Construction and Building Materials 25 (2011) 3583–3590

Contents lists available at ScienceDirect

Construction and Building Materials

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

How sulfates and increased temperature affect delayed ettringite formation (DEF)
in white cement mortars
E. Adamopoulou a, P. Pipilikaki b, M.S. Katsiotis a, M. Chaniotakis b, M. Katsioti a,⇑
a
School of Chemical Engineering, Laboratory of Inorganic and Analytical chemistry, National Technical University of Athens, 9 Heroon Polytecniou St. Zografou, Athens 15780, Greece
b
TITAN Cement Co. S.A., Group R&D and Quality Department, 18-19200 Eleusina, Greece

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

Article history: In this study, white cement CEM I and white limestone cement CEM II-LL A and D with 15%, 25% and 35%
Received 20 August 2010 limestone substitution were studied. The way delayed ettringite is forming due to exposure to increased
Received in revised form 2 January 2011 temperature (50 °C) and external sulfate attack was examined in mortar samples which were immersed
Accepted 1 March 2011
for 90 days in three different solutions: (a) saturated solution Ca(OH)2 at 50 °C, (b) saturated solution of
Available online 25 March 2011
Ca(OH)2 at 20 °C and (c) 5% w/w Na2SO4 solution at 50 °C. During this period mortar samples were visually
observed regularly while their expansion was estimated on a weekly basis by measuring the change of
Keywords:
length with a micrometer. At the end of the 90-days period the compressive strength of the mortars was
DEF
Limestone
determined and the deterioration products were identified through means of X-ray diffraction (XRD), Scan-
Sulfates ning Electron Microscopy (SEM) and thermogravimetric analysis (DTG). The results of this study show that
White cement DEF occurred in two forms. Samples cured at increased temperature contained DEF type I, which caused
Mortar mediocre expansion and damage. Samples cured at increased temperature in the presence of sulfates pro-
duced DEF type II, which caused significant damage on the surface and exhibited high expansion.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction which at large quantities can cause expansion effects through

Delayed ettringite formation (DEF) is the ettringite-related
deterioration of concrete regardless of whether or not the concrete
member has been subjected to steam curing [1]. It is a phenome-
non in which ettringite is generated by accumulating in the con-
crete after hardening, eventually leading to its expansion and
thus, ultimately to the destruction of the concrete or mortar. DEF
is promoted by high-temperature steam curing, while normal
ettringite is formed as a consequence of the setting regulation dur-
ing the plastic stage of fresh concrete and is destroyed by steam
curing [2–5]. Furthermore, ettringite appears in two crystalliza-
tions, different in habit and size. The first one (type I) of long
lath-like crystals which range from 10 to 100 lm and several lm
thick, formed at low hydroxyl ion concentrations (i.e. with low
pH values in the pore solution). In the case of hydrated cements
containing significant amounts of type I ettringite crystals, these
lead to high strengths and non-expansion, thus the classification
of type I ettringite as ‘‘not expansive’’. The second type (type II)
of ettringite is rod-like crystals which range only from 1 to 2 lm
long and 0.1 to 0.2 lm thick, which are formed at high hydroxyl
ion concentrations. These are present during the hydration of
Portland cements and are classified as type II ettringite crystals,
water adsorption [6].
The way ettringite formation and damage mechanism is initi-
ated in the cement matrix has been the point of controversy for
the past years. After carrying extensive investigations, different
hypotheses were developed by many authors on the matter. It is
said that the formation mechanism and ettringite modification
play an important role. It has been found that heat treatment at
higher temperatures, frost or freeze–thaw, carbonation, moisture
effects and with it natural occurring changes in moisture and tem-
perature can cause ettringite formation [6].
Finally, most researchers agree that DEF is created when the fol-
lowing three conditions coincide and apply: excessive sulfates,
high-temperature steam curing and sufficient water supply [7–
12]. Delayed ettringite is formed into the pores or the cracks of
the cement paste due to high sulfate presence. The overall reac-
tions characterizing the delayed ettringite formation process are
described by following reactions.
CaðOHÞ2 þ Na2 SO4 þ H2 O ! CaSO4  2H2 O þ 2NaOH ð1Þ
C3 A þ 3ðCaSO4  2H2 OÞ þ 36H2 O
! 3C3 A  CaSO4  14H2 O ðmonosulfateÞ ð2Þ

