNTSL had the poorest perfor-
mance, showing only up to 4% increase of modulus when com-
pared to the 11% improvement attained with the same aspect
ratio CNTPL, due to the difference in their dispersion process.
Again, greater amount of short CNTs (CNTSS) could be better dis-
persed than lower amounts of CNTSL with higher aspect ratio, pro-
viding reduced CNT free volume of cement paste and better filling
of nanosized voids. As for the use of functionalised CNTs
(CNTCOOH and CNTOH), increments of about 12% were found as
compared to that of RCM.
Unlike for flexural and compressive strength, the maximum
increment of dynamic modulus of elasticity was obtained for mor-
tars with w/c of 0.45, regardless of the type of CNTs. This can be
partly attributed to the greater amount of sand in mortars with
w/c of 0.35 than in mixes with w/c of 0.45, being 59% and 56% in
volume, respectively. Therefore, any paste improvement has a
slightly greater influence in mortars with w/c of 0.45.
It is recognized that the quality of the aggregate-paste ITZ
affects the modulus of elasticity, namely by influencing the micro-
crack propagation in this region [36–38]. According to Neville [39],
rigid aggregates can better participate in modulus of elasticity
when improved aggregate-paste bond is achieved. From the
obtained results, it is only possible to conclude that the incorpora-
tion of CNTs helped to improve the global modulus of elasticity.
The effect of CNTs on the ITZ is better analysed in 3.7. A high linear
correlation was found between the modulus of elasticity and com-
pressive strength (Fig. 4), even taking into account different types
of CNTs, w/c and testing ages. This high correlation shows that
compressive strength was essentially affected by the characteris-
tics of the cement paste and ITZ. Otherwise, the aggregates partic-
ipation would lead to different trends and lower R2.
3.4. Ultrasonic pulse velocity
The UPV mainly depends on the square root of the modulus of
elasticity and density of materials [40]. Therefore, since these
two properties were slightly influenced by the incorporation of
CNTs (Table 2), the same is expected for UPV. On the other hand,
UPV (K
4.2
4.0
RCM
3.8 CNT
3.6
20 30 40 50 60 70 80 90 100
fcm (MPa)
Fig. 5. Ultrasonic pulse velocity vs compressive strength of RCM and CNT-mortars.
the UPV was measured in uncracked specimens with minor contri-
bution of the bridging effect of CNTs. For these reasons, the incre-
ment of UPV in reinforced mortars was only up to 6% (Table 2). In
many cases the difference of UPV between mortars was even lower
than 3%, which is within the test variability.
As the compressive strength and UPV are both affected by the
paste microstructure, a good correlation (R2 = 0.89) was obtained
between these properties (Fig. 5). This is valid, because no signifi-
cant differences were introduced on the type and volume of aggre-
gates between mixes. The major variations occurred at the w/c and
type of CNTs, which only affected the microstructure of the cement
paste.
3.5. Fracture toughness
Relevant increments up to 54%, 65% and 31% in rf, Gf and Es
were obtained in CNT-reinforced mortars, respectively, when com-
pared to RCM (Table 3). The strength increase can be attributed to
the various factors previously mentioned, namely the effects of fil-
ler, nucleation and bridging. However, the bridging effect should
be the mechanism that contributes most to the fracture energy.
The highest Gf of CNT-reinforced mortars was essentially justi-
fied by the increment obtained in the flexural peak strength. Con-
sequently, higher residual strength for a given net deflection and
CMOD were obtained in CNT-mortars. The higher strain capacity
of CNT-reinforced mortars is also shown in Fig. 6. In addition,
greater increase of the fracture strength (rf) than of fctm (Table 2)
was attained in CNT-reinforced mortars when compared to RCM.
This means that CNTs were able to improve the post-cracking
behaviour of mortars, showing their positive action on the reten-
tion of crack-propagation (Fig. 6).
The post-peak behaviour was similar between different mor-
tars, showing low deformations and hence low ductility, regardless
of the type of CNTs. Nevertheless, as mentioned, reinforced mortars
evidenced greater post-peak residual strengths (Fig. 6). In sum, the
greatest reinforcement contribution of CNTs occurred in the pre-
peak behaviour of mortars. In fact, CNTs are only able to provide
their bridging effect at nanoscale, involving very small microcracks
with up to about 1 lm wide ([12], see 3.7). In these terms, the first
microcracks can be retained and their propagation and develop-
ment of macrocracks are delayed. However, as soon as the maxi-
mum strength is attained and macrocracks are developed, CNTs
 A. Hawreen et al. / Construction and Building Materials 168 (2018) 459–470 465

