Materials and Design 103 (2016) 1–9

Contents lists available at ScienceDirect

Materials and Design

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

The influence of calcium content on the performance of metakaolin-

based geomaterials applied in mortars restoration
F. Allali a, E. Joussein b, N. Idrissi Kandri a, S. Rossignol c,⁎
a
Laboratoire de la Chimie Appliquée, Faculté des Sciences et Techniques B.P. 2202—Route d'Imouzzer, Fes, Morocco
b
Université de Limoges, Groupement de Recherche Eau Sol Environnement (GRESE) EA 4330, Faculté des Sciences et Techniques, 123 avenue Albert Thomas, 87060 Limoges cedex, France
c
Université de Limoges, Science des Procédés Céramiques et Traitement de Surface (SPCTS), UMR CNRS 7315, Ecole Nationale Supérieure de Céramique Industrielle (ENSCI), 12 rue Atlantis, 87068
Limoges cedex, France

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

Article history: The architectural heritage needs to be preserved, especially coatings, because they communicate the historical
Received 23 February 2016 building as it was created. The aim of this work is to develop a stable and durable restore coating mortar
Received in revised form 7 April 2016 which would be compatible with the traditional elements of masonry, based on metakaolin-geopolymer binder
Accepted 9 April 2016
mixed with traditional material of restoration. To do this pure metakaolin substitution in the Si-Na and Si-K
Available online 13 April 2016
geopolymers formulation were done by different amount of calcium hydroxide (lime) and calcium carbonate
Keywords:
(calcareous sand). The feasibility of their use in synthesizing geopolymers and investigations of their properties
Cultural heritage were performed using FTIR, XRD, and compressive strength. Briefly, the results traduce the role of CaCO3 as
Restoration binder quasi without dissolution in alkaline solution whatever the cation used (Na- or K-based), the polyconden-
Mortars sation reaction is effective in this case. Using the calcium hydroxide Ca(OH)2, the reaction is dominated by hydra-
Geopolymers tion process which clearly limit the polycondensation one for the short time (up to 1 h). The potential use of
Calcium geopolymer binder as restoration mortar for historical building is then effective since the reaction seems to be
FTIR quite effective with calcareous sand easily accessible with a low cost.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction
The rehabilitation of historical buildings is very important in terms
of history and culture of cities that reflect the characters of each region.
According to this, architectural heritage preservation has reached a ris-
ing interest for scientists, architects, engineers and archaeologists. Since
a long time, the most widely used binder in the construction activity
were lime based mortars (e.g. calcium lime, pozzolan lime, hydraulic
lime; [1]. So quasi all the historic monuments are lime based binder
and need same composition for their restoration one. However, several
authors indicated that the use of concrete mortars in our cultural heri-
tage showed several disadvantages. Indeed, the crystallization pro-
cesses of salts present in the cement-based mortars can induces
damage of the stone, or also the presence of structural problems,
when movements of building is effective, due to the low flexibility of
the cement-based mortars [1,2,3,4]. According to these facts, it will be
interesting to obtain a palliative of cement-based concrete having nice
properties (acidic resistance, no leaching materials, high compressive
strength…). In Morocco, an important number of monuments are
inscribed on the World Heritage List. Unfortunately these built heritages
are subjected to a number of potentially harmful deterioration agents
⁎ Corresponding author.
E-mail address: sylvie.rossignol@unilim.fr (S. Rossignol).
during their long lifecycle. As well known, the long-term exposition to
air pollution, rainfall, freeze–thaw cycles and, most of all, salt crystalliza-
tion may cause significant damages to the constituent materials of the
masonries [5,6], especially coatings because they are the most exposed
to environmental degradation factors. It is therefore necessary to de-
velop strategies of monument restoration based on new insights. More-
over, it is well known that the restoration of cultural heritage is realized
using natural raw materials which are localized near the monument. So
the new concrete can be realized with classical raw material largely
found around the world. The kaolin is a well done candidate. One of
the most innovative applications for kaolin's is their use as inexpensive
raw materials in the synthesis of geopolymers. Indeed, there has been
increasing interest in producing eco-friendly geopolymers from natural,
inexpensive clayey materials [7,8,9]. Geopolymers are new materials
that are obtained through the activation of aluminosilicates and
alumina- and silica-rich materials such as calcined clays which are
used as precursors [10] and appear to be a promising alternative for
conventional cementitious materials due to their hydrothermal stability
and resistance to acidic environments and/or reducing agents [11,12].
They are also known for their promising mechanical and thermal prop-
erties [13,14].These new materials exhibit similar properties of natural
rocks. So, it's could be a good alternative of classical material in restora-
tion work [15]. However, the presence of secondary minerals in kaolin
ore deposit is quite effective inducing no pure kaolin except specific

