Applied Clay Science 161 (2018) 82–89

Contents lists available at ScienceDirect

Applied Clay Science

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

Research paper

Red-clay ceramic powders as geopolymer precursors: Consideration of T

amorphous portion and CaO content
Martin Kepperta, Eva Vejmelkováa, Petr Bezdičkab, Magdaléna Doleželováa, Monika Čáchováa,
⁎
Lenka Scheinherrováa, Jaroslav Pokornýa, Martin Vyšvařilc, Pavla Rovnaníkovác, Robert Černýa,
a
Department of Materials Engineering and Chemistry, Faculty of Civil Engineering, Czech Technical University in Prague, Thákurova 7, 166 29 Prague 6, Czech Republic
b
Institute of Inorganic Chemistry AS CR v.v.i., Husinec – Řež 1001, 250 68 Řež, Czech Republic
c
Institute of Chemistry, Faculty of Civil Engineering, Brno University of Technology, Žižkova 17, 602 00 Brno, Czech Republic

A R T I C LE I N FO A B S T R A C T

Keywords: Geopolymers are conventionally prepared from metakaolin, fly ash, or blast furnace slag. In this paper, geo-
Geopolymers polymer synthesis using two different red-clay ceramic powders and a varying amount of Na2O/SiO2 activator is
Red-clay ceramics presented. Experimental results show that despite of the high content of crystalline minerals and lower con-
Composition centration of amorphous matter, both tested precursors enable to prepare geopolymers with satisfactory me-
Amorphous content
chanical properties. The discussion of measured values of compressive strength with respect to SiO2/Al2O3 and
Calcium oxide
Al2O3/(Na2O + K2O) ratios in the prepared geopolymer mixes reveals that the consideration of only amorphous
portion of ceramics is appropriate. A higher content of CaO in the precursor is found to cause a broader pore size
distribution and higher porosity of resulting geopolymers, which is due to the formation of calcium silicate
hydrates.

1. Introduction materials with possible environmental and durability benefits (which

Geopolymers (or aluminosilicate inorganic polymers, alkaline acti-
vated materials) are materials synthesized from a suitable aluminosili-
cate precursor (i.e., monomer) by polymerization induced in alkaline
environment. The prepared materials consist of fundamental SiO4 and
AlO4− structural units connected through oxygen bridges (e.g., poly
(sialate) structure eSieOeAleOe) to polymer network (Brus et al.,
2016). Since the AlO4− unit is negatively charged, the presence of ca-
tions in the structure is necessary. The geopolymer synthesis, in a
simplified way, lies in: 1) dissolution of aluminosilicate precursor in
alkaline activator; 2) polycondensation of monomers to the final
polymer network; 3) solid state transformation to final hard solid (Yun-
Ming et al., 2016). As activators, a wide range of alkaline solutions is
applied: NaOH, Na2CO3, sodium silicate (water glass), sodium alumi-
nate, or their potassic or lithic equivalents (Hosan et al., 2016; Peys
et al., 2016; Hu et al., 2017). The “overall” composition of the geopo-
lymer can be described as follows (Davidovits, 1991):
nM2O·Al2O3∙xSiO2∙yH2O where M is Na, K, or Li, n ranges between 1 and
1.6, y between 3 and 7 and x is higher than 2 (or, the Si/Al molar ratio
is higher than 1). Obviously, the formula is applicable just for geopo-
lymers which do not contain CaO and MgO. Traditionally, geopolymers
are supposed to be first of all an alternative to Portland cement based
are paid by more sensitive mix design and technology, as compared to
ordinary Portland cement). Nevertheless, geopolymers can also be used
in other applications, frequently advantaging their thermal stability: for
production of thermally resistant construction elements, as adhesives,
for solidification of hazardous wastes, or as catalyst support (Pacheco-
Torgal et al., 2014; Topçu et al., 2014; Lemougna et al., 2016; Kovarik
et al., 2017; Provis, 2017).
There is a wide range of possible aluminosilicate precursors of
geopolymers, as well as activation solutions; obviously the most im-
portant geopolymer properties (setting/hardening kinetics, strength)
depend, besides other, on the composition of both principal raw ma-
terials. Probably the most widely studied aluminosilicate precursor of
geopolymers is metakaolin (dehydroxylated kaolin, MK). As the raw
material, either quarried clay or an industrial clayey waste (sludge from
paper or ceramics production) is used. The optimum calcination tem-
perature depends on the composition of particular clay and ranges from
600 to 900 °C (Rashad, 2013; Mohammed, 2017). The initial crystalline
structure of present clay minerals is disrupted during the dehydrox-
ylation and the material becomes reactive; obviously the higher level of
dehydroxylation (i.e., the amorphousness of the material) results in
better reactivity. The mechanochemical treatment of clay is also pos-
sible (MacKenzie et al., 2007; Yun-Ming et al., 2016). The important