Table 3
Fracture strength (rf), fracture toughness (Gf) and static elastic modulus (Es) of mortars at 28 d.

Mix w/c CNT CNT rf Gf Es

Type % MPa CV (%) D (%) N/m CV (%) D (%) GPa CV (%) D (%)
RCM_0.45 0.45 – 0 3.2 5.7 0.0 70.7 1.7 0.0 13.8 4.0 0.0
CNTSS_0.45 0.45 CNTSS 0.1 4.7 2.7 45.0 103.3 2.1 46.0 16.2 5.2 17.0
CNTSL_0.45 0.45 CNTSL 0.05 4.0 3.9 22.0 87.7 8.2 24.1 12.7 3.5 8.0
CNTPL_0.45 0.45 CNTPL 0.05 4.9 4.5 49.8 106.2 1.9 50.2 17.9 3.9 30.0
CNTCOOH_0.45 0.45 CNTCOOH 0.05 5.0 3.1 54.4 116.9 4.6 65.2 18.1 2.9 31.0
CNTOH_0.45 0.45 CNTOH 0.05 4.5 5.9 40.0 94.6 1.7 33.8 16.7 3.8 21.0

(a) 1.4 (b) 1.4

1.2 1.2
1.0 RCM 1.0

Load (kN)
CNTSS
Load (kN)

0.8 0.8
RCM
CNTSL
0.6 0.6 CNTSS
CNTPL CNTSL
0.4 0.4 CNTPL
CNTCOOH
0.2 CNTOH 0.2 CNTCOOH
CNTOH
0.0 0.0
0 1 2 3 4 5 0 0.01 0.02 0.03 0.04 0.05
Deflection (mm) COMD (µm)

Fig. 6. (a) Load-deflection and (b) load-CMOD curves of mortars with w/c of 0.45 and 0.1% of CNTSS or 0.05% of other types of CNT.