http://dx.doi.org/10.1016/j.matdes.2016.04.028
0264-1275/© 2016 Elsevier Ltd. All rights reserved.
2 F. Allali et al. / Materials and Design 103 (2016) 1–9
Table 1
Characteristics of the raw materials used.
Raw material Chemical composition (wt%)
Metakaolin M1 55 SiO2; 40 Al2O3; 0.8 (K2O + Na2O);
1.4 Fe2O3; 1.5 TiO2; 0.3 (CaO + MgO)
Calcitic sand 84.9 CaO; 6.4 SiO2; 4.5 Al2O3; 1.4 Fe2O3
site. So it is difficult to obtain a constant elemental composition. More-
over, the use of high purity metakaolin in the geopolymer for restora-
tion can induce some differences in term of colors, and aspects which
is a disadvantage. One of the solutions is to use not pure natural raw ma-
terial. The objective of this work is to develop a stable and durable repair
coating mortar which would be compatible with the elements of ma-
sonry, using a metakaolin-geopolymer binder mixed with traditional
material of restoration (lime and calcareous sand: Ca-rich component).
Since geopolymer is traditionally devoid of calcium, it is necessary to
focus on the effect of pure metakaolin substitution in the Si-Na and Si-
K geopolymers formulation by different amount of calcium hydroxide
(lime) and carbonate in the formulation in the way to understand the
calcium behavior in the geopolymer binder. First, the physical and
chemical properties of the raw materials were briefly investigated,
and then introduced in the geopolymer formulation. The feasibility of
their use in synthesizing geopolymers and investigations of their prop-
erties were performed using FTIR, XRD, and their compressive stress
was also evaluated.
2. Materials and methods
2.1. Mortar designs and mixing procedures
Two raw materials designed for the coatings restoration were used
to perform two sets of consolidate materials: a calcareous sand
(CaCO3) and calcium hydroxide (Ca(OH)2). The calcareous sand is not
pure calcite and contains quartz and a little amount of dolomite and
muscovite. Moreover, a metakaolin M1 supplied by Imerys was also
used in the formulation of mortar. The characteristics of the raw mate-
rials used in this study are described in Table 1. Various types of geoma-
terials were synthesized by substitute metakaolin M1 with
(i) calcareous sand or (ii) calcium hydroxide in the formulation of an or-
dinary geopolymer using a Na or K alkaline silicate solution supplied by
Woellner. In the way to compare, the same Si/M ratio (Si/M = 1.7; M =
Na or K) was used for the two silicate solution. The protocol used is de-
scribed in Fig. 1. Briefly, pellets of NaOH (VWR, 97% pure) or KOH (VWR,
85.2% pure) were dissolved in sodium silicate or potassium silicate in-
ducing a Si/M ratio near 0.89 and 0.58 respectively. Then, silicate solu-
tion synthesized was mixed with metakaolin M1 and alternative
materials (calcareous sand or calcium hydroxide) in the way to increase
Fig. 1. Synthesis protocol of consolidated materials.
the Ca/Si ratio for each component. The reactive mixture was homoge-
nized and then placed in an opened sealed mold in ambient air (25 °C,
40% R.H.). The nomenclature used for the consolidate materials ob-
tained with calcareous sand is MK1 − x(CaCO3)x SiNa and MK1 −
x(CaCO3)x SiK for a MK1 – CaCO3 substitution with sodium or potassium
silicate solution, respectively. Same nomenclature is applied with cal-
cium hydroxide – MK substitution i.e. MK1 − x(Ca(OH)2)x SiNa and
MK1 − x(Ca(OH)2)x SiK.
2.2. Characterization technics
Chemical composition of each raw material was performed from X-
ray fluorescence (XRF) analyses using an XMET 5100 device commer-
cialized by OXFORD Instruments. Acquisitions proceeded from pressed
pellets for 300 s time analyses. The synthesized materials were charac-
terized from CarMaLim platform (University of Limoges, France) using X
Ray diffractometer with a BRUKER AXS D8 Advance powder diffractom-
eter using CuKα radiation (λKα = 0.154186 Å) over the 2θ range from
5° to 70° with an acquisition time of 2 s and a step size of 0.04°. Files
from the Joint Committee on Powder Diffraction Standards (JCPDS)
were used for phase identification. Infrared spectroscopic investigations
were realized from a Thermo Fisher Scientific 380 infrared spectrometer
(Nicolet) using the attenuated total reflection (ATR) method. The IR
spectra were gathered over a range of 400–4000 cm−1 with a resolution
of 4 cm−1. The acquisition of spectra has started just after the mixing.
The atmospheric CO2 contribution was removed with a straight line be-
tween 2400 and 2280 cm−1. To monitor the geopolymer network for-
mation, a software was used to acquire a spectrum (64 scans) every
10 min for 13 h. The software realized a spectral acquisition in an auto-
matic way at regular intervals.
The compressive strength of samples stored for 7 days were deter-
mined using LIyodEZ20 machine moving at constant cross-head dis-
placement of 0.2 mm/min. Test tubes, used for the compression tests,
were cylindrical in shape with a diameter (Φ) of 15 mm and a height
(h) of approximately 35 mm and were aged in closed mold at room
temperature. The samples were demolded and rectified just before the
tests. Samples were rectified by means of a diamond wheel to obtain
perfectly parallel and flat faces. All values presented in the current
work are an average of 10 samples. The experimental error is obtained
from the average of standard deviation.
3. Results and discussion
3.1. Calcareous sand-based geomaterials
3.1.1. Calcareous sand and metakaolin mortar from sodium silicate solution
The infrared spectra of the various mixtures from metakaolin – cal-
careous sand with sodium silicate solution are reported in the Fig. 2a.
 F. Allali et al. / Materials and Design 103 (2016) 1–9 3

