Construction and Building Materials 146 (2017) 350–359

Contents lists available at ScienceDirect

Construction and Building Materials

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

Thermal performance, freeze-and-thaw resistance, and bond strength of

cement mortar using rice husk-derived graphene
Inkyu Rhee a,⇑, Jun-Seok Lee b, Jin Hee Kim b, Yoong Ahm Kim c,⇑
a
Dept. of Civil Eng., Chonnam National University, 77 Yongbong-ro, Buk-gu, Gwangju 61186, Republic of Korea
b
Faculty of Eng., Chonnam National University, 77 Yongbong-ro, Buk-gu, Gwangju 61186, Republic of Korea
c
School of Polymer Science and Eng. & Alan G. MacDiarmid Energy Research Institute, Chonnam National University, 77 Yongbong-ro, Buk-gu, Gwangju 61186, Republic of Korea

h i g h l i g h t s

 Thermal performance of GRH was fair among other carbon-based materials.

 GRH mortar showed a better performance against freezing-and-thawing attack.
 Bond strength of thin GRH overlay exhibited better on 0.75 wt% specimen.

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

Article history: We explored the thermal performance, bond strength, and freeze-and-thaw resistance of mortar compos-
Received 29 March 2017 ite incorporating rice husk-derived graphene-like material (GRH). These three features are key for road
Received in revised form 12 April 2017 pavement application, especially against ice and snow on its surface. Groups of GRH mortar prisms
Accepted 13 April 2017
(20  20  50 mm, 50  50  100 mm) were tested to examine heating and cooling performance, in com-
Available online 21 April 2017
parison with other carbon-based materials: multi-walled CNT, graphene platelets, and carbon fibers.
Additionally, thermal performance was assessed in terms of GRH wt% insertion. GRH cubes of 50 mm
Keywords:
with different wt% were tested to capture freeze-and-thaw resistance for 300 cycles. XRD, Rietveld,
Rice husk-derived graphene
Thin overlay
and SEM-EDS analyses of calcium silicate hydrate (CSH) gels of undamaged/damaged cubes with/without
Thermal conductivity GRH were compared at the age of 120 days after casting. A group of U 100  200 mm slanted concrete
Bond strength cylinders was tested with a 10-mm-thick overlay with different wt% GRH. The compressive failure load
Freeze-and-thaw resistance and fractured area were used to evaluate the bond strength of the GRH thin overlay.
Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction of concrete pavements. Because the target temperature of the thin

Traditional snow and ice control techniques focus mainly on
chemical reactions between deicer (chloride) and water in many
countries. Alternatively, a direct heat transfer method for rapid
deicing using electricity and a conductive medium, such as copper
wire, has been applied in some countries. Automatic spraying of
chloride deicer has been used widely to prevent road freezing in
an efficient and economical way. However, the dosage amount of
deicer varies from year to year because, recently, various effects
due to climate change have emerged. Additionally, deicer may
induce a decrease in bond strength and cause surface scaling, as
well as environmental concerns about the washed-out chloride.
For these reasons, direct heat transfer methods have been consid-
ered, by introducing copper wire with electrical charges on the top
⇑ Corresponding authors.
E-mail addresses: rheei@jnu.ac.kr (I. Rhee), yak@jnu.ac.kr (Y.A. Kim).
overlay in the melting process is 0 °C, thermally conductive mate-
rials can be mixed in a thin cementitious overlay onto the surface
of a concrete pavement to enhance the thermal efficiency. How-
ever, this method has some drawbacks: (a) energy loss when elec-
trical potential energy turns into thermal energy, (b) re-installation
is needed when the thin overlay cracks and is damaged under
mechanical loads and the harsh environmental load, and (c) nor-
mally, 30 min of pre-heating is required for a 5 °C temperature ele-
vation [8]. To overcome these disadvantages, some researchers
have mixed graphite powder [22] into concrete mix to enhance
heat conduction. However, a large amount of this is required and
it leads to a lower compressive strength than the normal mix.
The reason of such low strength may be attributed to the graphite
crystal structure itself: hexagonal and planar structure. The atoms
are weakly-bonded to this graphite sheet above and below and all
lie in a plane. Naturally, graphite is very lubricious because of this
[26]. Some research has evaluated adding microsilica and meta-

