Construction and Building Materials 124 (2016) 558–565

Contents lists available at ScienceDirect

Construction and Building Materials

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

Durability evaluation of cement exposed to combined action of chloride

and sulphate ions at elevated temperature: The role of limestone filler
N. Mavropoulou a, N. Katsiotis a, J. Giannakopoulos b, K. Koutsodontis b, D. Papageorgiou b,
E. Chaniotakis b, M. Katsioti a, P.E. Tsakiridis c,⇑
a
School of Chemical Engineering, National Technical University of Athens, 9 Iroon Polytechniou Street, 15780 Zografou, Athens, Greece
b
Titan Cement Company S.A., P.O. Box 18-19200, Elefsina, Greece
c
National Technical University of Athens, School of Mining and Metallurgical Engineering, Laboratory of Metallurgy, 9 Heroon Polytechniou St, 15780 Athens, Greece

h i g h l i g h t s

 Limestone filler role on cement durability exposed to SO4/Cl ions, at elevated temperature.
 Limestone presence led to less expansion and higher compressive strengths.
 Non-expansive Friedel’s salt was formed in chloride baths reducing DEF.
 Limestone filler reduced DEF and led to refinement of paste pore structure.

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

Article history: In the present investigation, the influence of sulphate and chloride ions combined action on cement
Received 12 April 2016 durability and the role of limestone filler were investigated. Samples of a reference white CEM I42.5
Received in revised form 18 July 2016 (CRef) and the corresponding of a white CEM II42.5 (CL15) with a 15 wt% limestone replacement were
Accepted 27 July 2016
exposed for a period of 90 days into three different curing baths (saturated Ca(OH)2 solution, 5 wt%
Available online 30 July 2016
Na2SO4 solution and a composite solution of 5 wt% Na2SO4 and 5 wt% NaCl), at 50 °C. Their compressive
strengths were determined after 28, 56 and 90 days of curing, whereas their expansion was estimated on
Keywords:
a weekly basis by the length change measurement. The deterioration products characterization was
Durability
Sulphate attack
carried out by X-ray Diffraction (XRD), Scanning Electron Microscopy (SEM), Fourier Transform
Chloride Infrared spectrometry (FT-IR), Thermal Analysis (TG-DSC) and Mercury Intrusion Porosimetry (MIP).
Blended cements According to the results the addition of limestone filler led to the durability increment, as the correspond-
Limestone filler ing mixtures presented lower expansion and higher compressive strengths, while in all cases the Cl ions
presence mitigated the external sulphate attack, a fact that was mainly attributed to the Friedel’s salt
formation.
Ó 2016 Elsevier Ltd. All rights reserved.

1. Introduction
The deterioration caused by the sulphate ion ingress from the
environment, called ‘‘external sulphate attack”, plays a major role
in the durability of concrete structures. The key issue of sulphate
attack’s effect on cement paste continues to cause significant
concern globally, because the delayed ettringite formation (DEF)
due to sulphate interaction could lead to the internal cracking of
concrete and to the corresponding loss of strength [1–3].
There is extensive scientific literature on the matter of DEF,
suffice herein to point out that ettringite appears in two forms,
⇑ Corresponding author.
E-mail address: ptsakiri@central.ntua.gr (P.E. Tsakiridis).
which are different not only in size but also in type. Lath-like
crystals’, ranging from 10 lm to 100 lm, is the main characteristic
of Type I, whose formation takes place at relative low concentra-
tions of hydroxyl ion (relatively lower pH values) and its presence
during hydration implies higher compressive strengths and lower
expansion values. For that reason the ettringite of the above type
has been classified as ‘‘non-expansive”. On the other hand, rod-
like crystals of 1–2 lm in length and 0.1–0.2 lm in thickness
(ettringite of Type II), are formed in higher pH values of the pore
solution, which could lead in large quantities, through water
adsorption, in expansion phenomena. It should be also noticed that
the delayed ettringite formation can be accelerated in case of
higher temperature treatment, whereas the formation of normal
ettringite, which is the result of calcium aluminate hydration in

http://dx.doi.org/10.1016/j.conbuildmat.2016.07.134
0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
 N. Mavropoulou et al. / Construction and Building Materials 124 (2016) 558–565 559

the presence of gypsum during setting regulation, is destroyed Table 1

with the temperature increment [4–6]. Chemical analysis and physical characteristics of cement used.