In the way to facilitate the understanding, only the initial IR spectra at t0
and at the end of the reaction tfinal) are shown for each MK – calcareous
sand mixtures realized (MK1 − x (CaCO3)x SiNa with x = 0, 0.47, 0.65
and 1). Whatever the mixtures realized, there is no difference between
the initial and the final IR spectra. This fact clearly traduces the no reac-
tive mixture.
IR spectra present a broad band between 4000 and 3000 cm− 1
assigned to the vibration of OH bonds [16] and a band at 1655 cm−1
characteristic of the presence water [17]. The strong band localized
near 1420 cm−1 is assigned to the asymmetric stretching vibration
out of the plane of the C\\O bond of the calcium carbonate. Two others
bands which are characteristics of carbonate are effective at 875 and
713 cm−1 [18,19,20]. Finally, the intense band around 985 cm−1 is spe-
cific of the Si\\O\\Si bond. Regardless of the geopolymer literature, the pres-
ence of a Si\\O\\Si bands (Q2) shift at 980 cm−1 with time associate with a
decrease in the H2O bands at 3255 and 1620 cm−1 is characteristic of a
polycondensation reaction [21].
The evolution of Si\\O\\M bands were plotted against time in the Fig. 3a.
As expected, there is no evolution of the band with time whatever the
composition. Then, the addition of calcareous sand in the composition
of a sodium-based geopolymer does not change the position of the
Si\\O\\M band with time. This trend evidence that calcareous sand is
quite stable in sodium silicate alkaline solution. The consolidated mate-
rial is relative to the coating of calcareous sand particles by siliceous spe-
cies. Due to the presence of sodium alkali solution as reactive mixture,
the formation of sodium carbonate (Na2CO3) which can be act as poten-
tial binder may be possible like in natural sedimentary environment for
calcareous rocks. According to this, the Fig. 4a shows the evolution of
the intensity ratio bands of 1420/1390 (i.e. CaCO3/Na2CO3 characteristic
bands [22]) plotted as a function of time. There is no differences

Fig. 2. FTIR spectra at t = 0 and t = 400 mn for MK1 − x(CaCO3)x with (a) SiNa for x = 1, x = 0.65 and x = 0.47 and (b) with SiK for x = 1, x = 0.52 and x = 0.3.
4 F. Allali et al. / Materials and Design 103 (2016) 1–9

Fig. 3. Evolution of the shift of the Q2 position from FTIR spectra versus time for MK1 − x(CaCO3)x mixture with (a) SiNa and with (b) SiK silicate solution.

whatever the composition, the intensities ratio of band are N 1 which
means that the contribution of the characteristic bands at CaCO3 is
larger than that of Na2CO3. This is in accordance with the fact that
there is no dissolution of the calcareous sand, the consolidation of the
material may be due to the geopolymer networks relative to the Si\\O\\Al
band as classically observed in the literature.
3.1.2. Calcareous sand and metakaolin mortar from potassium silicate
solution
The infrared spectra of the various mixtures from metakaolin – cal-
careous sand with potassium silicate solution (MK1 − x (CaCO3)x SiK
with x = 0, 0.32, 0.52 and 1) are reported in the Fig. 2b. The results
give show differences in the IR spectra with the increase of the CaCO3

Fig. 4. Evolution of the intensity ratio of carbonate bands from FTIR spectra versus time for MK1−x(CaCO3)x mixture with SiNa (a), and with SiK (b) in function of 1420/1360 ratio intensity
band, and finally (c) SiK mixture in function of 1420/1390 ratio intensity band.
 F. Allali et al. / Materials and Design 103 (2016) 1–9 5

amount, differences being effective for a substitution rate N 0.32. Glob- Table 2
ally, the major bands observed for SiK materials mixtures is similar to Mechanical properties at 7 days cure of the MK and (i) calcareous sand mixtures MK1 −
x(CaCO3)x and (ii) calcium hydroxide mixtures MK1 − x(Ca(OH)2)x with sodium silicate
the SiNa one (see before) but differs drastically in position and with
solution (SiNa) or potassium silicate solution (SiK).
time. The broad band at about 3400 cm−1 at the initial state (t0) de-
creases with time due to polycondensation [23]. Same behavior is ob- Mixtures Substitution rate Maximum compressive
served for the carbonate bonds. The calcite one at 1420 cm− 1 (x values) strength in MPa (7 days)