http://dx.doi.org/10.1016/j.conbuildmat.2017.04.109
0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.
 I. Rhee et al. / Construction and Building Materials 146 (2017) 350–359
kaolin into concrete mixes, but this introduces a more complex
production environment [5,24]. Previous research also examined
the thermal properties of cement-mortar composites using multi-
walled and single-walled carbon nanotubes (CNTs) [11,12,15,21]
and graphene oxide and graphene nanoplatelets [9,10], in compar-
ison with an ordinary mortar composite. Fundamental studies on
the carbon material itself were reported to determine the multi-
scale effects on strength and electrical and thermal conductivity
[4,9]. Muramatsu et al. [14] proposed a new method for synthesiz-
ing graphene from rice husk. Low-cost graphene materials could be
readily and cheaply synthesized on an industrial scale due to its
abundance. Rice husk ash has already received much attention as
an inclusion material for generating high-pozzolanic action inside
cementitious materials. Previous studies on mechanical properties
(compressive strength, electrical conductivity enhancement of
cement mortar composite mixed with CNT, graphene platelets,
graphene oxide [7,13,17–19], and rice husk-derived graphene
[16,20]) showed ample enhancement, with the aid of fumed silica
and a superplasticizer, compared with ordinary cement mortar. In
this study, we sought to explore the enhancement of the thermal
properties of cementitious mortar using rice husk-derived gra-
phene (GRH) and then performed feasibility tests of GRH mortar
in terms of bond strength between an existing concrete base and
a thin GRH overlay and freeze-and-thaw resistance under cyclic
environmental fatigue load. These three properties, conductive
medium, bond strength, and freeze-and-thaw resistance of thin
overlay, are the key to performance in actual application to a road
pavement surface.
(a)
(c)
Intensity (a.u.)
50nm
Fig. 1. (a), (b) Scanning electron microscopic (SEM) and (c) transmission electron microscopic (TEM) images of rice husk-derived graphenes and (d) its Raman spectrum
50 nm taken with a 532 nm laser line.
351
2. Rice husk-derived graphene materials
The graphenes used in this study was prepared via a chemical
activation of a typical agricultural waste (e.g., rice husk) with the
help of potassium hydroxide, as we demonstrated earlier
[14,16,20]. A typical synthetic procedure is described as follows.
Rice husk (10 g) was thermally treated at a temperature of
600 °C to obtain rice husk ash (RHA). Note that the yield of RHA
from rice husk is ca. 30%. Then, RHA (2.5 g) was mixed with KOH
(10 g) using a mortar and placed the mixture in an alumina pot,
which was fully covered with ceramic fiber wool. To generate char-
acteristics of graphenes from RHA, the alumina pot was encased in
a silicon carbide melting pot covered with carbon powder and
ceramic wool. The fully covered carbon powder was used to pro-
tect oxidation against air at high temperature. Then, the crucible
was thermally treated at 850 °C for 2 h in an atmospheric condi-
tion. The resulting powder (ca. 1 g) was washed with distilled
water for several hrs and then dried for 24 h under vacuum. Based
on SEM observation, the sample we synthesized exhibited particle-
like morphology with large sized holes (Fig. 1(a, b)), which is very
similar to that of the conventional activated carbons. To identify
the basic difference of our sample compared to the commonly
known activated carbons, we carried out TEM and Raman studies.
It is interesting to note that our sample consists of the large sized
but few layered graphenes with corrugated paper-like texture
(Fig. 1(c)). In addition, an intensified G’ band around 2850 cm 1
also supports the two or three layered stacking nature of our gra-
phene sample (Fig. 1(d)). Thus, our sample is relatively crystalline,
but few layered graphene with a high specific surface area (ca.
(b)
(d)
1000 1500 2000 2500 3000
Raman frequency (cm-1)
352 I. Rhee et al. / Construction and Building Materials 146 (2017) 350–359
2943 m2g 1), as previously demonstrated in Muramatsu et al. [14]
and Rhee et al. [16,20].
3. Thermal performances of GRH composite mortar
Two groups of carbon nanomaterials were used in this study.
The first consisted of various kinds of carbon nanomaterials [e.g.,
MWCNTs, COOH-decorated MWCNTs, graphene platelets (M15,
C650), PAN-based/pitch-based carbon fiber (Table 1)]. Control
specimens were selected: (a) MWCNT (>10 nm; Carbon NanoTech,
South Korea), (b) COOH-MWCNT (>50 nm; Cheaptube, USA), (c)
xGnPs M-15 (15:15 lm, 6 nm thickness; XG Science, USA) for ther-
mal conductivity, with a surface area ranging from 120 to 150 m2/
g, (d) xGnPs C-650 (650: 650 m2/g, 1–2 lm width, 2 nm thickness;
XG Science, USA) for electrical conductivity, with a surface area
ranging from 300 to 750 m2/g, (e) PAN-based carbon fibers (T300,
Toray Japan), and (f) pitch-based carbon fiber (diameter in the
range of 12–20 lm; GS Caltex, South Korea). Ordinary Portland
cement (Type 1), dry densified fumed silica (Elkem Materials,
Inc., Microsilica Grade 940-U), and ISO standard sands (SNL;
0.08–1.6 mm size), in accordance with EN 196-1, conforming to
ISO 679, were used for the mortar prism mix. The specific surface
area of the cement and fumed silica was 15–25 m2/g and the aver-
age diameter of a fumed silica particle was 0.15 lm. The by-
products of ultrafine particles of silicon dioxide are used in the pro-
cess of manufacturing silicon metal and ferro-silicon. Standard
sands, with controlled particle sizes, were distributed in the range
of 0.08–1.6 mm by sieving and used as fine aggregates in a dried
condition. A polycarboxylic acid-based admixture was applied to
enhance the flow of the carbon nanomaterials/cement composites
and to facilitate dispersion of the nanomaterials. In total, eight
mixtures (20  20  50 mm prisms) were prepared for Group #1
Table 1
Mix design for mortar composites (control group #1, 20  20  50 mm prism).
Sample No. Additives Weight (wt%) Cement (g) Sand (g)
P – – 30 24
GRH GRH 0.8 (0.24 g)
CF-pitch CF-pitch
CF-pan CF-pan
CG xGnP C650
MG xGnP M15
M MWCNT
COOH-C COOH-MWCNT
Fig. 2. The 50-mm prism of carbon mortar composite (section: 20 mm-sq., electrode: steel rod U 3 mm) and seven thermocouples were monitored under 60 A electrical
charge.
(Table 1). DC current (60 A) from a power supply was applied to
electrodes of each composite prism (Table 1). The total duration
time of current charge was 120 min. Six thermocouples (K-type,
copper-constantan thermocouples, max. 260 °C, 0.6  1.0 mm, wt.
3.33  10 3 g/mm) recorded temperature elevation as a function
of time with a digital multimeter (Agilent Technologies 34972A;
Fig. 2). Ambient temperature was also recorded during the test.
The first 6 min of charging showed rapid temperature rises in the
specimens. In the subsequent 54 min, the temperature was held
steady, under 60 A control, and finally in the last 60 min, the spec-
imens cooled down after turning off the power supply (Figs. 3 and
4).
In Fig. 3(a–c), the temperature change between the heat source,
Q, which was averaged from T5 and T6, onto the steel rod (elec-
trodes) and surface temperature, Ts, which was averaged from T1
and T2 onto the composite surface, is plotted against time.
Carbon-based material composites including GRH showed better
thermal conductive performance than other specimens. The smal-
ler range of temperature change, DT, indicated faster conduction
under the same heating conditions. The windowed graph of the
cooling process, between 60 and 120 min, is shown in Fig. 3(d).
The largest thermal shock, when the power was cut off, was
observed in the plain specimen and other specimens in a similar
range. The second test group consisted of GRH materials with dif-
ferent weight percentages, between 0 and 1.0 wt% (Table 2). In
total, five mixtures were prepared for Group #2. The size of the
specimens was enlarged, to a 50  50  100 mm prism (Fig. 4a).
Five different quantities of GRHs (0%, 0.25%, 0.5%, 0.75%, and 1.0%
by weight of cement) were added, and for each mixture, fumed sil-
ica amounting to 30% of the cement weight was applied to assist
possible dispersion inside the GRH/cement composites. The
amount of water in the GRH/cement composites was minimized
Water (g) Blast furnace slag (g) Viscosity agent (mL) Superplasticizer (mL)
15 9 – –
0.015 0.11
0.015 0.11
0.030 0.30
0.015 0.12
0.015 0.16
0.030 2.60
0.015 0.21
Top
Bottom
Bottom side Top side
: TS (Thermocouple Sensor)
Power supply
Data logger
Agilent 34972A
Ambient Temp.
 I. Rhee et al. / Construction and Building Materials 146 (2017) 350–359 353