2CaSO4 þ C3 A  CaSO4  14H2 O ! C3 A  3CaSO4

⇑ Corresponding author. Tel.: +30 2107723206; fax: +30 2106397771.
E-mail address: katsioti@central.ntua.gr (M. Katsioti).  32H2 O ðettringiteÞ ð3Þ

0950-0618/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2011.03.051
3584 E. Adamopoulou et al. / Construction and Building Materials 25 (2011) 3583–3590
C3 A  6H2 O þ 3ðCaSO4  2H2 OÞ þ 20H2 O
! C3 A  3CaSO4  32H2 O ðettringiteÞ ð4Þ
Pore distribution and shape, the transport of sulfates and external
conditions such as temperature and pressure play a major role on de-
layed ettringite formation. DEF may exist in different sizes (ranging
from amorphous colloidal particles to well defined crystals) depend-
ing on the conditions of formation. Usually DEF is found in the shape
of crystals roughly 1–2 lm long and 0.1–0.2 lm wide. It must be
noted that DEF can occur as either types I or II ettringite (as described
previously). According to existing literature, DEF created in high pH
environment (12.5–12.9) causes expansion and does not contribute
in strength development [13–16].
The External Sulfate Attack (ESA) induces damage which is
mainly determined by the chemical interaction of a sulfate-rich
soil or water, with the cement paste. ESA is a generic name for a
set of complex and overlapping chemical and physical processes
caused by reactions of numerous cement components with sulfates
originating from external or internal sources. Soils containing so-
dium, potassium, magnesium and calcium sulfate are the main
sources of sulfate ions in groundwater. Additionally, solid indus-
trial waste, fertilizers, SO2 found in the atmosphere or in liquid
industrial wastes are also regarded high sulfate sources which pro-
mote ESA. For ESA to occur the following three conditions must be
fulfilled at the same time: high permeability of concrete, sulfate-
rich environment and the presence of water [17–20]. As a result
of sulfates presence, different hydration products occur. These
products appear very rapidly, usually in few hours, and they are in-
volved in the expansion of concrete.
Aim of this work is to study the way cement pastes with high
limestone content act when exposed to high sulfate environments
and increased temperature treatment. The study is mainly focused
on determining deterioration products and effects while also verify-
ing the presence of DEF due to the aforementioned conditions. A sim-
ilar study has been performed by the authors and has been recently
published [21]. The previous work is focused more on the time evo-
lution of sulfate induced DEF expansion in white cement mortars.
Based on the conclusions drawn from that work, the authors wish
to give additional information regarding the expansion of mortar
due to sulfate attack, by including an additional mixture (35% lime-
stone substitution) and an additional set of conditions (curing in
20 °C). Accordingly, the effect of temperature on sulfate induced
expansion is the main focus in this paper. It must be noted that these
papers are complementary and readers are advised to read both
works in order to acquire a complete understanding of the sulfate in-
duced expansion mechanism. All in all, the authors believe that the
two papers offer a novel insight on sulfate induced DEF expansion,
thus assisting in the improvement process of the durability of white
cement mortars with high limestone content.
2. Sample preparation and methods
Initially mortars were prepared by mixing white cement CEM
I-52,5 with limestone at three different ratios (15%, 25% and
35%), thus creating white cement CEM II. Paste and mortar samples
were produced according to TN 196-1. A reference paste and mor-
tar sample was also prepared. Initial and final setting times were
determined for paste samples along with normal plasticity water,
according to EN 196-1.
Consequently, the cement mixtures were mixed with standard-
ised siliceous sand and water for the creation of mortar specimens.
The amount of sand used per specimen was 1350 g of standard sili-
ceous sand whose characteristics are described in EN 196-3. 450 g of
cement were added and all samples were formulated with a w/c ratio
equal to 0.6. The standard w/c ratio of 0.5 was not applied as the addi-
tion of limestone caused the samples to be less workable, while the
0.6 ratio proved to be the optimum regarding the rheology of the
fresh mortar samples. The mortars were molded in prismatic moulds
with dimensions 4  4  16 cm. Nine specimens were created from
each mixture in order to achieve statistical repeatability and were
preserved according to EN 196-1 instructions: for the first 2 days
all the samples were preserved in the moulds inside a wet chamber
(humidity >95%), then they were unmolded and preserved in a satu-
rated Ca(OH)2 solution for 12 days (total of 14 days).
After the initial preservation period the mortars were grouped
into three separate groups. Three specimens were taken from each
mixture for each group which corresponded to a different set of
condition, as explained below. First group was cured in a saturated
Ca(OH)2 solution at 20 ± 2 °C for 90 days. The second group was left
to hydrate in a saturated Ca(OH)2 solution at 50 ± 2 °C for 90 days
in order to evaluate the effect of increased temperature. Finally, the
third group was left to hydrate in a 5% w/w Na2SO4 solution at
50 ± 2 °C for 90 days to emulate an environment with high external
sulphate presence and increased temperature – both perfect condi-
tions for formation of type II ettringite (which causes expansion).
During this 90-days period the specimens were inspected and
their expansion was determined by measuring their length with
a micrometer on a weekly basis. It must be noted that soon after
the initial exposure to increased temperature and high sulfate
presence, deterioration was obvious on the surface of mortars. At
the end of the 90-days period, compressive strength of all mortars
was measured according to EN 196-1. Additionally, samples were
collected from the surface of all mortars and were studied using
X-ray diffraction analysis (XRD), TG thermogravimetric analysis
and Scanning Electron Microscopy (SEM).
X-ray powder diffraction (XRD) was carried out in a Siemens D-
5000 Diffractometer using Cu Ka radiation (k = 1.5405 Å, 40 kV,
30 mA). The diffraction spectra were collected in the range of 2–
60° 2h scale, with a step size of 0.03°/s. Powder samples (1 g) were
drilled from a layer of 2 mm thick (where damage from expansion
was evident) from the surface and were consequently analyzed.
Additional samples were also collected in order to determine the
amount of ettringite on the surface of each sample, using the Rietv-
ield quantitative method. In this case the spectra were collected in
the range 2–26° 2h scale, with a step size of 0.005°/sec. It must be
noted that ettringite appears in XRD at 2h angles 9.2° (main peak),
15.7° and 22.9°. Quantitative analysis was realized by measuring
the integral area at the aforementioned 2h angles.
Thermogravimetric analysis (TG/DTG) was carried out in a Met-
tler Toledo thermal analyzer, using a heating rate of 10 °C/min at
the range of 25–1000 °C, in air atmosphere. Specimens (10 lg)
were collected from a layer of 2 mm thick (where damage from
expansion was evident) from the surface and were consequently
analyzed. The DTG analysis was conducted in order to verify the
presence of calcite and determine the remaining portlandite in
the surface of each samples.
Finally, the microstructure of the surface as well as the forma-
tion and type of ettringite crystals of the treated mortars was stud-
ied under a FEI/Quanta 200 scanning electron microscope.
Specimens (1 ml) were collected from a layer of 2 mm thick
(where damage from expansion was evident) from the surface
and were consequently cast in resin and polished in order to be
examined under the microscope.
3. Results and discussion
3.1. Chemical analysis
Mortar specimens were chemically analyzed as shown in Table
1. The amount of C3A was determined according to Bogue’s equa-
tions for each cement mixture and is shown in Table 1.
 E. Adamopoulou et al. / Construction and Building Materials 25 (2011) 3583–3590 3585