decreases with time to give two main bands at 1460 and 1360 cm−1. MK1 − x(CaCO3)x SiNa x=0 2
These two bands can be attributed to the formation of potassium car- x = 0.47 77
x = 0.65 63
bonates (K2CO3) [24]. In the Si\\O bonds, there is a new band which ap-
x=1 2
pears at around 930 cm− 1 attributed to the vibration of Si\\O\\Ca [25] MK1 − x(CaCO3)x SiK x=0 42
from the beginning of the reaction (t0). After 400 min (tfinal) it seems x = 0.32 16
that the Si\\O\\M bands evolve. x = 0.52 41
To highlight the evolution of the Si\\O\\M band, the Fig. 3b shows the Q2 x=1 2
MK1 − x(Ca(OH)2)x SiNa x=0 2
shift (noted x) as well as the Si\\O\\Ca bands (noted x′) versus time. On one x = 0.41 10
hand, the Si\\O\\Ca band does not evolve with time, the band is effective x = 0.68 7
whatever the mixture (except for x′ = 0) from the beginning of the re- x=1 1
action (t0) traducing the relatively high solubility of CaCO3 in KOH me- MK1 − x(Ca(OH)2)x SiK x=0 42
x = 0.45 40
dium. This dissolution induces the presence of reactive Ca2+ leading the
x = 0.76 2
formation of the bond. For the Si\\O\\M, an important shift is observed from x=1 2
980 cm−1 to 945 cm−1 (x = 1) and about 960 cm−1 for the others mix-
tures. This is due to the formation of the Si\\O\\Al bonds as already explain
for polycondensation reaction [21,26,27]. SiK solution, the substitution of a part of Si\\O\\M by the Si\\O\\Ca allows the
The Fig. 4b and c show the evolution of carbonate bands against time consolidation of materials even with high amount of calcareous sand.
during the consolidation of the mixtures using the 1420/1360 intensity Contrary to SiNa solution, the calcareous sand is partially dissolved in
ratio (i.e. CaCO3/K2CO3 characteristic bands). The 1420/1360 intensity SiK media allowing the formation of Si\\O\\Ca bond leading to better me-
ratio is N1 between t0 and t150 mn. Beyond t150 mn the ratio is b 1 meaning chanical stress whatever the amount of calcareous sand add. This fact
that the band of calcium carbonate (1420 cm−1) had a majority of the can be shown by XRD (Fig. 6). Indeed, in the case of the SiNa-based sam-
beginning of the reaction whereas potassium carbonate subsequently ple, there is no dissolution of the calcareous sand which can be evi-
becomes the majority at the end. The results obtained clearly show denced by similar calcite peaks intensity. The presence of a small
that calcareous sand is dissolved in KOH-SiK media compare to NaOH- amorphous dome centered near 28°(2θ) traduce the polycondensation
SiNa one. The dissolution of calcareous sand allows the formation of of the geopolymer network [29]. This is not the case for the SiK one.
Si\\O\\Ca bonds and Si\\O\\Al one. This fact can play a main role in the working The XRD pattern give evidence of the decrease of the calcite intensity
properties of the synthesized materials. peaks associated with the presence of amorphous dome centered at
30°(2θ). This dome traduces the presence of amorphous and reactive
3.1.3. Mechanical properties of calcareous sand based-materials binder, i.e. the calcitic network and in little amount the geopolymer
The potential development of a stable and durable repair coating one. This fact explains also the increase of mechanical properties in
mortar compatible with the elements of masonry, using a metakaolin- the case of SiK-based materials.
geopolymer binder mixed with traditional material of restoration
(lime and calcareous sand) need good working properties such as me- 3.2. Calcium hydroxide (Ca(OH)2)-based geomaterials
chanical one. Then the performances of each synthesized materials
were tested from compression tests after 7 days curing. The results of 3.2.1. Calcium hydroxide (Ca(OH)2) and metakaolin mortar from sodium
these mechanical tests are shown in Table 2. In the case of the sodium silicate solution
silicate solution, the replacement of MK with calcareous sand decreases The protocol described in Fig. 1 was adopted to elaborate calcium
the mechanical properties from 80 to 2 MPa. These results are in accor- hydroxide-based geomaterials. In this case the metakaolin is substituted
dance with the IR data that evidenced the no dissolution of CaCO3 then by calcium hydroxide instead of calcareous sand. As for calcareous sand,
the consolidated form of synthesized materials: when the amount of IR spectra are performed from each mixtures realized MK1 −
unreactive species (i.e. CaCO3 vs MK) increases, the mechanical proper- x(Ca(OH)2)x with SiNa reactive solution and are reported in the
ties decreases. For the potassium silicate solution as reactive precursor,
the mechanical properties are globally lower than SiNa one but the me-
chanical performance are stable (near 40 MPa) from 32% of CaCO3 addi-
tion. This fact, in accordance with IR spectra (see before), reflect the
restructuration of the network relative to the dissolution of calcareous
sand toward SiK solution. The new materials formed show stable me-
chanical properties due to the Si\\O\\Ca bonds created. The maximum
stresses obtained from various CaCO3-based materials are plotted
against the Si/(Ca + Al + M) ratio, where M = Na or K in the Fig. 5.
Whatever the alkaline solution used, the graph clearly shows that the
mechanical strength increases with the increase of the Si/
(Ca + Al + M) ratio. However, the slope is not the same as already ex-
plained before (Table 2). So, the introductions of calcium element in a
SiNa-based composition have no impact: the calcareous sand acts as re-
inforcement without reacting with other components. As shown by [21,
28], the increase of the amount of reinforcement phase in the mixture
induces the decrease of mechanical properties. This fact can be ex-
plained by the non-reactivity of the calcareous component, then the ab- Fig. 5. Compressive strength (MPa) of synthesized geomaterials cured at 7 days-aged
sence of reactive calcium element in the sodium mixture. In the cases of based on MK1 − x(CaCO3)x with (▲) SiK and (♦) SiNa silicate solution.
6 F. Allali et al. / Materials and Design 103 (2016) 1–9