50

between Heat Source, Q and Surface Temperature, Ts

Plain GRH

Temperature Gradient ( T)
CF-pitch CF-pan
40
M15 MW-COOH

30

20

10

0
0 25 50 75 100 125
-10
Time (min)
(a) (b)

50 0

between Heat Source, Q and Surface Temperature, Ts

between Heat Source, Q and Surface Temperature, Ts

45
Temperature Gradient ( T)

Temperature Gradient ( T)
40 -2
35
30 -4
25
20 -6
Plain GRH Plain GRH
15 CF-pitch CF-pan
CF-pitch CF-pan
M15 MW-COOH
10 M15 MW-COOH -8
5
0 -10
0 1 2 3 4 5 60 75 90 105 120 135
Time (min) Time (min)
(c) (d)
Fig. 3. Heat transfer test for Group #1: (a) test apparatus, (b) temperature gradient between heat source, Q ( = avg. of T5 and T6), and surface temperature, Ts ( = avg. of T1 and
T2), of the specimen, (c) magnified graph of (b) in the range of initial time (heating) within 5 min, and (d) magnified graph of (b) in the range of tail time (cooling) after 60 min.

6
9
5

7
4 3

2
8

1 10

(a) (b)
Fig. 4. The 100-mm prism of carbon mortar composite (section: 50 mm-sq., electrode: steel rod U 3 mm) and 10 thermocouples were used to monitor the 60 °C temperature.

Table 2
Mix design for mortar composites (control group #2: 50  50  100 mm prism).