The mechanism, through which ettringite is formed, has been Oxides Chemical analysis (wt%)
well-studied and it has been concluded that the following three CRef CL15
conditions must simultaneously apply: a. excess of sulphate ions;
SiO2 21.87 19.06
b. elevated temperature and c. abundant supply of water [7–11]. Al2O3 5.06 4.65
In the cement production limestone filler has been extensively Fe2O3 0.2 0.23
used, as a substitute for clinker, as a less energy-demanding mate- CaO 65.28 68.03
rial; the percentage of limestone addition, according to European MgO 0.91 0.93
K2O 0.32 0.28
Standards (e.g. EN 197), ranges up to 35 wt%. According to various
Na2O 0.12 0.1
researchers, a certain amount of 5 wt% limestone addition reacts in SO3 3.03 2.98
order to form calcium-alumina carbonate phases as well as mono- Cl 0.018 0.017
carbonate, such as the more stable AFm phase (opposed to mono- LOI 1.02 3.98
sulphate) [12–14]. The latter results in a greater volume of Physical Characteristics
hydrates and as a result, more aluminates are initially bound in Specific Surface (cm2/g) 3400 4260
Specific gravity (g/cm3) 3.14 3.13
the form of ettringite, thus increasing the sulphate resistance.
Furthermore, it has been shown that the higher volume of hydrates
products physically lowers porosity [15].
Apart from research being conducted on sulphate-only environ- three different environment conditions: a) saturated Ca(OH)2
ments, there has been also an interest regarding the exposure on solution; b) 5 wt% Na2SO4 solution and c) combined solution of
both sulphate and chloride ions, representing the environment 5 wt% Na2SO4 & 5 wt% NaCl. The mortars specimens in the mould
met to marinas or to saline soils. It has been shown that there is were stored in a moist atmosphere (95% RH, 20 ± 1 °C) for 24 h
an increase in sulphate attack resistance of certain cement types and then the demoulded mortars were stored in tanks with the
when the solution also contains chloride ions, a fact that has been different solutions, until strength testing. Curing at elevated tem-
attributed to the aluminates hydrated phases transformation to the perature was chosen in order to accelerate the hydration phenomena
corresponding chloroaluminates phases, having as a consequence and the reactions between hydrated phases with the correspond-
the reduction of ettringite formation [16–18]. Furthermore, it has ing corrosive anions [2,3]. Inspection of the specimens was carried
been established that the ettringite solubility in the presence of out regularly, while the determination of the expansion took place
chlorides ions is three times higher than the corresponding one by measuring their length. After of 90 days of curing the mortars
in an aqueous solution [19,20]. compressive strengths were determined according to EN196-1
On the other hand, the rate of chloride ions diffusion has been standard.
found to be higher than the corresponding of the sulphate ones, Additionally, products from the surface of all mortars were col-
thus allowing the former to penetrate faster to the interior of lected and were determined by X-ray diffraction was used, using a
cement [21,22]. More specifically it has been reported that the dif- D8-Focus diffractometer (Bruker) at 40 kV and 40 mA with nickel-
fusion rate of chlorides in concrete is generally 10–100 times faster filtered CuKa radiation (k = 1.5406 Å). Semi-quantitative phases
than that of sulphate ions, while Rio and Turriziani have shown analysis was carried out by TOPAS software (Bruker-AXS), based
that the Cl diffusion rate in pure and blended cements is twice on Rietveld algorithm. The characterization of the above products
greater than that of the sulphate ions [23]. This higher diffusion was also carried out by a Spectrum GX (Perkin Elmer) Fourier
rate of the chloride ions can lead to their faster reaction with transform spectrophotometer, in the range of 4000–400 cm 1,
hydrated calcium aluminates phases and to the production of the using the KBr pellet technique. The pellets were prepared by press-
non-expansive Friedel’s salt (Ca3Al2O6CaCl2 10H2O) [24], whose ing a mixture of the sample and dried KBr (ratio, about 1:200) at
stability depends on pH, since it has been shown that as the pore 8 tons/cm2. A Setaram-Labsys thermal analyzer was used for TG/
solution’s alkalinity decreases, the solubility of the salt increases. DSC measurements, at a heating rate of 5 °C/min under static con-
On the other hand, the hydrated cement carbonation could lead dition in the range of 25–1000 °C. Scanning electron microscopy
to the Friedel’s salt dissolution [25–28]. was used for the hydration products morphology examination
The present work is part of a research program [29–31], which using a Philips XL-30 Environmental Scanning Electron Micro-
aims at the examination of limestone filler addition on the durability scope, equipped with an EDAX DX-4 Analyzer for spot chemical
evaluation of various cement mixtures exposed to the combined analyses. Finally, A Micromeritics AutoPore III 9410/9420 mercury
action of sulphate and chloride ions at elevated temperature, porosimeter was used for porosimetry measurements, in the range
focusing on the deterioration products determination and their of 1–4000 bar mercury intrusion in order to determine pore
effects on the produced mortars. volume and pore size distribution.

