8/8/2019 Mechanical characteristics of self-compacting concretes with different filler materials, exposed to elevated temperat
1/13
O R I G I N A L A R T I C L E
Mechanical characteristics of self-compacting concretes
with different filler materials, exposed to elevatedtemperatures
N. Anagnostopoulos K. K. Sideris A. Georgiadis
Received: 7 April 2008 / Accepted: 11 December 2008 / Published online: 24 December 2008
RILEM 2008
Abstract In this paper, the studies concern the
influence that different fillers have on the properties
of SCC of different strength classes when exposed to
high temperatures. A total of six different SCC and
two conventional concrete mixtures were produced.
The specimens produced are placed at the age of
180 days in an electrical furnace which is capable of
reaching 300C at half an hour and 600C at 70 min.
The maximum temperature is maintained for an hour.
Then the specimens are let to cool down in the
furnace. The hardened properties measured after fireexposures are the compressive strength, splitting
tensile strength, water capillary absorption and the
ultrasonic pulse velocity. Explosive spalling occurred
in most cases when specimens of higher strength
class are exposed to high temperatures. The spalling
tendency is increased for specimens of higher
strength class C30/37 irrespective of the mixture
type (SCC or NC) and the type of filler used.
Keywords Filler materials Glass filler
Mechanical characteristics Self-compacting concrete Slag Temperature
1 Introduction
1.1 General
An SCC is due to its various advanced properties
most useful regarding to the structure industry. The
ability to self compact without the use of any vibrator
allows SCC to pass through dense reinforcement and
fill in restricted sections, guaranteeing time superior
quality of the cast structure at the same. Moreover the
fact that the compaction takes place while casting,without any further delay, ensures a tight and
accurate construction schedule. The feature of self
consolidation is partly based on a new method for the
production and quality control of SCC [1] which
involves lower water to binder ratio, accumulates the
use of filler materials and the addition of superplast-
icizer in order to achieve the desired workability.
Many scientists have reported the similarity of SCC
with high performance concrete (HPC), which is also
produced with decreased water to cement ratio and
certain chemical admixtures [2, 3]. The problem thatoccurs is the behavior of such concrete mixtures
when exposed to high temperatures.
In general, concrete as a building material has a
reasonably good fire resistance. But when SCC or
HPC is used there are some complications. These
complications concern microstructure changes which
grow along with the increasing temperature [4]. At
certain temperatures there is apparent deterioration
mostly due to the dehydration of CSH gel and the
N. Anagnostopoulos K. K. Sideris (&) A. GeorgiadisLaboratory of Building Materials, Democritus University
of Thrace, P.O. Box 252, Xanthi 67100, Greece
e-mail: [email protected]
Materials and Structures (2009) 42:13931405
DOI 10.1617/s11527-008-9459-6
8/8/2019 Mechanical characteristics of self-compacting concretes with different filler materials, exposed to elevated temperat
2/13
increasing pore water pressure. The finer pore
distribution along with the poor pore connectivity
that characterizes SCC and HPC keeps the free and
chemically bound water trapped inside the structure,
leading to growing pore pressure [2, 3, 5]. When high
temperature and high heating rate are applied, the
concretes fire resistance is most likely to decreaseand thus spalling to occur.
1.2 Objective
Considering that SCC is a newer type of concrete
compared to the traditional concrete or even HPC, the
research performed on SCC after fire exposure is yet
limited. As reported in [2] the SCC mixtures which
are subjected to fire have an explosive spalling
tendency which is evident in concrete mixes of higher
strength classes, while SCC of lower strength classeshas a rather good fire resistance. In this contribution
the efforts are focused on producing SCC of different
strength classes which incorporate different filler
materials, in order to investigate their performance
after exposure at gradually up scaled temperature.
2 Materials and methods
2.1 Materials and mixtures
A total of six SCC and two NC mixes are produced
for this study. The same class of blended cement
(CEM II 42, 5 N) is used in all cases to produce
strength classes such as C25/30 and C30/37, accord-
ing to EN206-1 [6]. Coarse aggregates consisting of
crushed granite and limestone sand are used. A high
range water reducing carboxylic either polymer
admixture is added in different dosages to achieve
slump of 190 mm in the case of NC, or self
compactibility in the case of SCC. The filling
materials used for the production of all SCC mixesare respectively: limestone filler, slag and glass filler.