Fig. 7a. Firstly, spectra clearly evidences the absence of carbonate phase
whatever the time evolution and the formulations. As already explained
before, the IR spectra show the presence of a broad band between 3600
and 3200 cm−1 assigned to the vibration of OH bonds and a band at
1655 cm− 1 characteristic of the presence water. There is a vibration
band between 1010 et 970 cm−1 which differs according to the amount
of Ca(OH)2 in the mixture. This band is assigned to the Si\\O\\M (M = Si or
Al) vibration band. The presence of the 930 cm−1 band is effective from
t0 for all the mixture and is attributed to the Si\\O\\Ca vibration band. The
results evidenced the direct dissolution of the Ca(OH)2 component in
the SiNa media. Moreover, since there is no carbonate formed, this
fact traduce that the Ca cations are completely used to form the Si\\O\\Ca
bonds.
The evolution of the Si\\O\\M bands is plotted versus time in the Fig. 8a
(x for Si\\O\\M bonds and x′ for Si\\O\\Ca one). The results show a constant
Fig. 6. X ray diffractograms of synthesized geomaterials cured at 7 days based on MK1 −
evolution for Si\\O\\Ca band from the beginning of the reaction. However,
x(CaCO3)x with (a) SiK (x = 0.52) and (b) SiNa (x = 0.47).