3. Results & discussion

2. Materials & methods
CEM I reference cement, prepared by white cement clinker and
gypsum co-grinding, presented a Blaine value of 3400 cm2/g,
whereas the corresponding value for CEM II/A-L with 15 wt% lime-
stone filler was determined at 4260 cm2/g (Table 1). White cement
clinker was chosen to be used for the above mixtures production
due to its high C3A content (about 13 wt%), the phase with which
sulphate and chloride ions react in order to produce ettringite
and Friedel’s salt.
The produced cement mortars, prepared by mechanical mixing
and compaction in a mould using a jolting apparatus according to
EN196-1 standard, were exposed for a period of 90 days at 50 °C to
3.1. Physical characteristics
After 90 days of curing at 50 °C, all mortar samples’ surfaces
were examined using a Heine Delta 20 LED Illuminated Loupe
(10x) in order to assess their external condition. In all specimens
the first deterioration signs were observed at their edges and cor-
ners. It was found that CRef mortars displayed more cracks and
expansion in comparison to those with limestone filler. Also, it
was evident that mortar samples exposed to the Na2SO4 solution
displayed significantly greater deterioration in comparison to the
mortar samples exposed to the combined solution of sulphate
and chloride ions (Fig. 1).
560 N. Mavropoulou et al. / Construction and Building Materials 124 (2016) 558–565

Fig. 1. (a) CRef in 5 wt% Na2SO4 solution; (b) CL15 in 5 wt% Na2SO4 solution; c) CRef in 5 wt% Na2SO4 & 5 wt% NaCl solution; (d) CL15 in 5 wt% Na2SO4 & 5 wt% NaCl solution.

The length of both cement specimens was determined during
their curing via the use of a micrometer. Measurements were taken
throughout the 90 day period of their expansion, which was
expressed as the change in length (mm) per meter of sample.
The total measured expansion of all samples after 90 days of curing
is presented in Fig. 2. According to the results the specimens cured
in 5 wt% Na2SO4 solution displayed a greater expansion rate in
comparison with the mortar samples exposed to the combined
5 wt%Na2SO4 & 5 wt%NaCl solution. The above observation was
attributed to the presence of chloride ions and to the subsequent
chemical transformation of the calcium aluminates hydrated
phases to the non-expansive Friedel’s salt, instead of secondary
ettringite formation. Furthermore, it is clearly evident for all the
solutions tested that the samples with 15 wt% limestone filler dis-
played significantly less expansion in comparison to the reference
ones. This suggests that a 15 wt% replacement of clinker with lime-
stone can lead to a level of protection against delayed ettringite
formation, which is the main reaction product of calcium alumi-
nates hydrated phases in the presence of sulphate ions. The pres-
ence of limestone filler, which has been acted as additional
CRef _Ca(OH)2
nucleation sites, led to the partial consumption of the produced
calcium aluminates hydrated phases, due to their transformation/-
carbonation to the corresponding carboaluminates non expansive
derivatives.
3.2. Mechanical characteristics
The above observations were also confirmed and by the com-
pressive strengths results, after 90 days of curing (Fig. 3). The mix-
tures with 15 wt% limestone filler, which had been classified in the
same strength category (42.5 MPa) with the reference ones, clearly
presented greater compressive strengths, in all tested solutions. As
it was mentioned above, the carbonate filler accelerates cement
hydration, because of the number and the nature of contacts
achieved in cement-filler interface, on which nuclei of hydration
products are generated. On the other hand, the deterioration phe-
nomena, due to the secondary ettringite formation, have been
occluded, because of the calcium aluminates hydrated phases
CRef_Ca(OH)2
70
CL15 _Ca(OH)2
CRef_Na2 SO4

16000 60 CL15 _Na2 SO4

CL15 _Ca(OH)2
CRef_Na2 SO4 / NaCl
CRef _Na2 SO4
14000 CL15 _Na2 SO4 / NaCl
CL15 _Na2 SO4 50
Compressive Strength (MPa)

12000 CRef _Na2 SO4 / NaCl

Length Change (μm/m)