Sample No. Weight (wt%) Cement (g) Sand (g) Water (g) Fumed silica (g) Superplasticizer (mL)
P – 180 440 94 54 2.64
GRH1 0.25 (0.45 g)
GRH2 0.50 (0.90 g)
GRH3 0.75 (1.35 g)
GRH4 1.00 (1.80 g)
354 I. Rhee et al. / Construction and Building Materials 146 (2017) 350–359

at 40% by the weight of the cement. Different doses of superplasti- reacted faster than the plain mortar specimen. The optimal wt%
cizer were applied to achieve a limited flow value in accordance was 0.75 wt% of GRH mortar, rather than 1.0 wt%. The compressive
with ASTM C1437-15. The mixing process was performed using a strength of the GRH mortar cube showed a similar optimal wt% at
standard Hobart-like mixer (Kenwood mixer), as specified in ASTM 0.75 wt% in our previous work. This could be because the material
C 305-14. All dried materials, including cement, aggregates, carbon was cast more densely when GRH was added at 0.75 wt%. This
nanomaterials, and fumed silica, were mixed for 10 min. Water and leads to more pozzolanic activity during the process of hydration.
admixture were added and then mixed again for an additional Fig. 5(c) shows the surface temperature, which was the average
5 min. The specimens were demolded from the copper mold after value of sensors #2 and #5, minus the ambient temperature his-
1 day of curing and then cured in water for another 27 days at tory. Fig. 5(d) shows the temperature gradients of sensors #2, #5
room temperature (21 °C). In previous tests, the GRH mortar spec- and #7. The lower temperature gradient indicates a smaller tem-
imens showed competitive performance when compared with perature change between heat source Q and the outer surface tem-
other carbon-based materials, such as graphene nanoplatelet and perature, Ts. The 0.75 wt% and 1.0 wt% specimens showed the
carbon fibers. We next tried to determine the optimal wt% of lowest temperature change distributions. This is consistent with
GRH mortar for heat conductivity. To this end, five mixes were pre- the heat source history (Fig. 5b). Table 3 shows the infrared ther-
pared, with 0, 0.25, 0.5, 0.75, and 1.0 wt% GRH (Table 2). In total, mal images of cross sections of specimens when the application
10 thermocouples were attached on the surfaces of the specimen time was 10, 25, 50, and 85 min.
(#1–#6), electrode (#7, #8, and #9), and ambient air (#10)
(Fig. 4b). Sensor #7 was attached in the center of the electrode, 4. Freeze-and-thaw resistance and bond strength of GRH
which was embedded in the mortar specimen. The heat source, composite mortar
Q, was applied continuously up to 60 °C at the sensor #7 location
at 1.0 °C/min. Fig. 5(a) shows three different heating rates: 0.2, A thin GRH overlay on an existing concrete pavement may
1.0, and 3.67 °C/min. With the lower heating rate, 0.2 °C/min, it assist in the early elimination of surface snow and ice, with the
took a considerable amount of time to reach 60 °C inside the prism. aid of efficient heat conductivity. To meet this ideal situation, the
In contrast, at the faster heating rate, 3.67 °C/min, the heating per- freeze-and-thaw resistance of the GRH cement mortar would be
formance among the different GRH specimens was hardly distin- an essential performance characteristic for durability in the harsh
guishable. Thus, we selected the 1.0 °C/min heating rate, which environment. To explore the freeze-and-thaw resistance of the
allowed a 4-h test period: 1 h of heating to 60 °C, another 1 h of GRH mortar composite, six different mixes were prepared in the
holding time at 60 °C, and finally 2 h of cooling time. Fig. 5(b) form of a 50-mm cube using fumed silica and superplasticizer
shows the temperature history of sensor #7 for five different spec- (Table 4). The water-to-binder ratio was set to 40%. Then, 0, 0.2,
imens. As the wt% of GRH was increased, the heat conductivity 0.4, 0.6, 0.8, and 1.0 wt% of GRH mortar specimen were cast. The
Temperature by Heating Rate, Q (#7)

Temperature at Heat Source, Q (#7)

70
65
60 60
(Heating up-Holding-Cooling down)

(Heating up-Holding-Cooling down)

55
50
50
40 45
30 40
35
20 Plain 0.25wt% 0.5wt%
0.75wt% (0.20°C/min) 1.00wt% (0.20°C/min) 30 0.75wt% 1.0wt%
10 0.75wt% (1.00°C/min) 1.00wt% (1.00°C/min)
25
0.75wt% (3.67°C/min) 1.00wt% (3.67°C/min)
0 20
0 1 2 3 4 5 0 0.5 1 1.5 2 2.5 3 3.5 4
Time (hours) Time (hours)
(a) (b)
35
Temperature Gradients between Q & Ts
Temperature at Surface, Ts-Ta

Plain 0.25wt% 15
(Avergaed surface temp. of T2 and T 5)

Plain 0.25wt% 0.5wt%

30 0.5wt% 0.75wt%
(Heating up-Holding-Cooling down)

0.75wt% 1.0wt%
1.0wt% 13
25
10
20
8
15
10 5

5 3

0 0
0 0.5 1 1.5 2 2.5 3 3.5 4 0 0.5 1 1.5 2 2.5 3 3.5 4
Time (hours) Time (hours)
(c) (d)
Fig. 5. (a) Test apparatus, (b) temperature gradient between heat source, Q ( = T7), (c) surface temperature, Ts ( = avg. of T2 and T5) of the specimen, and (d) temperature
gradients between Q and Ts.
 I. Rhee et al. / Construction and Building Materials 146 (2017) 350–359 355

Table 3
Thermal gradient distribution at different times due to heat conduction using an infrared camera (FLIR E8).