The cement-filler material chemical compositions as
well as the aggregate grading curves are listed in
Tables 1, 2 respectively. In all cases the water/
cement ratios as well as the cement content are kept
relatively the same for each strength class. Moreover
the slump flow tests and slump tests with reference to
SCC and NC correspondingly were attempted to be of
the same order of value and thus to present respective
properties while in fresh state. The mix proportions of
all concretes are presented in Table 3.
2.2 Specimens and temperatures
The mixing is carried out in a pan mixer according tothe European Guidelines for SCC [7]. Right after the
mixing is completed the SCC is tested accordingly as
instructed in EFNARC specifications [8]. A number
of 150-mm cubes are prepared in order to assess the
compressive strength and the water capillary absorp-
tion at the age of 28 days. The water capillary
absorption is measured according to the procedure
described by RILEM TC116 [9]. The 150-mm cubes
are tested for compressive strength after a period of
Table 1 Chemical composition of cement and filling materials
(%)
Sample CEMII-A/M
42.5 N
Limestone
filler
Ladle
furnace slag
Glass
filler
SiO2a 23.85 1.8 32.5 62.1
Al2
O3
5.22 0.45 2.5 1.6
Fe2O3 4.13 0.08 0.1
FeO 1.72
CaO 58.2 54.8 54.1 18
MgO 3.2 0.68 5.55 2.4
SO3 3.3 0.05 0.2
K2O 0.68 0.04 0
Na2O 0.32 0.34 12.4
TiO2 0.24 0.17
P2O5 0.06 0.02 0.1
SrO 0.03
Cr2O3 0.02 ZnO 0.01
LoIb 1.57 40.5 3.19 0.4
SGc (g/cm3) 3.1 2.65 2.59 2.51
a All the samples are expressed by weigh percentageb Loss of ignitionc Specific gravity
Table 2 Composition of materials used (%)
Aggregate passing 0.25 0.5 1 2 4 8 16 32Limestone sand 6 18 56 81 100 100 100 100
Coarse aggregates 0 0 0 0 12 57 100 100
All fine materials pass through the 0.125-mm sieve
1394 Materials and Structures (2009) 42:13931405
8/8/2019 Mechanical characteristics of self-compacting concretes with different filler materials, exposed to elevated temperat
3/13
28 days of moist curing (20 2C, RH C 95%), as
the mean value of three specimens. Fire resistance is
measured on 100-mm cubes and 150 9 300-mm
cylinders. Those specimens also go through moist
curing for 28 days and are then left at ambience
(20 2C, RH C 65%), and not tested till the age of
fire tests (180 days).
Right before fire exposure at the age of 180 days
three 100-mm cubes of each mixture are dried to
constant mass at 105C. Table 4 shows the moisture
content that is afterwards determined as following:
W= (m0 - md)/mdmd = mass of the test specimen after drying at
105C
m0 = mass of the specimen before drying
At the age of 180 days the specimens which meant
to be exposed to high temperatures are heated in an
Table 3 Mix design proportions and fresh properties of self-compacting concretes (SCC) and normal concretes (NC)
Mix design proportions (kg/m3) Self compacting concrete Normal concrete
L-filler Slag Glass filler
SCC SCC SCC SCC SCC SCC NC NC
C25/30 LF C30/37 LF C25/30 SL C30/37 SL C25/30 GF C30/37 GF C25/30 NC C30/37 NC
CEMII-A/M 42.5 N 335 375 340 375 340 380 330 375
Filler 135 100 0 0 0 0 0 0
Slag 0 0 135 100 0 0 0 0
Glass filler 0 0 0 0 130 100 0 0
Sand 915 900 825 862 845 862 940 870
Coarse aggregates 800 800 800 800 800 800 927 955
Water 185 186 188 189 190 194 183 186
Super plasticizer (%)a 1.63 1.88 1.29 1.74 1.16 1.17 1.0 1.0
W/C 0.55 0.50 0.55 0.50 0.56 0.51 0.55 0.50
W/P 0.39 0.39 0.40 0.40 0.40 0.40 0.55 0.50
Air content (%) 1.70 1.60 1.90 1.70 1.40 1.20 2.10 1.80Slump (cm) 19 20
Slump flow D (cm) 75.5 75 75.5 75.5 74 73.5
t50 (s) 2 1.72 4.72 4.25 1.66 1.25
V funnel 1 (s) 10.5 10 8.49 9.18 4.38 6.06