Fig. 7. FTIR spectra at t = 0 and t = 400 mn for MK1 − x(Ca(OH)2)x with (a) SiNa and with (b) SiK silicate solution.
 F. Allali et al. / Materials and Design 103 (2016) 1–9
Fig. 8. Evolution of the shift of Q2 position from FTIR spectra versus time for MK1 − x(Ca(OH)2)x mixture with (a) SiNa and with (b) SiK silicate solution.
in the case of the Si\\O\\M band, the band shift rapidly to the higher wave-
number (from 980 to about 1000 cm−1). This fact can be explained by
the creation of news Si\\O\\Ca bands which can be relative to the linked
bands between each precursor of calcium silicate hydrate [30]. So this
band can traduce the restructuration of the calcium silicate hydrate net-
work. In the case of the x = 0.41 mixture, there is previously a shift from
1000 to 980 cm−1 traducing the polycondensation process then an in-
crease of the shift band to the higher wavenumber due to the Si\\O\\Ca
band (see before). This fact can traduce that the polycondensation pro-
cess is more rapid than the hydration one which need about 50 min.
This is in accordance with the literature [31]
3.2.2. Calcium hydroxide (Ca(OH)2) and metakaolin mortar from potas-
sium silicate solution
The Fig. 7b shows the IR spectra of the MK1 − x(Ca(OH)2)x with SiK
reactive solution. There is no real difference with the MK1 −
x(Ca(OH)2)x SiNa see before (Fig. 7a) except the presence of carbonate
bonds at 1390 cm− 1 attributed to a mixed Ca,K-carbonate after
400 mn. There is no change with time in the position of the bands but
their amount of carbonate increase with the increase of Ca(OH)2. Fi-
nally, the presence of Ca(OH)2, namely portlandite at 3630 cm− 1, is
also evidenced in each formulation and already increase with the
amount of x. This is due to the no complete dissolution of Ca(OH)2 in
the SiK-KOH media. In the 900–1100 cm−1 range, the main change is
the large increase of the band located at 930 cm−1 with time curing.
As for the others mixtures, the evolution of the Si\\O\\M bands versus
time is show in the Fig. 8b (x for Si\\O\\M bonds and x′ for Si\\O\\Ca one). As
for the SiNa media, the Si\\O\\Ca band is effective from the beginning of
the reaction (no shift; 930 cm−1) and stable with time. The Si\\O\\M
band shift rapidly to the higher wavenumber (from 980 up to
1100 cm−1 for x = 0.70). As before, this is probably due to precursor
of calcium silicate hydrate [30]. The presence of the Si\\O\\Al band being al-
ready present but with a low intensity and superposed with the Si\\O\\Ca
one. As for the MK1 − x(Ca(OH)2)x with SiNa reactive solution, there is
also dissolution of the Ca(OH)2 in the SiK one.
3.2.3. Mechanical properties of calcium hydroxide based-materials
The performances of each synthesized materials were evaluated
from mechanical tests after 7 days curing (Table 2). In the case of the so-
dium silicate solution, the addition of calcium hydroxide decrease dras-
tically the mechanical properties from 42 MPa for pure MK to 10, 7 and
1 MPa for x = 0.41, x = 0.68 and x = 1 respectively. The poor mechan-
ical strength can be explained by the uncompleted reaction or the fast
setting nature of the system. Indeed, high calcium content in the system
7
(x = 0.41, x = 0.68 and x = 1) induced fast setting hindered subse-
quent strength development [32] and then lower mechanical strength
properties. For the SiK media, the mechanical properties are quite stable
up to 45% of calcium hydroxide in mixture (about 42 MPa) then de-
crease to 2 MPa. The decrease of the mechanical properties can be at-
tributed to the increase of carbonation in the synthesized materials as
evidenced by FTIR spectra. This fact can be explained by (i) the carbon-
ate which growth in the porosity of the material induces constraint in
the network structure then a decrease of mechanical properties, and/
or (ii) the fast setting of the reaction as already explained by [32] in a
paper about alkali-activated cement, mortar and concrete. A well done
formulation of alkali-activated metakaolin mixture (geopolymer), de-
sirable strength and setting time of the mixture can be achieved under
the selected curing regime/temperature. When using the best optimum
of curing temperature of alkali-activated metakaolin (e.g. 60 °C) give the
best geopolymerization process. When the treatment (mixture) is real-
ized for another temperature or with a fast setting time, the strength de-
velopment is quite higher in few days but led to lower 28-day
mechanical properties. This fact is due to the increase in pore size and
pore volume of the paste as well as from specific surface area of
metakaolin used. Moreover, in the case of fly ash-based geopolymer,
there is an influence of the CaO content onto compressive strength
and setting time. The more the CaO amount increase, more the setting
time decreases. All these facts can explained the results obtained. The
Fig. 9. Compressive strength (MPa) of synthesized geomaterials after 7 days cured based
on MK1 − x((Ca(OH)2)x) with (▲) SiK and (♦) SiNa silicate solution.
8 F. Allali et al. / Materials and Design 103 (2016) 1–9
Fig. 10. X ray diffractograms of synthesized geomaterials at 7 days cured based on MK1 −
x(Ca(OH)2)x with (a) SiK (x = 0.45) and (b) SiNa (x = 0.41).
maximum stresses obtained from various Ca(OH)2-based materials are
plotted against the Si/(Ca + Al + M) ratio, where M = Na or K in the
Fig. 9. As for CaCO3-based material, the graph clearly shows that the me-
chanical strength increases with the increase of the Si/(Ca + Al + M)