are no longer able to retain them and a suddenly abrupt failure
occurs. This suggests that the global fracture behaviour could be
rather improved with the additional incorporation of fibers, more
adequate to retain the subsequent development of bigger cracks.
In addition, comparing to RCM, a relevant increase of the Es was
found in CNT-reinforcement mortars, especially in those produced
with CNTCOOH, CNTOH and CNTPL. This increase was up to about
3 times higher than that of dynamic modulus of elasticity, Ed. Note
that Es was determined in loaded and cracked specimens, which
contributed to a better effectiveness of the bridging mechanism
provided by CNTs. Moreover, for cracked specimens under tensile
stresses, the aggregates have less influence on the beam stiffness
and the effect of CNTs is more noticed.
Unlike flexural and compressive strength tests, maximum rf, Gf
and Es of CNT-mortars were obtained with the addition of
CNTCOOH, being 54%, 65% and 31% higher than those of RCM,
respectively. These increments were slightly higher than those
found in mortars with pristine CNTs (CNTPL) of the same aspect
ratio, being 50%, 50% and 30% higher than RCM, respectively. This
suggests that CNTs with higher aspect ratio and greater surface
area to volume ratio allowed higher pull-out tensile strengths
and greater strain capacity, being more effective in crack retention.
In addition, functionalised CNTs may develop higher bond
strengths with the surrounding hydration products of cement
pastes [7], leading to higher pull-out strengths. However, this
was not confirmed in the compressive and flexural strength tests,
in which other factors, such as the nucleation and filler effect were
also relevant.
Despite its slightly worse toughness behaviour, high fracture
energy increments were also achieved in mortars with CNTSS of
low aspect ratio. As mentioned, higher amounts of this type of
CNTs are easier dispersed in the cement matrix. The highest flexu-
ral strengths (Table 2) found in mortars produced with this type of
CNTs confirms that the dispersion of greater amounts of nanotubes
was effective and hence, the filling and nucleation effects could be
more relevant in these mixes. However, in toughness tests the
bridging effect is more determinant than the other effects of filler
and nucleation.
Again, the lowest improvement was found in CNTSL-mortars,
which confirms the inadequate dispersion of these CNTs. Mortars
with CNTSS and CNTOH showed similar increments of rf, Gf and
Es, being about 40–45%, 34–46%, 17–21%, respectively, when com-
pared to RCM.
3.6. Shrinkage
The shrinkage of CNT-reinforced mortars produced with differ-
ent amounts (0.1–1%) and types of CNTs and w/c (0.35, 0.45, 0.5)
are presented in Fig. 7. Shrinkage values at 2, 7, 90, 180 and 365
d are summarized in Table 4. As expected, shrinkage was lower
in mortars with lower w/c, which had lower evaporable water,
denser microstructure and greater paste content.
Shrinkage of cementitious materials can be controlled by reduc-
ing the source of shrinkage (cement hydration and evaporation of
capillary pore water) or increasing the restriction effect to free
shrinkage. This restriction effect is provided by the stiffness of
the cement matrix and, most of all, by the aggregate skeleton. In
general, the incorporation of CNTs was effective to reduce the
shrinkage of mortars. On the one hand, the small diameters of CNTs
(10–50 nm), due to their filler and nucleation effect, promote the
decrease of the amount of fine pores between the hydration prod-
ucts. As a consequence, the capillary stresses are reduced, resulting
in lower shrinkage [10]. In this case, CNTs can act as reducing
source of shrinkage in mortar. On the other hand, the microcrack
bridging effect provided by CNTs increases the matrix stiffness,
restricting the shrinkage deformation.
The incorporation of CNTs was more effective at early ages,
which is in accordance with the mechanical results (Table 2). In
general, the greater shrinkage reduction occurred before 7 d,
being 10–90% at 2 d and up to 62% at 7 d, depending on the type
and amount of CNTs and w/c. After 7 d, the reduction of shrinkage
in CNT-mortars was less relevant and suffered little variations
until 1 year. The long term shrinkage reduction was 2–21%
(Table 4). Again, this behaviour can be related with the accelera-
tion of the hydration process at early ages in CNT-reinforced
mixes [6,41].
466 A. Hawreen et al. / Construction and Building Materials 168 (2018) 459–470

Age (day) Age (day) Age (day)

(a) (b) (c)

0 100 200 300 400 0 100 200 300 400 0 100 200 300 400
0 0 0
RCM_0.35 RCM_0.5
-100 -100 RCM_0.45 -100
CNTSS(0.1) CNTSS(0.1) CNTSS(0.1)
-200 -200 -200
CNTCOOH(0.1) CNTSS(0.5) CNTSS(0.5)

ε s (µm/m)
ε s (µm/m)
-300 -300
εs (µm/m)

-300
CNTPL(0.1) CNTSS(1) CNTSS(1)
-400 -400 CNTPL(0.1) -400
-500 -500 CNTCOOH(0.1) -500
-600 -600 -600
-700 -700 -700
-800 -800 -800
-900 -900 -900
-1000 -1000 -1000

Fig. 7. Shrinkage of mortars with different types and amounts of CNTs and w/c of (a) 0.35, (b) 0.45 and (c) 0.5.

Table 4
Shrinkage, Ɛs, of mortars after 2, 7, 90, 180 and 365 d.