CL15 _Na2 SO4 / NaCl 40

10000

8000 30

6000
20

4000

10
2000

0 0

Fig. 2. Expansion after 90 days in different curing conditions at 50 °C. Fig. 3. Compressive strengths after 90 days in different curing conditions at 50 °C.
 N. Mavropoulou et al. / Construction and Building Materials 124 (2016) 558–565 561

consumption and the consequent carboaluminates formation. As a
result, the limestone filler addition resulted to an internal
microstructure refinement, lowering the porosity of the cement
matrix and thus leading not only to the compressive strength
increment, but also to a higher sulphate attack protection. It is also
noteworthy that the curing in chloride baths seemed to delay the
sulphates ions external attack and the formation of secondary
ettringite, as the compressive strengths were increased in all cases.
This observation has been also attributed to the reaction of calcium
aluminates hydrated phases in the presence of Cl ions (due to
their faster diffusion) and to their subsequent entrapment with
the form of non-expansive Friedel salt, thus avoiding the more
expansive secondary ettringite formation.
3.3. Mineralogical characterization
The X-ray diffraction patterns of the samples exposed for
90 days to 5 wt% Na2SO4 and to the composite 5 wt% Na2SO4/5
wt% NaCl solutions at 50 °C are presented in Fig. 4. The semiquan-
titative analysis, based on X-ray diffraction patterns, is presented
in Table 2. The principal phases determination was carried out
by TOPAS software (Bruker-AXS), based on Rietveld algorithm, tak-
ing into account the assumption that the sum of the identified
crystalline phases has to add up to 100%. Regarding the presence
of ettringite, in each sample and set of conditions it is obvious that
the samples that had been exposed to the Na2SO4 solution pre-
sented higher final content, in comparison with the samples
exposed to the composite solution of 5 wt% Na2SO4 & 5 wt% NaCl.
This was mainly attributed to the additional secondary ettringite
formation in the presence of sulphate anions, whose generation
on the other hand was disinclined because of the chloride binding
in the matrix of calcium aluminates hydration products and the
consequent formation of Friedel’s salt. The samples with limestone
filler presented lower ettringite content, as it had been transformed
during hydration to the more stable carboaluminates outgrowth.
This was also confirmed and by the portlandite amount detected
in case of 15 wt% substitution, which was found to be relatively
lower in comparison with the corresponding of reference
cement. The carbonate particles of the filler were acted as
nucleation sites, where further reaction with Portlandite was
carried out to produce more carboaluminates and carbonated
C-S-H, thus leading to the corresponding relative decrease of CaCO3
peaks intensity.
The FT-IR measurements, carried out on the deteriorated
mortar samples, were in accordance with the X-ray diffraction
findings. According to the obtained spectra (Fig. 5), it is obvious
from Ca(OH)2 characteristic absorption band (3644 cm 1) that in
case of the combined action of both chloride and sulphate ions
the consumption of Portlandite is higher than the corresponding
of Na2SO4 bath, mainly due the production of Friedel’s salt. This
was also confirmed by the detected peaks related with
ettringite at 1120 cm 1 (associated with SO4 2) and at
850 cm 1 (associated with AlO6), which were lower in case of
Na2SO4/NaCl curing bath.
On the other hand, in case of limestone filler addition the detec-
tion of carboaluminates was attested by the absorptions bands at
714, 875, 1424 cm 1, whereas the corresponding at 714, 875,
1473, 1795 cm 1 and 2513 cm 1 (CO23 vibration) are characteristic
for CaCO3. It should be noticed that in case of chloride ions there
was a slight decrease in the peaks that correspond to carboalumi-
nates, as they were not stable in Cl solution, contrary to ettringite,
which was not dissolved. Regarding the Friedel’s salt, its formation
was attested by the absorption band at 532 cm 1 (vibration of M–O
or M–O–H), as the characteristics bands at 3640 cm 1, 3480 cm 1
(mOH), 1621 cm 1 (dH2O), 785 cm 1 and 620 cm 1, were over-
lapped and by other hydrated phases.

1200
1. Ca(OH) 2
3
1 2. Ca 6Al2(SO4)3(OH)2 26H2O
3. CaCO 3
1
4. Ca 1.5 SiO3.5 xH2O
5. MgO
1000 7. Ca 4Al2O6(CO3)0.5(OH) 11.5 H2O
8. Ca 3Al2O6CaCl2 10H2O

800
1
4
2,3

2
Intensity (cps)

1 4
3 1
3 3 3
2 2 3 2
600 2 2 2
7 2 5 4
CL15 (SO 4 /Cl) 8
1

2
7
400 CL15 (SO 4 ) 2
2

8 2
CRef (SO 4 /Cl) 2
200

CRef (SO 4 )

0
5 10 15 20 25 30 35 40 45 50 55
2θ

Fig. 4. XRD of samples exposed to different curing conditions after 90 days at 50 °C.
562 N. Mavropoulou et al. / Construction and Building Materials 124 (2016) 558–565

Table 2
Phases composition by Rietveld analysis after 90 days of curing at 50 °C.