From Table 1, the following observations can be made. Samples
with limestone substitution exhibit a decreased amount of SiO2
and Al2O3 due to decreased cement presence. Accordingly, C3A
was calculated to be in inverse proportion to limestone percentage
due to decreased Al2O3 presence. On the contrary, samples with
limestone substitution exhibited higher LOI due to limestone calci-
nation. Finally, it should be mentioned that sulfate presence seems
to be relatively high for all samples, which fulfils one of the funda-
mental conditions (mentioned in the Introduction) for delayed
ettringite formation.
3.2. Setting time and normal plasticity water
Initial and final setting time and normal plasticity water were
determined for paste samples. In each case, 500 g of cement mix-
ture were mixed with the appropriate mass of water as determined
from the Vicat testing method. Upon successful mixing with water,
the pastes were left to hydrate in a wet chamber and initial and fi-
nal setting times were determined using the Vicat testing method.
The results are shown in Table 2.
3.3. Expansion
The length of mortar specimens was determined during their
curing with use of a micrometer and their expansion was esti-
mated. The latter was expressed as the change in length in mi-
cro-meters per meter of sample. Results are shown in Fig. 1 from
four specific ages (14, 28, 56 and 90 days). It is observed that spec-
imens cured in 50 °C and under external sulfate attack present
higher expansion with a higher rate than specimens cured in dif-
ferent conditions.
Furthermore, mortars containing limestone expand with a high-
er rate (proportional to limestone substitution) under high temper-
ature and sulfate presence. This is in accordance with the optical
inspection results. Images of the treated samples are presented in
Fig. 2, where the deterioration is visible, especially for samples
treated with sulfates at 50 °C. On the other hand, the mortars im-
mersed in saturated Ca(OH)2 solution have low expansion. This
means that heat induced ettringite formation is limited by the
amount of C3A and SO3 that exist in cement. Regarding the time
evolution of expansion, it is shown in the authors’ previous work
[21] that the expansion phenomena continue to apply even after
the 90 days period, thus indicating that as long there is a constant
supply of sulfates expansion will occur.
It is remarkable that in the case where high temperature and
external sulfate attack are applied the expansion increases while
C3A is decreased in the cement. This can be seen clearly in Fig. 3,
where expansion is plotted versus C3A content. This is attributed
to chemical reaction between calcite and Na2SO4 which promotes
formation of gypsum, as explained below.