Mix w/c CNT (%) Ɛs,2d Ɛs,7d Ɛs,90d Ɛs,180d Ɛs,365d

mm/m D mm/m D mm/m D mm/m D mm/m D CV
(%) (%) (%) (%) (%) %
RCM_0.35 0.35 0 55 – 207 – 430 – 531 – 594 – 4.3
CNTSS(0.1) 0.35 0.1 47 14.3 211 1.9 375 12.7 438 17.6 484 18.4 2.7
CNTCOOH(0.1) 0.35 0.1 31 43.5 156 24.7 409 4.8 445 16.2 495 16.6 4.5
CNTPL(0.1) 0.35 0.1 47 14.9 150 27.5 422 1.9 453 14.8 500 15.8 3.6
RCM_0.45 0.45 0 109 – 390 – 773 – 812 – 922 – 2.0
CNTSS(0.1) 0.45 0.1 43 60.6 150 61.5 566 26.7 641 21.1 727 21.2 2.2
CNTSS(0.5) 0.45 0.5 52 52.2 219 44.0 688 11.1 722 11.1 750 18.6 1.0
CNTSS(1) 0.45 1.0 10 90.4 250 35.9 742 4.0 805 0.9 844 8.4 0.7
CNTCOOH(0.1) 0.45 0.1 78 28.4 217 44.5 672 13.1 742 8.6 789 14.4 1.8
CNTPL(0.1) 0.45 0.1 47 57.0 291 25.4 656 15.1 734 9.6 750 18.6 2.2
RCM_0. 5 0.5 0 78 – 359 – 773 – 875 – 984 – 1.7
CNTSS (0.1) 0.5 0.1 47 40.0 250 30.4 680 12.1 750 14.3 797 19.0 1.9
CNTSS (0.5) 0.5 0.5 70 10.0 328 8.7 828 7.1 875 0.0 891 9.5 2.2
CNTSS(1) 0.5 1.0 68 12.9 328 8.6 868 12.2 922 5.4 969 1.6 1.4

For mortars with w/c of 0.45, the best balance between a rea-
sonable dispersion and an effective reinforcement was found for
0.1% of CNTSS, having 1 year shrinkage reduction of 21%. The
shrinkage reduction did not significantly vary for CNT concentra-
tions up to 0.5%, but decreased to only 8% when 1% CNTSS was
incorporated. A similar trend was found for w/c of 0.5. This is
attributed to the greater tendency to agglomerate of high CNTs
concentrations.
The shrinkage of mortars was slightly affected by the type of
CNTs. For mortars with w/c of 0.45, the incorporation of 0.1% of
CNTSS, CNTPL and CNTCOOH resulted in 21%, 19% and 14% lower
shrinkage as compared to that of RCM, respectively, at 1 year.
Smaller differences between these mixes were found for w/c of
0.35, in which the reduction varied between 16% and 18%. In gen-
eral, it can be concluded that all tested types of CNTs were equally
efficient and shrinkage curves showed the same configuration over
time. Only mortars with w/c of 0.45 and CNTCOOH showed slightly
higher shrinkage than those produced with other types of CNTs,
which is in line with the mechanical results.
3.7. Microscopic analysis
In general, CNTs were reasonably dispersed (Figs. 8 and 9), with
only some few cases of punctual agglomeration. Nevertheless, it
was not possible to achieve highly uniform dispersions, being
found some regions with almost no CNTs and others with closely
spaced CNTs (Fig. 8d). The mechanism of crack bridging was clearly
identified for all types of CNTs, namely CNTSS (Fig. 8a–c), CNTPL
(Fig. 8d), CNTCOOH (Fig. 8e) and CNTOH (Fig. 8f). In general, the
predominant fracture mode involved the CNT pull-out (Fig. 8b
and d) or the CNT breakage after some pull-out (Fig. 8a and e).
At low crack widths, CNTs could bridge the cracks without failure,
being able to effectively transfer the force across them (Fig. 8c and
f). In fact, it was found that this active bridging could only occur for
crack openings up to around 1 mm width.
Although all mortars showed ettringite, slightly higher amounts
were found in those with functionalised CNTOH and CNTCOOH
(Fig. 8e and f). The acid treatment of functionalised CNTs leaves
sulfate ions behind that may persist on the surface or inside the
nanotubes, even after the neutralization treatments [32]. The for-
mation of ettringite crystals may cause localized debonding, fol-
lowed by strength reduction [42], which may justify the slightly
lower mechanical strength found in mortars with functionalised
CNTs. However, mortars with CNTOH and, in especial, CNTCOOH
showed high performance in fracture toughness. As shown in
Fig. 8f, this indicates that microcrack bridging was relevant, prob-
ably due to the high aspect ratio and good nanotube/matrix inter-
action of functionalised CNTs.
 A. Hawreen et al. / Construction and Building Materials 168 (2018) 459–470 467