V funnel 2 (s) 28 15 14.4 11.25 5.16 13
J ring H (cm) 0.3 0.3 1 0.9 0.6 0.5
J ring D (cm) 68 68 67 68 66 68
LBOX (h2/h1) 0.88 0.88 0.83 0.85 0.82 0.84
t200 (s) 1 1 2.5 3.41 1.2 1.35
t400 (s) 2.01 3 5.5 5.1 1.4 2.25
fc28 (Mpa)b 37.1 54 37.7 53.5 38.3 49 36 52.7
a SP (super plasticizer) value is measured by % percent by weight of the entire fines amount (cement and filler materials)b
Compressive strength at the age of 28 days is measured in specimens of 150-mm edge cubes
Table 4 Moisture content of all the mixtures at the age of 180 days
Mixture C25/30
CC-LF
C30/37
SCC-LF
C25/30
SCC-SL
C30/37
SCC-SL
C25/30
SCC-GF
C30/37
SCC-GF
C25/30
NC
C30/37
NC
Moisture content (%)a 4.41 3.39 3.17 3.18 4.01 4.11 3.98 3.93
a Moisture content is measured in specimens of 100-mm edge cubes
Materials and Structures (2009) 42:13931405 1395
8/8/2019 Mechanical characteristics of self-compacting concretes with different filler materials, exposed to elevated temperat
4/13
electrical furnace Fig. 1. Two peak temperatures are
examined to determine the specimens fire resistance:
300 and 600C. The heating rate applied is 10C/min
until the target temperature is reached, and this is
maintained for a period of 1 h, Fig. 2. When the
heating period finishes, the furnace remains sealed for
24 h in order to cool down the specimens down to theambient temperature. Then the specimens are tested
to determine properties such as compressive strength,
splitting tensile strength and pulse velocity.
2.3 Compressive strength
The compressive strength is measured according to
EN 12390-3 [9]. The original compressive strength is
measured on 150-mm cubes which are tested at the
age of 28 days in order to specify the concretes
strength class. Residual compressive strength ismeasured after the fire tests on 100-mm cubes. A
Buehl & Fabel compression testing machine with
3000 KN capacity is used in all cases.
2.4 Splitting tensile strength
Splitting tensile strength is determined by measuring
the tensile strength on 150 9 300-mm cylinders at
different peak temperatures so that the residual
tensile strength is assessed in each case. According
to EN 12390-6 [10] a splitting attachment (CON-TROLS Model 50-C9000) is adjusted on the
laboratory compression testing machine.
2.5 Water capillary absorption
The water capillary absorption is measured according
to the procedure described by RILEM TC116 [11].
That property is measured on pre-weighted 150-mm
cubes. Specimens are placed on adjusted plastic
plates filled with water, so that only one surface of the
specimen is getting wet. Then the specimens areweighted in regular intervals and the absorbed water
quantity is estimated.
2.6 Stressstrain curves
100-mm cubes are used in order to calculate stress
strain curves at the age of 180 days. The specimensFig. 1 Indicative seating plan of the specimens in the furnace
Fig. 2 Temperature development of 300C (left) and 600C (right)
1396 Materials and Structures (2009) 42:13931405
8/8/2019 Mechanical characteristics of self-compacting concretes with different filler materials, exposed to elevated temperat
5/13
are placed on the laboratory compression testing
machine and stressstrain sensors are adjusted on the
upper metallic plate of the pressing device. As soon
as the specimen is loaded, the sensors transmit
electric signal to the data logger which is converted
through a computer programme into stressstrain
curves.
2.7 Pulse velocity
Pulse velocity is measured on 100-mm cubes accord-
ing to the procedure described by EN 12504-4 [12].
Specimens are tested at the age of 180 days before
and after fire tests using a PUNDIT ultrasonic pulse
velocity testing device.