10 min 25 min 50 min 85 min

Plain

GRH
0.25 wt%

GRH
0.50 wt%

GRH
0.75 wt%

GRH
1.00wt%

Table 4
Mix design for GRH mortar composites (control group #3: 50-mm cube).

Sample no. Weight (wt%) Cement (g) Sand (g) Water (g) Fumed silica (g) Superplasticizer (mL)
P – 280 686 147 84 4.1
GRH1 0.2 (0.56 g)
GRH2 0.4 (1.12 g)
GRH3 0.6 (1.68 g)
GRH4 0.8 (2.24 g)
GRH5 1.0 (2.80 g)

Thermometer

Specimens
Thermocouple

Freezer Thermoregulator
( -20°C) Pan heater

Timer controler

(a) (b)
Fig. 6. Freeze-and-thaw resistance test. (a) In-house freeze-and-thaw machine (temp. range: 24 to 15 °C, single cycle: 4 h), (b) 50-mm cubes of GRH mortar composite
under 100 cycles of the freezing-and-thawing process.
356 I. Rhee et al. / Construction and Building Materials 146 (2017) 350–359

size and shape of these specimens did not meet the ASTM specifi- single cycle consisted of (a) ramping down for 30 min, from 15 °C
cation for the standard freeze-and-thaw test method. Limited and holding at 24 °C for freezing, and (b) ramping up for
amounts of GRH were synthesized so that the relatively large spec- 30 min, from 24 °C and holding at 15 °C for 75 min for thawing.
imen needed for the standard test could not be used. An in-house Each freeze-and-thaw process had a 15-min power-off time before
freeze-and-thaw machine was prepared (Fig. 6a). The temperature turning from freezing to thawing, and vice versa. In total, 18 spec-
change for the freeze-and-thaw process was 24 to 15 °C. The 4-h imens (Fig. 6b) were cast, cured in water for 28 days, and used for
this test. Because three specimens were available per mix (Table 4),
two specimens of each mix were placed in the freezing-and-
120
thawing machine for cyclic temperature tests. The last specimen
Compressive Strength (MPa)

100 of each mix was placed and cured in a laboratory environment

(21 °C and 70% RH). In Fig. 7, the compressive strength of the 50-
80 mm cube GRH composite after 300 cycles of the freeze-and-thaw
test and an environmentally undamaged specimen, cured in the
60
laboratory environment, is plotted. The strength reduction was
40 Aer 300 cycles #1 estimated using the compressive strength of the undamaged spec-
Aer 300 cycles #2 imen: plain (16.7%), 0.2 wt% (23.4%), 0.4 wt% (25.1%), 0.6 wt%
Cured under room env.
20 Aer 300 cycles (avg.) (17.7%), 0.8 wt% (17.6%), and 1.0 wt% (9.4 wt%). The ‘best’ perfor-
mance among the specimens was identified for the 0.6 wt% GRH
0 composite because of the higher strength of the undamaged spec-
0 0.2 0.4 0.6 0.8 1.0
wt% imen and the smaller reduction in the damaged specimen due to
the freeze-and-thaw process.
Fig. 7. Compressive strength of 50-mm cube GRH composite after 300 cycles of XRD test results for plain and 0.4 wt% specimens cured in the
freeze-and-thaw test and environmentally undamaged specimen cured in the laboratory environment and 0.4 wt% specimen after 300 cycles of
laboratory environment.
Q

Plain (cured in lab env.)

0.4wt% (freeze-and-thaw w fo
for 300 cycle)
Diffracted Intensity (a.u.)

0.4wt% (cured in lab env.)v.)

Q

E+CSH
E+CSH
B+E+CSH
E + P + CSH

B+E+CSH
Q
E+CSH
CSH
E+CSH
E+CSH
P+CSH

CSH
Q
E+B
C
B

Q
E

5 10 15 20 25 30 35 40 45 50 55 60
Position(2θ CuKα)
(a)

Plain (cured in lab env.)

CSH

Eringite + CSH

0.4wt% (fre
(freeze-and-thaw for 300 cycle)
0.4wt% (cur
(cured in lab env.)
Diffracted Intensity (a.u.)
Diffracted Intensity (a.u.)