Table 1
Chemical analysis of cement mixtures (before hydration). 3.4. Compressive strength

Sample Component
At the end of the 90-days period, compressive strength of all
SiO2 Al2O3 Fe2O3 CaO MgO SO3 LOI C3A mortars was determined. The results are shown in Fig. 4. It must
0% LL 20.82 4.98 0.23 66.98 0.60 2.91 3.05 12.8 be noted that due to limestone substitution, the cements are not
15% LL 18.56 4.41 0.22 65.54 0.60 2.77 7.12 11.3 considered to be of the same CEM type. Accordingly, strength is re-
25% LL 15.60 3.84 0.23 65.86 0.49 2.54 10.96 9.5
duced where limestone is present, and this reduction is propor-
35% LL 13.36 3.25 0.21 66.07 0.50 2.37 13.73 8.3
tional to cement substitution. Another important observation is
that samples cured in a solution of 5% Na2SO4 have lower strength
than those cured in a saturated solution of Ca(OH)2 as external sul-
Table 2
Setting time and normal plasticity water. fate attack in combination with increased temperature resulted in
significant surfacial damage. However, one should observe that
Sample Setting time (min) Normal plasticity
samples with limestone exhibit a lower loss of strength than the
Initial–final Water (g) reference sample. In specific, the reference sample lost 14.3% of
0% LL 100–140 140 its strength due to external sulfate attack and increased tempera-
15% LL 110–140 135 ture, while samples with 15%, 25% and 35% limestone lost 13.7%,
25% LL 100–120 130
12.7% and 13.5% of their strengths respectively. The authors have
35% LL 100–130 132
observed a similar behaviour in other cases where limestone was

Fig. 1. Expansion of mortar specimens under three different curing conditions.

3586 E. Adamopoulou et al. / Construction and Building Materials 25 (2011) 3583–3590

Fig. 2. Images of mortar samples with 35% LL after 90 days curing at (A) 20 °C in Ca(OH)2, (B) 50 °C in Ca(OH)2, (C) 50 °C in 5% Na2SO4.

80

70
Compressive Strength (MPa)

60

50

40

30
50 oC, Na2SO4

50 oC, Na2SO4

50 oC, Na2SO4

50 oC, Na2SO4
20
20 oC

20 oC
50 oC

20 oC
50 oC

50 oC

20 oC
50 oC

10

0
0% LL 15% LL 25% LL 35% LL

Fig. 3. Expansion plotted versus respective C3A percentage for each mortar sample. Fig. 4. Compressive strength after 90 days of curing at different conditions.