3 Results and discussion
3.1 Original compressive strength
The compressive strength results at the age of
28 days are presented at Table 3 for all prepared
concrete mixes. While studying the compressive
strength results it emerges that in almost all cases
SCC develop higher values as compared with NC of
the same strength class. This is attributed to the
changes of the interfacial transition zone (ITZ)
caused by the different filler materials [13]. Asreported in Zhu and Bartos [14] ITZ is denser and
significantly more uniform in SCC than in NC.
Moreover as Traghard points out, the porosity of ITZ
is much lower in SCC than in NC of the same w/c
ratio, as the hydrated phases and unhydrited particles
appear to be more evenly distributed between the ITZ
and bulk density of SCC [15].
In the case of SCC, it appears that there are slight
variations as it regards their compressive strength
values. The use of different filler materials for SCC
mixture production has everything to do with these
deviations. SCC mixtures which are produced usingladle furnace slag as a filler material have higher
water absorption than expected, resulting to a viscous
and rather slow concrete, while in fresh state. On
the other hand, when limestone filler is used the
mixture performance is excellent in terms of rheo-
logical and mechanical characteristics. Similar
performance is noted in the case of glass filler with
even better rheological features. As mentioned
before, all mixtures are produced by keeping the w/c
ratio relatively the same. That means that in relation to
their absorption requirements, certain porosity isdeveloped in each case, which is finally reflected in
the compressive strength values.
3.2 Residual compressive strength
What would be of great importance in a fire scenario
is definitely the state of the concretes mechanical
properties. The residual compressive strength for all
SCC and NC mixtures after heating in 300 and 600C
is presented in Fig. 3. As Chan reports in his
investigation there are three temperature ranges fromthe point of strength loss: 20400, 400800 and
8001200C [2, 5]. After exposing to fire HPC
and NC mixtures prepared with ordinary Portland
cement, Chan concludes that only a small part of the
original strength is lost up to 400C, while severe
Fig. 3 Compressive strength of C25/30 (left) and C30/37 (right) SCC and NC, after heating at different temperatures (300 and
600C) compared with the compressive strength at room temperature (20C)
Materials and Structures (2009) 42:13931405 1397
8/8/2019 Mechanical characteristics of self-compacting concretes with different filler materials, exposed to elevated temperat
6/13
compressive strength loss occurs within the 400
800C range. That is mostly the case for the SCC and
NC mixtures prepared in this contribution, which are
exposed to slightly different temperature ranges,
though.
Regarding the 20300C range, there is a com-
pressive strength reduction for the SCC 25/30 whichfluctuates between 12% and 15% of its initial value,
while for the SCC 30/37 the strength loss percentage
reaches 18%. The corresponding strength loss per-
centages for the equivalent 25/3030/37 NC are 18%
and 17.6% respectively. At 600C the reduction in
compressive strength ranges from 52% to 57% for all
mixtures and explosive spalling occurs in cylindrical
specimens in all cases. The phenomenon is more
intense when ladle furnace slag is used as filler
material and spalling occurs in 100-mm cubes as
well. As there was no way for visible inspection,spalling is identified by hearing the series of pop outs,
happening in most cases when the distinctive tem-
perature of 300C is overrun.
The use of different filler materials in the case of
SCC does not seem to make any difference as it
regards explosive spalling Fig. 4. During fire the
humidity of the concrete increases along with tem-
perature and fluid water is formed [16]. The water is
transported inwards to the center of the specimen
where the space is limited [16]. At a certain point the
region becomes saturated and the entrapped water iseventually released in the form of steam by explosive
spalling [17, 18]. The filler content used is relatively
high (150 kg/m3) to ensure low levels of concrete
moisture and eventually to avoid spalling due to
steam pressure. The concept of using glass filler is the
formation of micro-cracks which develop because of
the thermal expansion of glass that is greater than the
concretes [16]. Thus greater pore connectivity is
thought to provide canals for the steam to escape.
Ladle furnace slag is used as a filler material as itpresents cementitious behavior mainly due to its high
content in CaO [19]. According to Piasta et al. the
development of micro-cracks increases beyond
300C and firstly occurs around calcium hydroxide
Ca(OH)2 crystals [20]. Hence slag due to its cemen-
titious properties is expected to ensure the
development of micro-cracks which will lead to
greater porosity.