Sample Curing conditions Mineralogical phases (wt%)

Ettringite Ca(OH)2 CaCO3 Ca1.5SiO3.5xH2O Carboluminates Friedel’s salt MgO
CRef 5 wt% Na2SO4 18.7 34.9 16.8 24.7 2.5 – 2.4
CL15 5 wt% Na2SO4 8.2 25.3 39.6 15.2 10.2 – 1.5
CRef 5 wt% Na2SO4 & 5 wt% NaCl 12.4 33.5 15.3 25.4 2.3 8.6 2.5
CL15 5 wt% Na2SO4 & 5 wt% NaCl 7.2 24.4 38.7 13.4 8.5 6.4 1.4

300
CL15_Na 2SO4/NaCl
250

CRef _Na 2SO4/NaCl 200

Transmittance
150
CL15_Na 2SO4

100

CRef_Na 2SO4
50

0
4000 3600 3200 2800 2400 2000 1600 1200 800 400
Wavenumbers (cm-1)

Fig. 5. FT-IR of samples exposed to different curing conditions after 90 days at 50 °C.

100 30
CRef_Na2 SO4
25
C L15 _Na2 SO4

95 CRef_Na2 SO4 / NaCl 20

C L15_Na2 SO4 / NaCl
15
Mass Loss (%)

DSC (μV)
90 10

85 0

-5

80 -10

-15

75 -20
0 200 400 600 800 1000
o
Temperature ( C)

Fig. 6. TG-DSC of samples exposed to different curing conditions after 90 days at 50 °C.

The results of thermal analysis (TG-DSC) of the samples exposed
for 90 days to 5 wt% Na2SO4 and to the composite 5 wt% Na2SO4/5 wt
% NaCl solutions at 50 °C are presented in Fig. 6. According to the
TG curves three major mass losses in the range of 120, 460 and
740 °C were detected, corresponding to the dehydration of C-S-H/
ettringite, to the Portlandite dehydroxylation and to the CaCO3
decomposition, respectively. Between 30 °C and 105 °C the evap-
orable water and part of the bound water have escaped. In the
range of 100–150 °C the broad peak was evidence of the existence
of ettringite and colloidal C-S-H gel. However, the addition of lime-
stone filler led to the decrement of both primary and secondary
ettringite mainly because of the carbonation of the calcium
aluminates hydrated phases. This was also confirmed and by the
shift to lower temperatures of AFt peaks, because of the hydrated
cement matrix enrichment with carboaluminates. In case of
composite curing bath (5 wt% Na2SO4 & 5 wt% NaCl), a small
dehydration peak at about 350 °C confirmed the presence of
Friedel’s salt, whose formation relative inhibited the growth of
more expansive secondary etrringite. However, the presence
of limestone filler led to a relative decrease of the produced
chloroaluminates, due to binding of calcium aluminates phases
with the form of their carbonate derivatives. Between 440 °C
and 490 °C the second main mass loss was attributed the
dehydroxylation of Ca(OH)2, whose quantity was relatively lower
 N. Mavropoulou et al. / Construction and Building Materials 124 (2016) 558–565 563

in case of CL15 samples, as it was consumed during the transforma-
tion reaction of hydrated calcium aluminates and calcium silicates
to their carbonated derivatives. Finally, the decomposition of
calcium carbonate both from filler and from Ca(OH)2 carbonation
was detected by the presence of two peaks at 680 °C and 740 °C.
3.4. Microstructural characterization
Figs. 7 and 8 show scanning electron micrographs of the CRef
and CL15 hydration products, after 90 days of curing at different
solutions. In case of 5 wt% Na2SO4 curing bath, rod-like secondary

a
a

5wt% Na2 SO4

a b
b
a

b b
b a b

5wt% Na2SO4 & 5wt% NaCl

Fig. 7. Electron micrographs of CRef after 90 days in different curing conditions at 50 °C; a: rod-like Ettringite crystals; b: Friedel’s salt plates.