used as a substitute for cement [21]. This should be taken into ac-
count when designing a cementitious product containing
limestone.
Finally, it must be noted that samples exhibiting higher
strengths also exhibited low expansion values when treated with
sulfates at increased temperature. The authors have observed a
similar behaviour before [10,21] which shows that compressive
strength must be regarded as a key factor for durability in high sul-
fates environments.
3.5. X-ray diffraction
The XRD patterns of all mortars are presented in Fig. 5. Rietvield
results are shown in Table 3. The most important observation can
be made regarding the presence of ettringite in each sample and
set of conditions. In specific, one observes that under sulfate and
increased temperature curing conditions all samples exhibited
high percentages of ettringite (9.2° 2h). More accurately, samples
with limestone contained higher percentages of ettringite in rele-
vant proportion to limestone substitution, as the Rietvield method
has shown. In this case DEF formed was type II which caused the
surfacial damage on the specimens. The fact that samples with
limestone contained less C3A (as shown previously) should not
confuse the readers, as the increase in ettringite was attributed
to the chemical reactions caused by limestone presence. In specific,
reaction (5) shows that calcite reacts with Na2SO4 and gypsum is
created:
 E. Adamopoulou et al. / Construction and Building Materials 25 (2011) 3583–3590 3587

A B
E= ETTRINGITE E= ETTRINGITE
Q = QUARTZ
E PE Q Q PCE P Q QQQC QP CP Q Q Q Q = QUARTZ
E E PE Q E Q C P Q QQC QPC P Q Q Q
C = CALCITE 500 C = CALCITE 500
P = PORTLANDITE P = PORTLANDITE
450 450

400 400

350 350
Intensity [a.u.]

Intensity [a.u.]
300 300

250 O
250
0% LL-50 C- 15% LL-50 C-
O

200 Na2SO4 200 Na2SO4

150 150
O
0% LL-50 C
O 15% LL-50 C
100 100

50 50
O
0% LL-20 C
O 15% LL-20 C
0 0
5 10 15 20 25 30 35 40 45 50 55 60 65 5 10 15 20 25 30 35 40 45 50 55 60 65
ο
ο 2θ [ ]
2θ [ ]
C D
E= ETTRINGITE
E P Q C QE CE PEQQQQ P QPQCP Q P Q Q
E= ETTRINGITE
Q = QUARTZ
E EE E P Q E E Q C P Q QQQC QC CP Q Q Q
Q = QUARTZ E
C = CALCITE
500 C = CALCITE 500
P = PORTLANDITE P = PORTLANDITE
450 450

400 400

350 350

300 300
Intensity [a.u.]
Intensity [a.u.]

250 250
O
25% LL-50 C- O
35% LL-50 C-
200 Na 2SO4
200 Na 2SO4
150
25% LL-50 C
O 150
O
100 35% LL-50 C
100
50
50
O
25% LL-20 C O
0 35% LL-20 C
0
5 10 15 20 25 30 35 40 45 50 55 60 65
ο
5 10 15 20 25 30 35 40 45 50 55 60 65
2θ [ ] ο
2θ [ ]

Fig. 5. XRD patterns for mortars (A) 0% LL, (B) 15% LL, (C) 25% LL and (D) 35% LL in 90 days.

Table 3 i.e. expansion, which is caused by ettringite formation. Finally, it

Ettringite’s percentage according to Rietvield method in cured mortars after 90 days. should be observed that where ettringite was present, the portlan-
Sample Conditions dite peaks indicated a decrease in its amount, which verifies that
20 °C 50 °C 50 °C + Na2SO4
most of the Ca(OH)2 was consumed in the ettringite formation
reactions.
0% LL 1.07 4.2 5.99
In cases where only increased temperature was applied, DEF
15% LL 1.26 2.95 6.71
25% LL 1.24 2.66 7.36 type I was formed as all of the sulfates originated from the interior
35% LL 0.85 2.34 7.47 of the samples. It must be noted that in case of limestone substitu-
tion, DEF formation was limited in regard to the reference sample.
This must be attributed to decreased C3A presence due to cement
CaCO3 þ Na2 SO4 ! CaSO4 þ Na2 CO3 ð5Þ substitution as shown previously.