Belite

28 28.5 29 29.5 30 30.5 31 31.5 32 49 49.5 50 50.5 51 51.5 52

Position(2θ CuKα) Position(2θ CuKα)
(b) (c)
Fig. 8. (a) XRD test results for plain and 0.4 wt% specimens cured under laboratory conditions and 0.4 wt% specimen after 300 cycles of the freeze-and-thaw process,
amorphous CSH, (b) ranged from 28.5 to 31° (2h), (c) with a narrower full width at half-maximum (FWHM) ranging from 50 to 50.8° (2h).
 I. Rhee et al. / Construction and Building Materials 146 (2017) 350–359 357

the freeze-and-thaw process are depicted in Fig. 8(a) between 5 Table 5

and 60°. These results were taken 120 days later (50 days for the Averaged Ca/Si ratio from SEM-EDS analyses.

freeze-and-thaw test) from cast specimens. Most C3S (3CaO
SiO2, Specimen Air-dried Freeze-and-thaw cycled
alite) was transformed to CSH with other hydrates, such as ettrin- Plain 0.4 wt% 1.0 wt% Plain 0.4 wt% 1.0 wt%
gite and/or portlandite and scattered in various ranges. Amorphous
Avg. Ca/Si 1.13 1.30 1.31 1.24 1.42 1.36
CSH ranging from 28.5 to 31° (2h), was identified, and had a nar-
rower full width at half-maximum (FWHM) in the range of 50 to
50.8° (2h) as shown in Fig. 8(b)–(c). The development of CSH in
the environmentally damaged 0.4 wt% (green line) specimen was cement weight, was added to enhance the strength and durability.
lower than in the other undamaged 0.4 wt% and plain specimen. The fumed silica (microsilica) contains a considerable proportion of
This might explain the loss of compressive strength observed in Si in CSH form [24,25]. Thus, the Ca/Si ratio of these cube speci-
the damaged specimens. Some C2S (2CaO
SiO2, belite) was noticed mens was reduced versus those of ordinary hydrates. The lower
in the range of 31 to 31.5° (2h) and the damaged 0.4 wt% specimen Ca/Si ratio may be related to the higher compressive strength. It
showed less than the other specimens. is partially reflected in the compressive strength (Fig. 7). In
Backscattering electron microscope method is generally used Figs. 9(c) and (d), the Ca/Si ratio of the damaged specimens shifted
for extracting CSH from other hydration products or unreacted in the higher Ca/Si ratio direction and, for the undamaged speci-
cement. For the simplicity, SEM-EDS analysis used to measure men, a lower Ca/Si ratio was seen. This tendency was also consis-
the atomic wt% of CSH. However, this method cannot be decoupled tent with the compressive strength results in Fig. 7.
CSH fully among other hydrates and this ‘by-inspection’ approach A standard test method for assessing the bond strength of
much depends on the researcher’s personal experience. In order to epoxy-resin systems used with concrete by slant shear (ASTM C-
mitigate this possible error, several random windowed regions 882, [23]) was adopted to evaluate the bond strength of GRH com-
were analyzed to evaluate the weight % and atomic % of Ca and posites. The new material was used as bonding medium for two
Si. Fig. 9(a)–(b) shows CSH and GRH clusters embedded in the pieces of concrete (Fig. 10). This test is important for the applica-
hydrates. Fig. 9(c) and Table 5 show that the Ca/Si ratio varied tion of a thin bonded overlay on a concrete pavement. A modified
among the three mixes depending on the undamaged/damaged version of the testing method of ASTM C-882 was applied using a
status. Normally, the Ca/Si ratio is close to 1.7 [1–3,6,25], but it var- 46 ± 0.6° slant surface and a 10-mm-thick GRH mortar composite
ied from 1.24 to 1.42. This may have resulted from the addition of overlay, as shown in Figs. 10(a) and (b). In total, 15 slanted con-
fumed silica (Table 4). Indeed, 30% fumed silica, compared with the crete cylinders were prepared with different GRH wt%: 0, 0.25,

(a) (b)
65 1.7
P (F/T cycled) Air-dried
avg. Ca/Si =1.42 P (air-dried)
60 0.4wt% (F/T cycled) 1.6 F/T cycled
0.4wt% (air-dried)
Averaged Ca/Si Ratio

55 1.0wt% (F/T cycled)

1.5
1.0wt% (air-dried)
Ca (Atomic %)

50 avg. Ca/Si =1.30

1.4
45
1.3
40
avg. Ca/Si =1.31 1.2
35
avg. Ca/Si =1.36
30 avg. Ca/Si =1.13 1.1
avg. Ca/Si =1.24
25 1.0
25 30 35 40 45 Plain 0.4wt% 1.0wt%
Si (Atomic %)
(c) (d)
Fig. 9. SEM image of (a) GRH 0.0 wt%, (b) GRH 0.4 wt% of damaged specimen, (c) Ca/Si ratio of GRH composites and (d) averaged Ca/Si ratio with standard deviations.
358 I. Rhee et al. / Construction and Building Materials 146 (2017) 350–359

Ao = π (1002) / 4 = 7,854mm2

Concrete Φ 100mm

GRH mortar
Thin layer

θ = 46 ± 0.6

y
71 t2
x Ainclined = 4 × ∫ 50 1 − dt = 11,153mm 2
0 712

(a) (b)

(c) (d)
Fig. 10. Bond failure test: (a) slanted surface and dimensions of cylinder, (b) slanted surface versus original surface, (c) compressive strength test, and (d) partially failed
bonding surface.