Fig. 8. SEM images of mortars with w/c of 0.5 and (a),(b),(c) 0.1% of CNTSS and 0.05% of (d) CNTPL, (e) CNTCOOH and (f) CNTOH [arrows indicate CNTs (") and ettringite ( )].

Fig. 9 shows that, at least in some areas, nanotubes could be
well distributed in the cement paste with CNTs spacing
ranging 1–4 lm in mortars with w/c of 0.55 and 0.1% CNTSS.
According to a previous work [12], for the same type and amount
of CNTs, the theoretical spacing estimated in pastes with also
w/c of 0.55 was 3.8 mm, which is within the range obtained in this
study.
CNTs were reported to reduce the proportion of crystalline C–H
[12] and increase the amount of high stiffness C–S–H [10], which
may benefit the reduction of porosity at the aggregate-paste inter-
face. However, due to their nanosize, it may be difficult for CNTs to
significantly improve the aggregate-paste bonding strength or to
efficiently retain the large microcracks developed around aggre-
gates during loading. In fact, on average, within the first 5 mm of
the surrounding matrix, no CNT could be found in microcracks
(Fig. 10). Especially for dense aggregates, the cement paste could
not penetrate the few pores in the siliceous aggregate and CNTs
were not able to bridge the cracks developed very near the aggre-
gate surface. However, this does not mean that the cement matrix
in the ITZ was not globally improved in CNT-mortars. In fact, the
ITZ average size in common cementitious materials may vary
between 40 and 50 mm, although the weakest zone is limited to
5–10 mm [43,44]. On the one hand, as found in thermogravimetric
analysis [12], less large crystalline C–H seems to be formed in the
presence of CNTs, which would favor the development of preco-
cious cracks and preferential failure planes. The filler and nucle-
468 A. Hawreen et al. / Construction and Building Materials 168 (2018) 459–470

Fig. 9. (a) and (b) SEM images of CNT spacing in mortars with w/c of 0.55 and 0.1% of CNTSS [arrows indicate CNTs (")].

Fig. 10. (a–c) Aggregate-paste ITZ in mortars with w/c of 0.5 and 0.1% CNTSS [arrows (l): distance between the aggregate and the nearest identified CNTs]. No CNTs were
found near the aggregate surface. (d) Scheme of CNT microcrack bridging in ITZ at orthogonal cracks from aggregates [CNTs are pointed ("), which are not in the same scale of
the background].