⁎
Corresponding author.
E-mail address: cernyr@fsv.cvut.cz (R. Černý).

https://doi.org/10.1016/j.clay.2018.04.019
Received 1 March 2018; Received in revised form 11 April 2018; Accepted 13 April 2018
0169-1317/ © 2018 Elsevier B.V. All rights reserved.
M. Keppert et al.
Table 1
Selected parameters of metakaolins for geopolymer synthesis.
Ref. Content (wt%) d50 (um) Sg (m2/g)
SiO2 Al2O3
Kovarik et al., 2017 52.1 41.9 4.0 13.0
Borges et al., 2017 54.5 44.2 4.5 N/A
Belmokhtar et al., 2017 53.6 42.2 4.8 6.2
Lahoti et al., 2017 53.0 43.8 1.3 N/A
parameters of several calcined clays successfully used for geopolymer
synthesis are summarized in Table 1. Concerning the phase composi-
tion, metakaolin is usually believed to be fully amorphous, but it can
contain some crystalline residuals (kaolinite, muscovite, quartz, etc.)
(Belmokhtar et al., 2017). Other frequently used aluminosilicate pre-
cursors are coal fly ash (FA) and ground granulated blast furnace slag
(GGBS). The most important difference compared to MK is content of
CaO which in FA varies widely from 0 up to 16.5% (Reddy et al., 2016).
GGBS contains significantly higher amounts of CaO (30–40%) and MgO
(5–10%) (e.g. Fiala et al., 2017). FA and GGBS contain just minor
amount of crystalline minerals (< 20%, the most abundant are quartz
and mullite) (Font et al., 2010; Poole and Sims, 2016), the rest of
material is glassy. A geopolymer derived from GGBS and FA contains
both calcium silicate hydrates and “true geopolymer” structures (Li
et al., 2010).
In this paper, a relatively unusual aluminosilicate precursor, namely
red-clay ceramic grinding dust (RCC), which is an industrial by-product
generated in brick blocks calibration, is used for the preparation of
geopolymer mixes. The fundamental properties of two different RCC are
determined at first and compared with conventionally used alumino-
silicate precursors. Then, the properties of obtained geopolymers are
analyzed, taking into account the precursors' amorphous portion and
CaO content.
2. Experimental methods
Two kinds of red-clay based ceramic powders were used as geopo-
lymer precursors. They were generated as by-products of the calibration
of thermal insulating brick blocks and will be marked H and L in what
follows, according to the brick factories placed in Hevlin and
Libochovice localities (Czech Republic). The RCC powders were used
as-received, just the particles larger than 1 mm were removed by
sieving. The chemical composition of ceramic precursors was examined
via X-ray fluorescence (XRF) spectroscopy, using a Thermo ARL 9400
XP instrument. The obtained data were evaluated by Uniquant 4 soft-
ware. Diffraction patterns were collected using a PANalytical X'Pert
PRO diffractometer equipped with a conventional X-ray tube (CoKα
radiation, 40 kV, 30 mA, line focus) and a multichannel detector
X'Celerator with an anti-scatter shield. X-ray patterns were measured in
the range of 5 to 105° 2 Θ with a step of 0.0167° and 1000 s counting
per step. Conventional Bragg-Brentano geometry was used with the
following parameters: 0.02 rad Soller slit, 0.5° divergence slit, 1.0° anti-
scatter slit, and 15 mm mask in the incident beam, 5.5 mm anti-scatter
slit, 0.02 rad Soller slit and Fe beta-filter in the diffracted beam. Side
loaded sample holders were used to minimise the preferred orientation.
The duration of the scan was ca. 13.2 h. Quantification of the experi-
mental data was performed with the similar methodology as described
by Hradil et al. (2017) using the Rietveld method. The BGMN code was
used for all calculations (Bergmann et al., 1998). This program includes
a code which permits the use of structural models describing correctly
the disorder appearing frequently in the case of clay minerals (Ufer
et al., 2004, 2008). Models of other reference materials were described
as standard Rietveld models (ICSD, 2017). Refined parameters com-