a
a a
a

5wt% Na2 SO4

b a

a b

5wt% Na2SO4 & 5wt% NaCl

Fig. 8. Electron micrographs of CL15 after 90 days in different curing conditions at 50 °C: lath like carboaluminates crystals; b: Friedel’s salt plates.
564 N. Mavropoulou et al. / Construction and Building Materials 124 (2016) 558–565

ettringite crystals were detected, which being formed at the high
sulphate ions concentration, have caused expansion phenomena
through water adsorption. The delayed ettringite evolution, carried
out mainly by dissolution and recrystallization in voids, was also
promoted and by the high-temperature curing. On the other hand,
the presence of chloride ions, in conjunction with the sulphates
ones, seemed to mitigate the formation of secondary ettringite.
In case of 5 wt% Na2SO4 & 5 wt% NaCl mixed curing bath the pres-
ence of chlorite led to the formation of hexagonal plates of Friedel’s
salt, a less expansive phase, which seemed to have been also pro-
moted, instead of ettringite, and by the elevated temperature.
In case of CL15, it is obvious that the limestone fine fraction has
acted as a micro-filler, leading to a pore structure refinement and
thus improving mechanical properties. Due to filler partial dissolu-
tion, additional nucleation sites were produced, where reaction
with calcium aluminates hydrated phases led to the formation of
needle-like or lath like carboaluminates, which range from
10 lm to 20 lm, a phase that is more stable than secondary ettrin-
gite. However, the carbonization of the hydrated cement matrix led
to the consequent decrease of Portlandite content and to a relative
decrement of the cement paste pH [13,15]. That was the reason of
the observed reduction of Friedel’s salt formation, in cases of Na2-
SO4 & 5 wt% NaCl curing bath, as it is known that although Friedel’s
salt in higher pH is more stable instead of secondary ettringite,
when the alkalinity was decreased, because of limestone filler, its
solubility was increased and thus more carboaluminates were
formed [32,33].
The results of MIP are summarized in Table 3. Fig. 9 presents
the cumulative and the differential pore size distributions of the
samples exposed for 90 days to 5 wt% Na2SO4 and to the composite
5 wt% Na2SO4/5 wt% NaCl solutions at 50 °C. In all cases there was a
decrease in the porosity with the advancement of hydration period,
due to the gradual filling of larger pores with hydration products.
Regarding differential curves, a strong peak was observed at
around 20 nm, while a second peak appeared at smaller pore sizes
(2 nm), thus suggesting a bimodal distribution. The larger peak
corresponded to the pressure required to break through the struc-
ture formed by the hydration products, while the first peak was
related to the lowest size of pores, connecting a continuous matrix
[33]. Moreover, it should be mentioned the presence of a third peak
in the range of 20 lm, which should be attributed to the presence
of larger voids and cracks, due to the sulphate interaction and to
the consequent delayed ettringite formation. After 90 days of cur-
ing, the addition of limestone filler led to the refinement of the
paste pore structure. As it was mentioned above the carbonate par-
ticles, especially the finer fraction, promoted the cement hydration,
mainly because of the number and the nature of contacts achieved
in cement-filler interface, on which nuclei of hydration products
were generated, gradually filling the larger pores and converted
them into smaller ones. On the other hand, the carbonates parti-
cles, binding the calcium aluminates hydrated phases with the
form of the corresponding carbonate derivatives, led to a consider-
able reduction in the volume of larger voids/cracks, which had
been ascribed to the secondary ettringite formation expansion
phenomena, and therefore enhanced the final compressive
strengths. As a results, the CL15 mixtures presented smaller total
pore area and total porosity, thus increasing their durability
against external sulphate attack, whereas the reference cement

Table 3
MIP after 90 days of curing at 50 °C.

Sample Curing Intrusion results

conditions
Total pore Bulk density Total intrusion Average pore Porosity
area (m2/g) (g/mL) volume (mL/g) diameter (nm) (%)
CRef 5 wt% Na2SO4 25.88 2.04 0.081 16.8 16.59
CL15 5 wt% Na2SO4 17.59 2.06 0.074 13.8 15.25
CRef 5 wt% Na2SO4 15.85 2.09 0.068 16.4 14.24
& 5 wt% NaCl
CL15 5 wt% Na2SO4 14.32 2.12 0.062 12.5 13.24
& 5 wt% NaCl

80 50
CRef_Na2 SO4

CL15 _Na2 SO4

CRef_Na2 SO4 / NaCl 40

Derivative ((mm3/g)/log(r, nm))

60 CL15 _Na2 SO4 / NaCl

Cum. pore vol. ( mm3/g)

30

40

20

20
10

0 0
1 10 100 1000 10000
Pore Radius (nm)