3.3 Moisture content and water absorption
Since the use of different filler materials is expected
to alter the porosity of SCC mixtures and thus to
prevent the effect of spalling, the moisture content
and the capillary water absorption for all mixtures are
presented in Tables 4 and 5 respectively. As Bostrom
et al. points out, SCC has a high probability of
spalling when exposed to fire compared to conven-
tional concrete [21]. Considering the low
permeability of SCC due to its denser structure,
water vapor is very limited to evaporate out of SCC.
Lower moisture content is therefore of great signif-icance, since the accumulated pore pressure is
accordingly minimized.
The concrete mixtures produced in this investiga-
tion vary with reference to their capillary water
absorption since they belong to different strength
classes. Concrete mixtures of the lower strength
classC25/30appear to have greater water capillary
absorption values compared to C30/37. Lower w/c
ratios as well as the higher cement content that is
used for the C30/37 production forms a tighter
structure which eventually results in lower perme-ability. Water capillary absorption values are in all
cases lower in SCC compared to NC of the same
strength class, probably due to a more efficient
packing that is achieved by the use of filler materials.
Among SCC mixtures the one produced with glass
filler seems to have greater water absorption values,
since the w/c ratio in that occasion is somewhat
higher. SCC produced with ladle furnace slag appears
to be less permeable than the remaining mixturesFig. 4 Spalled specimens after exposure at high temperature
(600C)
1398 Materials and Structures (2009) 42:13931405
8/8/2019 Mechanical characteristics of self-compacting concretes with different filler materials, exposed to elevated temperat
7/13
mainly due to its denser microstructure, while
limestone filler performed similarly. In any occasion
there is explosive spalling when the peak temperature
of 600C is maintained when all concretes of both
strength classes are tested, and this is valid irrespec-
tive of the strength class and the filler material used.
3.4 Pulse velocity
The significant changes which concrete specimens
undergo as far as their pore structure is concerned,
when heated at different peak temperature are
assessed by the pulse velocity test. The results are
plotted in Fig. 5 where the residual pulse velocity is
expressed as the ratio of the pulse velocity after
exposure to each peak temperature to the initial value
at ambient temperature. Using pulse velocity propa-
gation is possible to figure out through a non
destructive method, the extent of deterioration due
to elevated temperatures or even the presence of
cracks and voids [22]. A decrease in velocity
indicates the initiation of cracks in the concrete mass
and increase in the porosity [2]. According to Piasta
the development of micro-cracks in cement paste
increases significantly beyond 300C [20]. Lin et al.
further confirms that the majority of the cracks and
the extremely large cracks are formed between 300
and 500C [23].
The relative pulse velocity results in this research
coincide well with the above made statements, as
there is a slight reduction in the slope up to 300C
and then the angle of gradient is greater for the
temperature range of 300600C for all mixtures
Fig. 5 (C25/30-left and C30/37-right). For the con-
crete mixtures which belong to the lower strength
classC25/30the relative pulse velocity is almost
identical for all mixtures after exposure at 300C. At
the following temperature range (300600C) more
severe degradation occurs. After exposure to 600C
SCC produced with limestone filler have the best
performance and SCC with ladle furnace slag appear
to have suffered greater deterioration. That is mostly
the case for the C30/37 mixtures. After exposure at
600C all SCC specimens containing slag suffer
severe deterioration due to explosive spalling. With
the exception of all mixture produced with slag,
SCCs of the higher strength class perform slightly
higher residual values compared to ordinary concrete
after exposure to 600C.
3.5 Residual tensile strength
With a close look at the tensile splitting strength in
Fig. 6 it becomes evident that there is a sharp loss of
tensile strength compared to a rather smoother declin-
ing curve which corresponds to the compressive
Table 5 Capillary water absorption (g/cm2
)
Time/mix C25/30
SCC-LF
C30/37
SCC-LF
C25/30
SCC-SL
C30/37
SCC-SL
C25/30
SCC-GF
C30/37
SCC-GF
C25/30
NC
C30/37
NC
T /10 min 0.1578 0.1022 0.1428 0.1122 0.1569 0.1448 0.1444 0.0933
T /24 h 0.5133 0.4133 0.5019 0.4003 0.5180 0.4536 0.5444 0.4667
Fig. 5 Ratio of residual pulse velocity [V(T)] after peak temperature to the pulse velocity at room temperature [V(20C)] of C25/30
(left) and C30/37 (right) for SCC and NC
Materials and Structures (2009) 42:13931405 1399
8/8/2019 Mechanical characteristics of self-compacting concretes with different filler materials, exposed to elevated temperat
8/13
strength loss at different peak temperatures [2, 5].