ation effect of CNTs may also contribute to the greater quality of
the ITZ around the aggregate. On the other hand, the effective
bridging provided by CNTs was confirmed in the orthogonal cracks
that propagated from the aggregate surface (Fig. 10).
As shown in the scheme of Fig. 10d, the CNTs can assist the
retention of crack propagation in the ITZ, which improves the
mechanical and durability performance of mortars. In sum, the
incorporation of CNTs can delay the aggregate-paste microcrack
development, which is usually the ignition for the mortar failure.
However, as mentioned, the CNTs should not be able to signifi-
cantly improve the bond between the surface of the aggregate
and the cement paste.
 A. Hawreen et al. / Construction and Building Materials 168 (2018) 459–470
4. Conclusions
The reinforcement efficiency of mortars with different types of
pristine and functionalised CNTs was investigated. The main fol-
lowing conclusions were drawn:
- The incorporation of small amounts of CNTs (0.05–0.1%) was
effective to increase the flexural and compressive strengths at
28 d, up to 29% and 23%, respectively. Comparing to short CNTs
(CNTSS), only half amount of CNTs of higher aspect ratio
(CNTPL) was needed to obtain mortars with similar mechanical
performance.
- The incorporation of CNTs was more effective in low w/c mor-
tars, possibly due to the higher concentration of CNTs and the
addition of superplasticizer that assisted the cement and CNT
dispersion.
- CNTs essentially contributed to improve the mechanical
strength and shrinkage at early ages, increasing the hydration
rate due to their nucleation effect.
- The contribution of CNTs to the mechanical strength showed to
be lower in mortars than in cement pastes of identical compo-
sition. This was attributed to the poorer dispersion of CNTs and
the influence of the aggregate-paste ITZ in the mechanical
strength.
- The dynamic modulus of elasticity and ultrasonic pulse velocity
were little affected by the incorporation of CNTs. Up to about 3
times higher increase was obtained in the static modulus of
elasticity determined in cracked conditions.
- Relevant improvements, up to 54% in the fracture strength
and up to 65% in the fracture energy showed the high potential
of CNTs to improve the cracking resistance of cementitious
composites. The greatest contribution of CNTs occurred in the
pre-peak behaviour of mortars, when bridging could be
effective at small microcracks. The most effective reinforce-
ment was found in mortars with CNTs of high aspect ratio, con-
firming their greater crack retention capacity. However, the
post-peak mortar deformation was low, regardless of the type
of CNT.
- CNTs were effective to reduce the shrinkage of mortars, espe-
cially at early ages. The shrinkage reduction after 7 d was up
to 62%, decreasing to up 21% after 1 year. Shrinkage was little
affected by the type of CNT. >0.1% of CNTs was not effective
to further reduce the long-term shrinkage.
- Mortars with functionalised CNTs showed similar to slightly
lower mechanical strength improvement than those with pris-
tine CNTs of equal aspect ratio. However, these mortars showed
the best performance in fracture toughness, where the bridging
effect and the CNT/matrix bond strength was determinant. The
incorporation of functionalised CNTs led to slightly larger
amount of ettringite.
- SEM analysis showed reasonable dispersion of CNTs, being able
to provide efficient bridging in cracks with up to around 1 lm
width. SEM analysis also suggests that the aggregate-paste
bond should not be significantly improved by the incorporation
of CNTs. However, it was shown that CNTs can improve the
quality of the aggregate-paste ITZ.
Acknowledgements
The authors wish to thank CEris–ICIST/IST for funding the
research, as well as the companies SECIL and BASF for supplying