ratio whatever the alkaline solution used. The introduction of Ca cation
in the mixture after dissolution of Ca(OH)2 phase induce large modifica-
tion of the reaction involved. Moreover, when Ca cation is rapidly avail-
able in the mixture, the better mechanical strength is obtain with SiK
media compare to the SiNa mixture. The XRD patterns obtained after
7 days of cure are shown in the Fig. 10 for two representative mixture
with an equivalent Ca amount. For each mixtures, the presence of an
amorphous dome centered near 30°(2θ) is effective. This dome traduces
the presence of hydrated phase (superposed with the geopolymer one).
However, the amorphous dome is higher in the case of K reactive mix-
ture than Na one. This is in accordance with the compressive strength
obtained which are better in the case of the K mixtures. Moreover, the
portlandite can be also evidenced in the case of the SiK formulation.
As already shown for the CaCO3 mixture, the SiK reactive solution is
more efficient than SiNa one for Ca(OH)2 geomaterials.
3.3. Explanation of the reaction involved
All the results evidenced that there is differences in the reactivity of
each mixtures traducing (i) the role of the Ca speciation: the Ca(OH)2
Fig. 11. Evolution of the shift of Q2 position from FTIR spectra for (●) MK1 − x(CaCO3)x
mixture with SiNa, (●) MK1 − x(CaCO3)x mixture with SiK, (♦) MK1 − x(Ca(OH)2)x
mixture with SiNa, (♦) MK1 − x(Ca(OH)2)x mixture with SiK, (▲) reference composition
SiNa, (▲) reference composition SiK.
mixture being more reactive than the CaCO3 one, and (ii) the role of
the activation solution since SiK-KOH is more reactive than SiNa-
NaOH one. According to this, and to better understand the reaction in-
volved for each MK1 − x(M)x mixtures at various Ca proportion, the
Si\\O\\M shift (Q2) after consolidation is reported relative to the Si/
(Ca + Al + M) ratio, where M = Na or K in the Fig. 11. Moreover, the
metakaolin MK poles are also reported for each solution. For the
MK1 − x(CaCO3)x synthesized materials, the Si/(Ca + Al + M) ratio de-
crease (equivalent to an increase of Ca amount) with the Si\\O\\M position
whatever the silicate solution used. However, the shift being more pro-
nounced for the SiK solution one. This fact traduces a better reactivity of
the SiK solution in accordance with the results obtained previously (see
before; Figs. 3 and 4). The reactivity is relative to the dominant
geopolymerisation reaction coupled with a small part of hydration pro-
cess (Si\\O\\Ca bonds) in the case of SiK-CaCO3 mixtures due to the low sur-
face dissolution of the calcareous sand (Figs. 4, 5 and 7). This point
induces better mechanical strength (Fig. 5). In the case of MK1 −
x(Ca(OH)2)x synthesized materials, same behaviors (but more pro-
nounced) is observed for the Si/(Ca + Al + M) ratio relative to the Si\\O\\M
position (from 950 to 925 cm−1 shift) whatever the silicate solution
used (Fig. 11). This result traduces clearly those main differences be-
tween the initial compositions with pure MK as well as the CaCO3-
based materials. Moreover the position of the band is much lower
near 930 cm−1. Since the Ca(OH)2 is highly reactive in alkaline media
(see before Figs. 4, 5 and 7), the presence of Ca2+ favoring the substitu-
tion of the Si\\O\\M band by the Si\\O\\Ca one.
Finally, the reaction involved in each case in alkaline silicate solution
media (CaCO3 vs Ca(OH)2) is quite different: (i) consolidation from
polycondensation process for CaCO3 mixture without calcareous disso-
lution or (ii) Si\\O\\Ca bonds with limited Si\\O\\Al one (beginning of hydra-
tion). In this later case, the presence of hydration processes can be
evidenced by infrared spectra traducing the duality reaction.
4. Conclusions
The effect of pure metakaolin substitution in the Si-Na and Si-K
geopolymers formulation by different amount of calcium hydroxide
(lime) and carbonate in the geopolymer formulation were realized in
the way to understand the calcium behavior in the geopolymer binder.
The results evidenced various behaviors:
(i) In the case of calcareous sand-based geomaterials, the results ev-
idenced the partial surface dissolution of the CaCO3 phases in the
K-silicate solution whereas no dissolution was observed for the
Na-silicate solution. These different behaviors induce differences
in the mechanical properties since the CaCO3 can play the role of
binder in Na-silicate solution. However, an increase of the me-
chanical strength is observed for the K-solution due to the dual
role of binder and Ca-reactive bond.
(ii) For the calcium hydroxide (Ca(OH)2)-based geomaterials com-
posite, the results show the partial (quasi-total) dissolution of
the Ca(OH)2 phase which react in alkaline media. The infrared in-
vestigations traduce that the polycondensation process is effec-
tive but limited and more rapid than the hydration one which
need about 50 min. After this time, the hydration reaction is
dominant governing the properties of synthesized materials.
Finally, the use of CaCO3 vs Ca(OH)2 can allow to govern the reaction
involved in the geomaterials: polycondensation process without calcar-
eous dissolution versus dominant hydration reaction when the Ca2+ is
available during the reactive mixture. The potential use of geopolymer
binder as restoration mortar for historical building is then effective
since the reaction seems to be quite effective with calcareous sand easily
accessible with a low cost.
 F. Allali et al. / Materials and Design 103 (2016) 1–9