Fig. 9. Cumulative and differential pore size distribution curves of samples exposed to different curing conditions after 90 days at 50 °C.
 N. Mavropoulou et al. / Construction and Building Materials 124 (2016) 558–565
was characterized by a higher porosity and larger pores. In cases of
mixed curing bath, the presence of chlorides and the subsequent
consumption of calcium aluminates hydrated phases for the Frie-
del’s salt formation, led to a further decrease of total porosity, shift-
ing the larger peak to smaller pore radius, although the final larger
voids content was higher in relation with the corresponding of CL15
mixtures.
4. Conclusions
The present research study was conducted in order to investi-
gate the durability and the role of limestone filler in cement mix-
tures during hydration, under the exposure of sulphate and
chloride ions combined action, at elevated temperature. All mortar
samples immersed into the 5 wt%Na2SO4 and to the corresponding
5 wt%Na2SO4 & 5 wt% NaCl curing baths displayed strength loss
and expansion, which was mainly attributed to delayed ettringite
formation. However, after 90 days of curing, the addition of lime-
stone filler resulted in higher compressive strengths and lower
total porosity. The carbonate particles promoted the cement
hydration, due to the additional nucleation sites generation, where
more hydration products were formed, gradually filling the larger
pores. Furthermore, the filler addition led to the partial consump-
tion of the produced calcium aluminates hydrated phases due to
their transformation to the corresponding carboaluminates non
expansive derivatives, leading to lower expansion values. On the
other hand the presence of chloride ions prohibited the external
sulphate attack, as their diffusion rate is higher, thus binding cal-
cium aluminates hydrated phases, creating the more stable form
of the Friedel’s salt, instead of expansive secondary ettringite.
However, the addition of limestone filler led to inhibition of the
Friedel’s salt formation, leading to the formation of more
carboaluminates.
Acknowledgements
The Authors would like to acknowledge Dr. C.G. Koutsodontis
for the Mercury Intrusion Porosimetry analyses.
References
[1] M. Collepardi, A state of the art review on delayed ettringite attack on concrete,
Cem. Concr. Compos. 25 (4–5 SPEC) (2003) 401–407.
[2] C.D. Lawrence, Mortar expansion due to delayed ettringite formation. Effects of
curing period and temperature, Cem. Concr. Res. 25 (4) (1995) 903–914.
[3] F.P. Glasser, The role of sulphate mineralogy and cure temperature in delayed
ettringite formation, Cem. Concr. Compos. 18 (3) (1996) 187–193.
[4] S. Diamond, Delayed ettringite formation – processes and problems, Cem.
Concr. Compos. 18 (3) (1996) 205–215.
[5] J. Stark, K. Bollmann, Delayed Ettringite Formation in Concrete, XVIII
Symposium on Nordic Concrete Research, Reykjavik/Island (1999).
[6] A. Neville, The confused world of sulfate attack on concrete, Cem. Concr. Res.
34 (8) (2004) 1275–1296.
[7] M. Santhanam, M.D. Cohen, J. Olek, Mechanism of sulfate attack: a fresh look:
part 1: summary of experimental results, Cem. Concr. Res. 32 (6) (2002) 915–
921.
[8] P.W. Brown, An evaluation of sulfate resistance of cements in a controlled
environment, Cem. Concr. Res. 11 (5–6) (1981) 719–727.
565
[9] P. Pipilikaki, M. Beazi-Katsioti, J.L. Gallias, Performance of limestone cement
mortars in a high sulphates environment, Constr. Build. Mater. 23 (2009)
1042–1049.
[10] H.-J. Kuzel, Initial hydration reactions and mechanisms of delayed ettringite
formation in Portland cements, Cem. Concr. Compos. 18 (3) (1996)
195–203.
[11] H.-J. Kuzel, H. Meyer, Mechanisms of ettringite and monosulfate formation in
cement and concrete in the presence of CO2, in: Proceedings of the 15th
International Conference on Cement Microscopy, Duncanville, Dallas, Texas,
International Cement Microscopy Association, 1993, pp. 191–203.