the materials used in the experiments. The first author would like
to thank Fundação Calouste Gulbenkian (Portugal) for the financial
support through scholarship (no. 125745).
469
References
[1] P.M. Ajayan, O. Stephan, C. Colliex, D. Trauth, Aligned carbon nanotube arrays
formed by cutting a polymer resin–nanotube composite, Science 265 (1994)
1212–1214.
[2] C. Gau, H.S. Ko, H.T. Chen, Piezoresistive characteristics of MWNT
nanocomposites and fabrication as a polymer pressure sensor,
Nanotechnology 20 (2009) 185503.
[3] F. Gojny, M. Wichmann, B. Fiedler, K. Schulte, Influence of different carbon
nanotubes on the mechanical properties of epoxy matrix composites – A
comparative study, Compos. Sci. Technol. 65 (2005) 2300–2313.
[4] S.R. Bakshi, D. Lahiri, A. Agarwal, Carbon nanotube reinforced metal matrix
composites - a review, Int. Mater. Rev. 55 (2013) 41–64.
[5] S.C. Tjong, Carbon Nanotube Reinforced Composites: Metal and Ceramic
Matrices, Wiley-VCH, 2009, p. 242.
[6] A. Hawreen, J. Bogas, M. Guedes, M. Pereira. Mechanical characterization of
cement pastes reinforced with pristine and functionalized MWCNTs, in
Materiais, XVIII Congresso da Sociedade Portuguesa de Materiais. 9-12/4/
2017: University of Aveiro, Aveiro, Portugal.
[7] G.Y. Li, P.M. Wang, X. Zhao, Mechanical behavior and microstructure of cement
composites incorporating surface-treated multi-walled carbon nanotubes,
Carbon 43 (2005) 1239–1245.
[8] A. Peigney, C. Laurent, E. Flahaut, A. Rousset, Carbon nanotubes in novel
ceramic matrix nanocomposites, Ceram. Int. 26 (2000) 677–683.
[9] M.M.J. Treacy, T.W. Ebbesen, J.M. Gibson, Exceptionally high Young’s modulus
observed for individual carbon nanotubes, Nature 381 (1996) 678–680.
[10] M.S. Konsta-Gdoutos, Z.S. Metaxa, S.P. Shah, Multi-scale mechanical and
fracture characteristics and early-age strain capacity of high performance
carbon nanotube/cement nanocomposites, Cem. Concr. Compos. 32 (2010)
110–115.
[11] A. Cwirzen, K. Habermehl-Cwirzen, V. Penttala, Surface decoration of carbon
nanotubes and mechanical properties of cement/carbon nanotube composites,
Adv. Cem. Res. 20 (2008) 65–73.
[12] A. Carriço, J.A. Bogas, A. Hawreen, M. Guedes, Durability of multi-walled
carbon nanotube reinforced concrete, Constr. Build. Mater. 164 (2018) 121–
133.
[13] A. Chaipanich, T. Nochaiya, W. Wongkeo, P. Torkittikul, Compressive strength
and microstructure of carbon nanotubes–fly ash cement composites, Mater.
Sci. Eng A 527 (2010) 1063–1067.
[14] S. Musso, J.-M. Tulliani, G. Ferro, A. Tagliaferro, Influence of carbon nanotubes
structure on the mechanical behavior of cement composites, Compos. Sci.
Technol. 69 (2009) 1985–1990.
[15] A. Tamimi, N.M. Hassan, K. Fattah, A. Talachi, Performance of cementitious
materials produced by incorporating surface treated multiwall carbon
nanotubes and silica fume, Constr. Build. Mater. 114 (2016) 934–945.
[16] M.S. Morsy, S.H. Alsayed, M. Aqel, Hybrid effect of carbon nanotube and nano-
clay on physico-mechanical properties of cement mortar, Constr. Build. Mater.
25 (2011) 145–149.
[17] F. Collins, J. Lambert, W.H. Duan, The influences of admixtures on the
dispersion, workability, and strength of carbon nanotube–OPC paste mixtures,
Cem. Concr. Compos. 34 (2012) 201–207.
[18] S. Kumar, P. Kolay, S. Malla, S. Mishra, Effect of multiwalled carbon nanotubes
on mechanical strength of cement paste, Mater. Civ. Eng. 24 (2012) 84–91.
[19] B. Zou, S.J. Chen, A.H. Korayem, F. Collins, C.M. Wang, W.H. Duan, Effect of
ultrasonication energy on engineering properties of carbon nanotube
reinforced cement pastes, Carbon 85 (2015) 212–220.
[20] W. Li, W. Ji, G. Fang, Y. Liu, F. Xing, Y. Liu, B. Dong, Electrochemical impedance
interpretation for the fracture toughness of carbon nanotube/cement
composites, Constr. Build. Mater. 114 (2016) 499–505.
[21] P. Stynoski, P. Mondal, C. Marsh, Effects of silica additives on fracture