prised unit cell parameters for all phases as well as size and micro-strain
broadening parameters. Preferred orientation was refined only for mica
83
Applied Clay Science 161 (2018) 82–89
Table 2
Composition of geopolymer mixes.
Geopolymer RCC (g) Water glass (g) NaOH (g) Water (g)
H/L1 200 70 2.57 41
H/L2 200 80 2.95 36
H/L3 200 90 3.33 31
(muscovite, illite). Spherical harmonic of the 2nd degree was used
(Bergmann et al., 2001). Additional refined non-structural parameters
were sample displacement error and a 6th degree polynomial for
background modelling. The amorphous portion of the materials was
quantified by help of added internal standard (ZnO) as described, e.g.,
by Madsen and Scarlett (2008). Zinc oxide was chosen as an internal
standard as its presence is not expected in all studied materials (no Zn
was found by the XRF spectroscopy). The particle size distribution was
determined using a laser diffraction analyzer (Fritsch, Analysette 22
MicroTec plus). Density of RCC powders was measured by helium
pycnometry (Pycnomatic ATC). Their pore system was characterized by
help of the nitrogen adsorption method (Sorptomatic 1990); the specific
surface area was determined by the BET method, the pore size dis-
tribution by the BJH method. The RCCs were also examined by TG/DSC
analysis (Labsys Evo, Setaram).
Geopolymers were prepared from the H and L precursors, with three
dosages of activating solution (Table 2) based on sodium water glass VS
1.6 (Vodní sklo, a.s., Czech Republic) and NaOH (p.a., Penta Chemicals,
a.s., Czech Republic). The silicate modulus of the activator was kept
constant (1.4); the water content – sum of dosed waster and water
present in sodium silicate solution – was also equal in all mixes. The
recipes were based on preliminary experiments performed with varying
activator composition and dosage. The varying dosage of liquid acti-
vator means varying Si/Al and Na/Al ratio in the prepared materials.
The mixed geopolymer paste was placed to 40 × 40 × 160 mm moulds,
covered by foil in order to prevent fast drying and cured in laboratory
conditions (23 °C) for 28 days.
The flow properties of pastes were characterized by help of a
Discovery HR-1 (TA Instruments) hybrid rheometer and TRIOS
4.0.2.30774 software for data evaluation. In order to avoid wall slip-
page, the building material cell and the paddle type rotor was adopted.
The gap thickness was chosen to be large enough (10 mm) for paste
dispersion. All the measurements were done at 25 °C. The mixed paste
was pre-sheared for 60 s at 100 s−1 in order to re-homogenize the
sample and to eliminate its shear history. After a period of rest of 60 s,
the rheological measurements were started. The testing routine com-
prised a shear rate increase (from 0.1 to 100 s−1) applied through 30
steps with 15 s of measuring time at each shear rate, followed by a
decrease of shear rate at the same conditions. The flexural strength (3
replicates) was measured by three point bending test (EN 196-1); the
compressive strength was determined by help of halves of prisms after
bending test (i.e., 6 replicates). The strength measurements were per-
formed by an FP 100 testing device (Heckert). The density, bulk den-
sity, and porosity were determined by help of helium pycnometry
(Pycnomatic ATC) and volume determination of prismatic samples
(from moulds 40 × 40 × 160 mm).
3. Results
3.1. Characterization of RCC precursors
The chemical composition of RCC is summarized in Table 3. The
line “total” means chemical composition, as it was determined by XRF.
The XRD analysis of RCC (Fig. 1, Table 4) revealed a high complexity of
these materials; the dominating mineral was quartz, followed by nu-
merous silicates which were partially contained in the loess and par-
tially synthesized during the ceramics firing. Terms microcline and
M. Keppert et al. Applied Clay Science 161 (2018) 82–89