Chan et al. [5] attribute this to the presence of many
micro or macro cracks that are produced in the
specimens due to thermal incompatibility.
The tensile splitting strength results for the C25/30
mixtures are plotted in Fig. 6 consisting of two
descending branches, exhibiting the strength loss at
20300 and 300600C temperature ranges respec-
tively. All mixtures of this strength class tend to a
similar decrease from 20 to 300C. SCC with
limestone filler appear to have the best perfor-
mance, from 300 to 600C. On the other hand SCC
with ladle furnace slag in its composition along
with SCC using glass filler follows similar decline
with NC and suffers the greater loss of tensilestrength at the range of 300600C. Splitting
tensile strength results are only plotted from 20
to 300C regarding to the higher strength class
concrete, since all the cylinders which belong to
that strength category spall explosively. SCC with
limestone filler manage in this case also to
maintain greater percentage of its original tensile
strength compared to the rest SCC and NC of the
same strength class.
3.6 Stressstrain curves
The stressstrain curve, representing the deformation
and mechanical characteristics, is an important
material characteristic of concrete [24]. It is also
important to extract results for the concrete attributes
from the stressstrain curves at elevated tempera-
tures, although many coexisting effects determine the
shape of the curve. In this study, an attempt is made
to observe the difference of the stressstrain curves
among SCC mixtures prepared with several filler
materials compared to normal concrete.
As it is observed in Fig. 7, there is no noticeable
difference between SCC and NC in the shape of the
curves, either between SCC with different filler
materials. Generally, the ascending phase of all the
curves as the temperature increases becomes
smoother especially at high temperatures (600C)
while the peak strain increases and the peak strength
decreases. It must be mentioned that there could be a
pronounced concave-up curve at the beginning of
loading due to the pre-existing cracks caused by
heating and cooling [25]. Comparing the curves
between the two strength classes it can be mentioned
that the higher the strength class, the more rapid theascending phase and the more linear the descending
one, due to the stiffness softening of the specimens at
all temperatures respectively. Also, the percent of
peak strain increasing is for the C25/30 mixtures 12
15% and for the C30/37 mixtures 710% at 300 and
at 600C the respective percent is 6065% and 50
55%. Finally, SL mixtures give the impression of
having more linear phases and rough curve distribu-
tion before and after the peak strength compared to
the GF mixtures that have the smoother distribution.
3.7 Model suggestion
The model equations which are proposed in order to
evaluate the mechanical characteristics for both
unheated and heated concrete are shown in the
following equations [25]. In this paper the residual
mechanic characteristics (only the peak values) are
evaluated and compared with the values that have
been produced by the experimental program at
Fig. 6 Tensile strength of C25/30 (left) and C30/37 (right) for SCC and NC after heating at different temperatures
1400 Materials and Structures (2009) 42:13931405
8/8/2019 Mechanical characteristics of self-compacting concretes with different filler materials, exposed to elevated temperat
9/13
different temperatures (T). In particular the residual
peak strength value (fcr) is evaluated by Eq. 1 using
original compressive strength (fc). Equation 2
expresses the peak strain value (eor) using the
unheated peak strain. The residual tensile strength
(ftr) is calculated by Eq. 3 using the experimentally
measured tensile strength (ft), whilst Eq. 4 evaluates
the residual modulus of elasticity (Ecr).
fcr=fc 1:008T
450ln T5800
!0:0; 20C\T800C1
Fig. 7 Stressstrain curves of C25/30 (left) and C30/37 (right) of NC (a, b) and SCC (ch) after heating at different temperatures
Materials and Structures (2009) 42:13931405 1401
8/8/2019 Mechanical characteristics of self-compacting concretes with different filler materials, exposed to elevated temperat
10/13
ftr=ft 1:05 0:025 T;
0:80;1:02 0:0011 T!0;
8