properties of carbon nanotube and carbon fiber reinforced Portland cement
mortar, Cem. Concr. Compos. 55 (2015) 232–240.
[22] M.S. Konsta-Gdoutos, Z.S. Metaxa, S.P. Shah, Highly dispersed carbon
nanotube reinforced cement based materials, Cem. Concr. Res. 40 (2010)
1052–1059.
[23] M. Guedes, A. Hawreen, J. Bogas, S. Olhero, Experimental procedure for
evaluation of CNT dispersion in high pH media characteristic of cementitious
matrixes, 1 Congress of Tests and Experimentation in Civil Engineering,
Instituto Superior Técnico, Lisbon, Portugal, 2016.
[24] BSEN 197-1, Cement. Composition, specifications and conformity criteria for
common cements, BSEN, London, 2011.
[25] LNEC E398, Concretes: determination of shrinkage and expansion, LNEC,
Lisbon, 1993.
[26] BSEN 1015-11, Methods of Test for Mortar for Masonry. Determination
of Flexural and Compressive Strength of Hardened Mortar, BSEN, London,
1999.
[27] BSEN 12504-4, Testing Concrete. Determination of Ultrasonic Pulse Velocity,
BSEN, London, 2004.
[28] ASTM E1876-01, Standard Test Method for Dynamic Young’s Modulus, Shear
Modulus, and Poisson’s Ratio by Impulse Excitation of Vibration, ASTM, 2015.
[29] RILEMTCS, Determination of the fracture energy of mortar and concrete by
means of three-point bend tests on notched beams, Mater. Struct. (1985) 285–
290.
[30] S. Shah, Determination of fracture parameters (K Ic s and CTODc) of plain
concrete using three-point bend tests, Mater. Struct. 23 (1990) 457–460.
470 A. Hawreen et al. / Construction and Building Materials 168 (2018) 459–470
[31] T. Nochaiya, A. Chaipanich, Behavior of multi-walled carbon nanotubes on the
porosity and microstructure of cement-based materials, Appl. Surf. Sci. 257
(2011) 1941–1945.
[32] R.K. Abu Al-Rub, A.I. Ashour, B.M. Tyson, On the aspect ratio effect of multi-
walled carbon nanotube reinforcements on the mechanical properties of
cementitious nanocomposites, Constr. Build. Mater. 35 (2012) 647–655.
[33] L.I. Nasibulina, I.V. Anoshkin, A.G. Nasibulin, A. Cwirzen, V. Penttala, E.I.
Kauppinen, Effect of carbon nanotube aqueous dispersion quality on
mechanical properties of cement composite, Nanomaterials (Basel). 2012
(2012) 1–6.
[34] R. Hamzaoui, S. Guessasma, B. Mecheri, A.M. Eshtiaghi, A. Bennabi,
Microstructure and mechanical performance of modified mortar using hemp
fibres and carbon nanotubes, Mater. Design 56 (2014) 60–68.
[35] S. Mindess, J. Young, D. Darwin, Concrete, 2nd edition., Prentice Hall, Pearson
Education, Inc, 2003.
[36] Z. Sun, E.J. Garboczi, S.P. Shah, Modeling the elastic properties of concrete
composites: experiment, differential effective medium theory, and numerical
simulation, Cem. Concr. Compos. 29 (2007) 22–38.
[37] C. Li, J. Zheng, X. Zhou, M. McCarthy, A numerical method for the prediction of
elastic modulus of concrete, Mag. Concrete Res. 55 (2003) 497–505.
[38] K.M. Lee, J.H. Park, A numerical model for elastic modulus of concrete
considering interfacial transition zone, Cem. Concr. Res. 38 (2008) 396–402.
[39] A. Neville, Aggregate bond and modulus of elasticity of concrete: ACI
materials, 1997.
[40] K.W. Nasser, A.A. Al-Manaseer, Comparison of Nondestructive testers of
hardened concrete, ACI Mater. 84 (1987) 374–380.
[41] S.J. Chen, F.G. Collins, A.J.N. Macleod, Z. Pan, W.H. Duan, C.M. Wang, Carbon
nanotube–cement composites: a retrospect, IES, J. Part A Civ. Struct. Eng. 4
(2011) 254–265.
[42] F. Sanchez, A. Borwankar, Multi-scale performance of carbon microfiber
reinforced cement-based composites exposed to a decalcifying environment,
Mater. Sci. Eng. A 527 (2010) 3151–3158.
[43] P. Mehta, P. Monteiro, Concrete: Microstructure, properties and materials,
2006.
[44] S. Chandra, L. Berntsson, Lightweight Aggregate Concrete, Noyes publications-
Wiliam Andrew Publishing Science, Tevhnology and applications, USA, 2003.