Table 6
Mix design for mortar composites (control group #4: U 100 mm  200 mm).

Sample no. Weight (wt%) Cement (g) Sand (g) Water (g) Silica fume (g) Superplasticizer (mL)
P – 280 686 147 84 3
GRH1 0.25 (0.7 g)
GRH2 0.50 (1.4 g)
GRH3 0.75 (2.1 g)
GRH4 1.00 (2.8 g)

400
Bond Strength by Slanted Shear (MPa)

0.5, 0.75, and 1.0 wt%. Each mix was used for three specimens
(Table 6). After 28 days of curing in water, compressive tests were 350
performed, as shown in Fig. 10(c) and (d). The bond strength was
300
calculated from the failure load divided by the detached slanted
area. Most specimens showed fully debonding shear failure 250
(Fig. 10c) and some showed partially debonding shear failure
(Fig. 10d). From digital images of the failure surface, the effective 200
bond failure was evaluated and used for calculating the bond
strength, which is failure load divided by effective bond failure 150
area. Fig. 11 shows the bond strength variation with respect to
wt% of GRH added. The bond strengths with the same mix exhib-
100
ited widespread results. Nevertheless, the tendency of average
50
bond strength showed better performance under a 0.75 wt% GRH 1 2 3 avg
mortar overlay. Beyond 0.75 wt% inclusion, bond strength may 0
decrease. This result is consistent with the compressive strength P 0.25wt% 0.5wt% 0.75wt% 1.0wt%
sensitivity of GRH mortar identified in previous work.
Fig. 11. Bond strength of concrete cylinder with 10-mm-thick GRH mortar overlay.
 I. Rhee et al. / Construction and Building Materials 146 (2017) 350–359
5. Conclusions
From the results of this study, we reached three main
conclusions:
1) The thermal performance of GRH mortars was explored
experimentally using two test groups: (a) with other
carbon-based materials, such as carbon nanotubes, carbon
fibers, and graphene platelets, and (b) with the addition of
different wt% levels of GRH. Of those tested, the 0.75 wt%
GRH prism showed the most efficient heat transfer (0–
1.0 wt% specimens). This 0.75 wt% GRH mix could have the
optimal wt% to satisfy the compressive strength enhance-
ment [16], electrical properties [20], and thermal perfor-
mance. The heat transfer ability, as a conductive medium,
was fair, in comparison with other carbon-based materials.
Thus, GRH could be a useful alternative conductive filler
for cementitious materials.
2) Freeze-and-thaw resistance tests of GRH mortar under 300
cycles of temperature change between z24 and 15 °C for
50 days showed compressive strength reductions: plain
(16.7%), 0.2 wt% (23.4%), 0.4 wt% (25.1%), 0.6 wt% (17.7%),
0.8 wt% (17.6%), and 1.0 wt% (9.4%). The ‘best’ performance
among the specimens tested was for the 0.6 wt% GRH com-
posite. The 0, 0.4, and 1.0 wt% specimens were used to
examine the development of CSH. According to XRD/phase
analyses, damaged specimens showed lower development
of these CSHs. Another aspect of the strength reduction after
300 cycles of the freeze-and-thaw process is that the Ca/Si
ratio was higher for damaged specimens, according to the
SEM-EDS analyses.
3) A modified slanted shear test with a 10-mm thin GRH over-
lay was used to evaluate bond strength. Bond strengths with
the same mix showed widespread results. Nevertheless, the
tendency of average bond strength showed better perfor-
mance under a 0.75 wt% GRH mortar overlay. Beyond
0.75 wt%, bond strength may decrease. This result is consis-
tent with the compressive strength sensitivity of GRH mor-
tar identified in previous work.
Acknowledgments
The first author acknowledges financial support under research
grant no. 2016-1033 from the National Research Funds, Korean
Ministry of Education.
References
[1] J.J. Thomas, J.J. Chen, H.M. Jennings, Ca-OH bonding in the C-S-H gel phase of
tricalcium silicate and white Portland cement pastes measured by inelastic
neutron scattering, Chem. Mater. 15 (2) (2003) 3813–3817.
359
[2] S. Hoshino, K. Yamada, H. Hirao, XRD/Rietveld analysis of the hydration and
strength development of slag and limestone blended cement, J. Adv. Concr.
Technol. 4 (3) (2006) 357–367.
[3] F. Pelisser, P.J.P. Gleize, A. Mikowski, Effect of the Ca/Si molar ratio on the