References
[1] A. Pozo, Evolution of mechanical properties and drying shrinkage in lime-based and
lime cement-based mortars with pure limestone aggregate, Constr. Build. Mater. 77
(2015) 472–478.
[2] J. Lanas, J.I. Alvarez, Masonry repair lime-based mortars: factors affecting the me-
chanical behavior, Cem. Concr. Res. 33 (2003) 1867–1876.
[3] M. Mosquera, D. Benitez, S. Perry, Pore structure in mortars applied on restoration:
effect on properties relevant to decay of granite buildings, Cem. Concr. Res. 32 (12)
(2002) 1883–1888.
[4] E.A.W. Hendry, Masonry walls: materials and construction, Constr. Build. Mater. 15
(8) (2001) 323–330.
[5] A. Charola, Salts in the deterioration of porous material. an overview, J. Am. Inst.
Conserv. 39 (3) (2000) 327–343.
[6] S. Siegesmund, R. Snethlage, Stone in Architecture. Properties, Durability, fourth ed.
Heidelberg, Berlin, 2011.
[7] H. Xu, J.S.J. Van Deventer, Effect of source materials on geopolymerization, Ind. Eng.
Chem. Res. 42 (2003) 1698–1706.
[8] W.K.W. Lee, J.S.J. Van Deventer, The effect of ionic contaminants on the early-age
properties of alkali-activated fly ash-based cements, Cem. Concr. Res. 32 (2002)
577–584.
[9] A. Elimbi, H.K. Tchakoute, D. Njopwouo, The effect of calcination temperature of ka-
olinite clays on the properties of geopolymer cements, Constr. Build. Mater. 25
(2011) 2805–2812.
[10] J. Davidovits, Geopolymer: Chemistry and Applications, 2nd ed. St-Quentin, France,
2008.
[11] T. Bkharev, Resistance of geopolymer materials to acid attack, Cem. Concr. Res. 35-4
(2005) 658–670.
[12] N.K. Lee, H.K. Lee, Setting and mechanical properties of alkali-activated fly ash/slag
concrete manufactured at room temperature, Constr. Build. Mater. 47 (2013)
1201–1209.
[13] P. Duxson, S.W. Mallicoat, G.C. Lukey, W.M. Kriven, J.S.J. Van Deventer, The effect of
alkali and Si/Al ratio on the development of mechanical properties of metakaolin-
based geopolymers, Colloids Surf. A Physicochem. Eng. Asp. 292 (2007) 8–20.
[14] Z. Zhanga, J.L. Provisb, A. Reidc, H. Wanga, Mechanical, thermal insulation, thermal
resistance and acoustic absorption properties of geopolymer foam concrete, Cem.
Concr. Compos. 62 (2015) 97–105.
[15] T. Hanzlícek, M. Steinerová, P. Straka, I. Perná, P. Siegl, T. Švarcová, Reinforcement of
the terracotta sculpture by geopolymer composite, Mater. Des. 30 (2009)
3229–3234.
[16] L.J. Post, L. Borer, Physical properties of selected illites, beidellites and mixed-layer
illite-beidellites from southwestern Idaho and their infrared spectra, Appl. Clay
Sci. 22 (2002) 77–91.
9
[17] D. Bacui, J. Simitzi, Synthesis and characterization of a calcium silicate bioactive
glass, J. Optoelectron. Adv. Mater. 9 (11) (2007) 3320–3324.
[18] A. Bakolas, G. Biscontin, A. Moropoulou, E. Zendri, Characterization of the lumps in
mortars of historic masonry, Thermochim. Acta 809 (16) (1995) 269–270.
[19] T.L. Hughes, C.M. Methven, T.G.J. Jones, S.E. Pelham, P. Fletcher, C. Hall, Determining
cement composition by Fourier transform infrared spectroscopy, Adv. Cem. Based
Mater. 2 (3) (1995) 91–104.
[20] C. Genestar, C. Pons, Ancient covering plaster mortars from several convents and Is-
lamic and Gothic palaces in Palma de Mallorca (Spain): analytical characterization, J.
Cult. Herit. 4 (4) (2003) 291–298.
[21] E. Prud'homme, P. Michaud, E. Joussein, J.M. Clacens, S. Rossignol, Role of alkaline
cations and water content on geomaterial foams: monitoring during formation, J.
Non-Cryst. Solids 357 (2011) 1270–1278.
[22] S. Chunming, L.D. Suarez, Infrared speciation of absorbed carbonate on aluminum
and iron oxides, Clay Clay Miner. 45 (6) (1997) 814–825.
[23] P. Innocenzi, Infrared spectroscopy of sol–gel derived silica-based films: a spectra-
microstructure overview, J. Non-Cryst. Solids 316 (2003) 309–319.
[24] R.L. Mills, Highly stable novel inorganic hydrides, J. New Mater. Electrochem. Syst. 6
(2003) 45–54.
[25] S.R. Federman, V.C. Costa, D.C.L. Vasconcelos, W.L. Vasconcelos, Sol-Gel SiO2-CaO-
P2O5 biofilm with surface engineered for medical application, Mater. Res. 10 (2)
(2007) 177–181.
[26] A. Autef, E. Joussein, A. Poulesquen, G. Gasgnier, S. Pronier, I. Sobrados, J. Sanz, S.
Rossignol, Influence of metakaolin purities on potassium geopolymer formulation:
the existence of several networks, J. Colloid Interface Sci. 408 (2013) 43–53.
[27] A. Autef, E. Joussein, G. Gasgnier, S. Rossignol, The role of the silica source on the
geopolymerization rate: a thermal analysis study, J. Non-Cryst. Solids 366 (2013)
13–21.
[28] R.H. Kupaei, U.J. Alengaram, M.Z. Jumaat, The effect of different parameters on the
development of compressive strength of oil palm shell geopolymer concrete, Sci.
World J. 2014 (2014) 16.
[29] N. Essaidi, B. Samet, S. Baklouti, S. Rossignol, The role of hematite in aluminosilicate
gels based on metakaolin, Ceramics-Silikáty 58 (1) (2014) 1–11.
[30] I. Garcia-Lodeiro, A. Fernandez Jimenez, A. Palomo, D.E. Macphee, Effect of calcium
additions on N–A–S–H cementitious gels, J. Am. Ceram. Soc. 93 (7) (2010)
1934–1940.
[31] I. García-Lodeiro, A. Fernández-Jiménez, A. Palomo, Hydration kinetics in hybrid
binders: early reaction stages, Cem. Concr. Compos. 39 (2013) 82–92.
[32] P. Chindaprasirt, T. CaO, Setting, segregation and bleeding of alkali-activated ce-
ment, mortar and concrete binders, in: F. Pacheco Torgal, Labrincha JA, C. Leonelli,
A. Palomo, P. Chindraprasirt (Eds.), Handbook of alkali-activated cements, mortars
and concretes, first ed.Woodhead Publishing, 2014.