Reaction (5) takes place faster than (1) and the formed gypsum
reacts at an early stage with the unhydrated C3A resulting in 3.6. Thermogravimetric analysis
ettringite being formed according to reactions (3) and (4). Addi-
tionally, calcite turned into portlandite also contributes to reaction Differential thermogravimetry analysis (DTG) curves of the
(1). This ettringite formation takes place with a higher rate as gyp- damaged surfaces of the mortars treated with 5% Na2SO4 and
sum exists in the solution and thus has increased mobility and 50 °C are presented in Fig. 6. The peak near 130 °C is attributed
reactivity. Similar observations have been made by Kawaura et to the dehydration (water loss) of ettringite. The peak at 450 °C
al. [22], who have showed that sulfate presence along with in- is attributed to Ca(OH)2 and the peak near 700 °C is attributed to
creased calcite and portlandite presence result in volume change, the loss of CO2 and is related to the existence of calcite. The DTG
3588 E. Adamopoulou et al. / Construction and Building Materials 25 (2011) 3583–3590

-0,009

-0,018

-0,027
400 440 480

Fig. 6. DTG results of mortar samples cured for 90 days at 50 °C, immersed in sulfate solution.

A 10 μm B 10 μm

Fig. 7. SEM image of mortar sample 0% LL cured at 50 °C for 90 days with sulfates (A) and without sulfates (B).

A 20 μm B 20 μm

Fig. 8. SEM image of mortar sample 15% LL cured at 50 °C for 90 days with sulfates (A) and without sulfates (B).