micro/nanomechanical properties of synthetic C-S-H measured by
nanoindentation, J. Phys. Chem. 116 (2012) 17219–17227.
[4] E. Pop, V. Varshney, A.K. Roy, Thermal properties of graphene: fundamentals
and applications, Mater. Res. Soc. 37 (2012) 1273–1281.
[5] F. Pelisser, A. Barcelos, D. Santos, M. Peterson, A.M. Bernardin, Lightweight
concrete production with low Portland cement consumption, J. Clean. Prod. 23
(2012) 68–74.
[6] S. Grangeon, F. Claret, Y. Linard, C. Chiaberge, X-ray diffraction: a powerful tool
to probe and understand the structure of nanocrystalline calcium silicate
hydrates, Acta Crystallogr., Sect. B 69 (2013) 465–473.
[7] I. Rhee, Y.S. Roh, Properties of normal strength concrete and mortar with
multi-walled carbon nanotubes, Mag. Concr. Res. 65 (16) (2013) 951–961.
[8] J. Heo, B. Park, T. Kim, A feasibility study on developing snow melting system
using CNT-cement composite, Int. J. Highway Eng. 15 (2) (2013) 29–37.
[9] A. Sedaghat, M.K. Ram, A. Zayed, R. Kamal, N. Shanahan, Investigation of
physical properties of graphene-cement composite for structural applications,
Open J. Compos. Mater. 4 (2014) 12–21.
[10] G. Fugallo, A. Cepellotti, L. Paulatto, M. Lazzeri, N. Marzari, F. Mauri, Thermal
conductivity of graphene and graphite: collective excitations and mean free
paths, Nano Lett. 14 (11) (2014) 6109–6114.
[11] Y.H. Zhao, Effect of CNT/CNF on thermal and mechanical properties of cement
mortars, Adv. Mater. Res. 1049–1050 (2014) 234–237.
[12] M.C. Camacho, O. Galao, F.J. Baeza, E. Zornoza, P. Garces, Mechanical properties
and durability of CNT cement composites, Materials 7 (2014) 1640–1651.
[13] X. Wang, I. Rhee, Y. Wang, Y. Xi, Compressive strength, chloride permeability
and freeze–thaw resistance of MWNT concretes under different chemical
treatments, Sci. World J. (2014) 1–8.
[14] H. Muramatsu, Y.A. Kim, K.S. Yang, R. Cruz-Silva, I. Toda, T. Yamada, M.
Terrones, M. Endo, T. Hayashi, H. Saitoh, Rice husk-derived graphene with
nano-sized domains and clean edges, Small 10 (14) (2014) 2766–2770.
[15] D.A. Exarchos, P.T. Dalla, I.K. Tragazikis, T.E. Matikas, The effect of CNTs
reinforcement on thermal and electrical properties of cement-based materials,
Proc. SPIE 9436 (2015).
[16] I. Rhee, Y.A. Kim, G.O. Shin, J.H. Kim, H. Muramatsu, Compressive strength
sensitivity of cement mortar using rice husk-derived graphene with a high
specific surface area, Constr. Build. Mater. 96 (2015) 189–197.
[17] B.A. Al Muhit, B.H. Nam, L. Zhai, Zuyus, Effect of microstructure on the
compressive strength of graphene oxide-cement composites, Nanotechnol.
Constr. (2015) 1–9.
[18] H.K. Kim, Chloride penetration monitoring in reinforced concrete structure
using carbonnanotube/cement composite, Constr. Build. Mater. 96 (2015) 29–
36.
[19] B. Wang, R. Jiang, Z. Wu, Investigation of the mechanical properties and
microstructure of graphene nanoplatelet-cement composite, Nanomaterials 6
(2016) 1–15.
[20] I. Rhee, J.S. Lee, Y.A. Kim, J.H. Kim, J.H. Kim, Electrically conductive cement
mortar: incorporating rice husk-derived high-surface-area graphene, Constr.
Build. Mater. 125 (2016) 632–642.
[21] G.M. Kim, F. Naeem, H.K. Kim, H.K. Lee, Heating and heat-dependent
mechanical characteristics of CNT-embedded cementitious composites,
Compos. Struct. 136 (2016) 162–170.
[22] P. Pan, S. Wu, X. Hu, G. Liu, B. Li, Effect of material composition and
environmental condition on thermal characteristics of conductive asphalt
concrete, Materials 10 (2017) 1–17.
[23] ASTM C882/C882M, Standard Test Method for bond Strength of Epoxy-Resin
Systems Used with Concrete by Slant Shear, ASTM International, PA, USA.
[24] A. Brykov, M. Voronkov, M. Mokeev, Ultrafine silica additives behavior during
alkali-silica reaction long-term expansion test, Mater. Sci. Appl. 5 (2014) 66–
72.
[25] A.C.A. Muller, K.L. Scrivener, J. Skibsted, A.M. Gajewicz, P.J. McDonald,
Influence of silica fume on the microstructure of cement pastes: new
insights from 1H NMR relaxometry, Cem. Concr. Res. 74 (2015) 116–125.
[26] S. Wu, L. Mo, Z. Shui, Z. Chen, Investigation of the conductivity of asphalt
concrete containing conductive fillers, Carbon 43 (7) (2005) 1358–1363.