[12] H. Kuzel, H. Pöllmann, Hydration of C3A in the presence of Ca(OH)2,
CaSO4*2H2O and CaCO3, Cem. Concr. Res. 21 (5) (1991) 885–895.
[13] B. Lothenbach, G. LeSaout, E. Gallucci, K.L. Scrivener, Influence of limestone
on the hydration of Portland cements, Cem. Concr. Res. 38 (6) (2008)
848–860.
[14] T. Matschei, B. Lothenbach, F.P. Glasser, The role of calcium carbonate in
cement hydration, Cem. Concr. Res. 37 (4) (2007) 551–558.
[15] T. Schmidt, B. Lothenbach, M. Romer, J. Neuenschwander, K. Scrivener, Physical
and microstructural aspects of sulfate attack on ordinary and limestone
blended Portland cements, Cem. Concr. Res. 39 (12) (2009) 1111–1121.
[16] O.S.B. Al-Amoudi, M. Maslehuddin, Y.A.B. Abdul-Al, Role of chloride ions on
expansion and strength reduction in plain and blended cements in sulfate
environments, Constr. Build. Mater. 9 (1) (1995) 25–33.
[17] H.A.F. Dehwah, M. Maslehuddin, S.A. Austin, Long-term effect of sulfate ions
and associated cation type on chloride-induced reinforcement corrosion in
Portland cement concretes, Cem. Concr. Compos. 24 (1) (2002) 17–25.
[18] V.V. Kind, Effect of Chlorides on the Speed of Sulfate Corrosion of Portland
Cement, Tsement, BRE No. 346, 1956, 22, Part 1, 3–6.
[19] J. Zuquan, S. Wei, Z. Yunsheng, J. Jiang, L. Jianzhong, Interaction between
sulfate and chloride solution of concretes with and without fly ash, Cem.
Concr. Res. 37 (8) (2007) 1223–1232.
[20] W.H. Harrison, Effect of chloride in mix ingredients on sulphate resistance of
concrete, Mag. Concr. Res. 42 (152) (1990) 113–126.
[21] R.E. Obserholster, Pore structure, permeability and diffusivity of hardened
cement paste and concrete in relation to durability: status and prospects, in:
Proceedings of 8th International Congress on the Chemistry of Cement, Rio de
Janeiro, Brazil, Sub-Theme 4.1, September 1986, pp. 323–335.
[22] R.F.M. Bakker, Permeability Of Blended Cement Concretes, ACI Special
Publication SP-79, American Concrete Institute, Detroit, 1983, 589–605,
December 1964 74–80.
[23] A. Rio, R. Turriziani, Il cement 80(1) 1983 37–48
[24] K. Ogawa, D.M. Roy, C4A3S hydration ettringite formation and its expansion
mechanism: III. Effects of CaO, NaOH and NaCl: conclusions, Cem. Concr. Res.
12 (1982) 247–256.
[25] C.L. Page, O. Venneland, Pore solution composition and chloride
binding capacity of silica-fume cement pastes, Mater. Struct. 16 (91) (1983)
19–25.
[26] S. Goni, A. Guerrero, Accelerated carbonation of Friedel’s salt in calcium
aluminate cement paste, Cem. Concr. Res. 33 (2003) 21–26.
[27] K. Kobayashi, M. Funato, R. Shiraki, Y. Uno, K. Kawai, Carbonation and
concentration of chloride in concrete containing chlorides, III, Kenkyin 41(12)
(1989) 930–932.
[28] A.K. Suryavanshi, R.N. Swamy, Stability of Friedel’s salt in carbonated concrete
structural elements, Cem. Concr. Res. 26 (5) (1996) 729–741.
[29] M. Beazi-Katsioti, N. Patsikas, P. Pipilikaki, N. Katsiotis, K. Mikedi, M.
Chaniotakis, Delayed ettringite formation (DEF) in mortars of white cement,
Constr. Build. Mater. 25 (2) (2011) 900–905.
[30] E. Adamopoulou, P. Pipilikaki, M.S. Katsiotis, M. Chaniotakis, M. Beazi-Katsioti,
How sulfates and increased temperature affect delayed ettringite
formation (DEF) in white cement mortars, Constr. Build. Mater. 25 (8)
(2011) 3583–3590.
[31] N. Patsikas, N. Katsiotis, P. Pipilikaki, D. Papageorgiou, E. Chaniotakis, M. Beazi-
Katsioti, Durability of mortars of white cement against sulfate attack in
elevated temperatures, Constr. Build. Mater. 36 (2012) 1082–1089.
[32] R. Luo, Y.B. Cai, C.Y. Wang, Effects of sulphates on chloride binding and
corrosion of GGBS concrete, paste and mortar, in: 26th Conference on Our
World in Concrete & Structures, 27–28 August 2001, Singapore.
[33] P. Pipilikaki, M. Beazi-Katsioti, The assessment of porosity and pore size
distribution of limestone Portland cement pastes, Constr. Build. Mater. 23 (5)
(2009) 1966–1970.