results furthermore support the observations made with XRD, as
portlandite is decreased for samples with limestone (see also the
inlayed plot).
3.7. Scanning Electron Microscopy (SEM)
The surface of damaged mortars was examined with SEM. In
specific, samples cured at 50 °C and samples cured at 50 °C in pres-
ence of sulfates were studied, as these were the samples where
damage and expansion occurred. A series of images were collected
in order to ensure that a representative and statistically sound
depiction of the damaged structure was obtained. Selected images
are shown in Figs. 7–10.
By observing the images one can see the form of ettringite
which causes the expansion problems. In all cases where limestone
was present, ettringite formed due to sulfates and increased tem-
perature (depicted as A in each sample) is found to be needle-like
and less than 10 lm in length. This verifies that under the afore-
 E. Adamopoulou et al. / Construction and Building Materials 25 (2011) 3583–3590
A 10 μm
Fig. 9. SEM image of mortar sample 25% LL cured at 50 °C for 90 days with sulfates (A) and without sulfates (B).
A 10 μm
Fig. 10. SEM image of mortar sample 35% LL cured at 50 °C for 90 days with sulfates (A) and without sulfates (B).
mentioned conditions, DEF formed is type II which is known to
cause expansion in mortars [6]. On the contrary, DEF formed due
to increased temperature is shown to be increased in size
(20 lm), however retaining its needle-like shape. As the expan-
sion results showed previously, type I DEF can also be responsible
for the damaging of mortar samples but in a limited degree rele-
vant to the external sulfate attack effect.
4. Conclusions
This study has showed how curing conditions can result in sur-
facial damage to white cement mortars containing limestone after
90 days. It is shown that increased temperature (50 °C) can result
in formation of DEF type I which causes mediocre expansion. In
the case of cements containing similar amounts of SO3, the amount
of DEF created is related to C3A amount, which, in the case of lime-
stone substitution, is limited. As SEM analysis showed, DEF created
under increased temperature was needle-like in shape and near
20 lm in length.
In the case of curing at 50 °C and in presence of sulfates, the
external sulfate attack phenomenon was observed as DEF occurred
in higher concentrations. Formation of DEF type II was increased
due to the high concentration of sulfates, as shown from the Rietv-
ield quantitative analysis. Additionally, the presence of limestone
contributed greatly to the formation of ettringite, as calcite reacted
with Na2SO4 to create additional amounts of gypsum, which was
consumed to create ettringite. It must be noted that even though
limestone substitution meant decreased amounts of C3A, mortars
3589
B 10 μm
B 10 μm
containing limestone exhibited higher concentration of ettringite
under 50 °C and sulfate attack due to increased calcite content.
Regarding ettringite in all, this work showed how the curing
conditions result in the formation of two different forms of the
crystal. In specific, type I DEF was created when increased temper-
ature was applied, while type II DEF was found when increased
temperature was applied and sulfate solution was present. Both
forms of ettringite resulted in externally damaged mortars and
had a similar shape. However, it was the type II ettringite that
caused higher expansion and greater damage, something that is
owed to its small size and – in this case – higher rate of formation.
The study of sulfate induced DEF and its effect on expansion of
white cement mortars containing limestone continues, as the
authors plan to investigate these phenomena in mortar samples
of the same strength class.
References
[1] Hanechara S, Oyamada T, Fukuda S, Fujiwara T. Delayed ettringite formation
and ALKALI aggregate reaction. In: Proceedings of the 8th int conference on
creep, shrinkage and durability mechanics of concrete and concrete structure,
Ise-Shima, Japan, vol. 2; 2008. p. 1051–6 [30 September–2 October].
[2] Collepardi M. A state of the art review on delayed ettringite attack on concrete.
Cem Concr Compos 2003;25(4–5):401–7.
[3] Lawrence CD. Mortar expansion due to delayed ettringite formation. Effects of
curing period and temperature. Cem Concr Res 1995;25(4):903–14.
[4] Glasser FP. The role of sulfate mineralogy and cure temperature in delayed
ettringite formation. Cem Concr Compos 1996;18(3):187–93.
[5] Diamond S. Delayed ettringite formation-processes and problems. Cem Concr
Compos 1996;18:205–15.
3590 E. Adamopoulou et al. / Construction and Building Materials 25 (2011) 3583–3590
[6] Stark J, Bollmann K. Delayed ettringite formation in concrete. Bauhaus-
University. <http://www.faggruppeba.no/ikbViewer/Content/738949/doc-23-
2.pdf>.
[7] Neville A. The confused world of sulfate attack on concrete. Cem Concr Res
2004;34:1275–96.
[8] Santhanam M, Cohen MD, Olek J. Mechanism of sulfate attack: a fresh look:
part 1: summary of experimental results. Cem Concr Res 2002;32(6):915–21.
[9] Brown PW. An evaluation of sulfate resistance of cements in a controlled
environment. Cem Concr Res 1981;11:719–27.
[10] Pipilikaki P, Katsioti M, Gallias JL. Performance of limestone cement mortars in
a high sulphates environment. Concr Build Mater 2009;23:1042–9.
[11] Kuzel H-J. Initial hydration reactions and mechanisms of delayed ettringite
formation in Portland cements. Cem Concr Compos 1996;18:195–203.
[12] Kuzel H-J, Meyer H. 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, Dallas 1993. Duncanville,
Texas: International Cement Microscopy Association; 1993. p. 191–203.
[13] Klemm Waldemar A. Ettringite and oxyanion – substituted ettringites – their
characterization and applications in the fixation of heavy metals: a synthesis
of the literature. Portland Cement Association; 1998 [chapter 10].
[14] Klemm Waldemar A. Ettringite and oxyanion – substituted ettringites – their
characterization and applications in the fixation of heavy metals: a synthesis
of the literature. Portland Cement Association; 1998 [chapter 6].
[15] Barnett SJ, Macphee DE, Lachowski EE, Crammond NJ. XRD, EDX and IR
analysis of solid solutions between thaumasite and ettringite. Cem Conc Res
2001;32(5):719–30.
[16] Hasan B, Sedat A. Ettringite formation in historic bath brick-lime plasters. Cem
Concr Res 2003;33:1457–64.
[17] Collepardi M. Sulfate attack and alkali–silika expansion. In: Proceedings of 2nd
international symposium on concrete technology for sustainable development
with emphasis on infrastructure, Hyderabad, India; 2005. p. 55–67 [27
February–3 March].
[18] Collepardi M. Damage by delayed ettringite formation––a holistic approach
and new hypothesis. Concr Int 1999;21(1):69–74.
[19] Klemm WA, Miller FM. Internal sulfate attack: a distress mechanism at
ambient and elevated temperatures? PCA R&D serial no. 2125; 1997.
[20] Santhanam M, Cohen MD, Olek J. Mechanism of sulfate attack: a fresh look.
Part 2. Proposed mechanisms. Cem Concr Res 2003;33(3):341–6.
[21] Katsioti M, Patsikas N, Pipilikaki P, Katsiotis N, Mikedi K, Chaniotakis M.
Delayed ettringite formation (DEF) in mortars of white cement. Concr Build
Mater 2011;25:900–5.
[22] Kawaura J, Morikawa R, Oshita H. Analytical Study of the ettringite re-
generation in concrete due to sulfate erosion. In: Proceedings of the 8th int
conference on creep, shrinkage and durability mechanics of concrete and
concrete structure, Ise-Shima, Japan, vol. 2; 2008. p. 1037–1043 [30
September–2 October].