Table 3
Chemical composition of H and L ceramic powders.
SiO2 Al2O3 Fe2O3 CaO MgO K2O Na2O TiO2 SO3 Sum

H Total 58.8 19.6 5.7 6.9 2.8 2.9 1.5 0.8 0.7 99.7
In minerals 51.1 10.0 2.3 2.3 1.2 2.9 1.5 0.0 0.0 99.7
Amorphous 7.7 9.6 3.4 4.6 1.6 0.0 0.0 0.8 0.7
L Total 49.9 20.4 5.0 15.4 2.8 3.3 0.5 0.8 1.5 99.6
In minerals 36.0 7.5 2.0 2.4 0.7 2.5 0.4 0.0 1.2 99.6
Amorphous 13.9 12.9 3.0 13.0 2.1 0.8 0.1 0.8 0.3

akermanite in Table 4 should be understood as representatives of very Table 4

similar minerals – microcline-orthoclase and akermanite-gehlenite. Phase composition of H and L ceramic powders.
Both RCC contained a residue of illite as the main clay mineral present H L
in the loess used for bricks production. Mullite was a product of high-
temperature decomposition of kaolinite. By a careful inspection of Quartz SiO2 26.2 20.0
measured X-ray pattern it was found that the basal lines of mica are Hematite Fe2O3 2.3 2.0
Albite NaAlSi3O8 13.0 3.5
split in two (Fig. 1, insert). Therefore, two mica phases were con- Microcline KAlSi3O8 7.1 10.7
sidered, i) illite that has not been fully decomposed during the pre- Muscovite KAl2(AlSi3O10)(OH)2 12.5 4.4
paration of RCC and ii) muscovite considered as an impurity in raw Illite K0.65Al2(Al0.65Si3.35O10)(OH)2 3.8 3.1
materials used for bricks firing. Neither calcite nor portlandite were Diopside CaMgSi2O6 4.4 1.6
Akermanite Ca2MgSi2O7 2.8 2.9
detected in any RCC. The most abundant “phase” in both RCC was
Mullite Al6Si2O13 – 2.9
amorphous material having origin in thermally decomposed illite and Anhydrite CaSO4 – 2.0
other clay minerals. The higher amorphous content in L was related to Amorphous – 27.8 47.0
the higher content of CaO indicating a higher content of calcium car-
bonate in the initial clay (Cultrone et al., 2004). The chemical com-
position of the sole amorphous matter is directly undetectable, but it (Table 4), but their position was typical for Ca(OH)2 and CaCO3 de-
can be calculated as the difference between the determined total che- composition (Sabeur et al., 2016). Expressed numerically, the material
mical composition of the whole matter (XRF) and the known content of H contained 0.8% of Ca(OH)2 and 0.9% CaCO3, while L contained 1.2%
individual oxides in present crystalline phases, as detected and quan- of Ca(OH)2 and 1.2% CaCO3. The reason for the absence of calcite and
tified by XRD. If one knows the content of amorphous matter in the portlandite in XRD results may be due to either their low amount (XRD
sample (Table 4), the composition of amorphous matter can be de- detection limit is about 1%), or their poor crystallinity. The summarized
termined (Williams and van Riessen, 2010). These two contributions to content of CaO in the Ca(OH)2 and CaCO3 was 1.1% in the material H
the total chemical composition are presented in Table 3 in rows “in and 1.6% in L. It means that unspecified portions of CaO remained in
mineral” and “amorphous”. The amorphous content was higher in L the amorphous part of both analyzed precursors.
than H. Obviously, the amorphous matter contained SiO2 and Al2O3, The particle size distribution of raw materials is shown in Fig. 3. The
but also relatively high amount of CaO (L in particular). curves were nearly identical what was obviously reflected by similar
An attempt to characterize the nature of CaO present in the amor- values of d50 and Blaine specific surface area (Table 5). The red clay
phous portion was performed by help of thermal analysis (Fig. 2). It ceramic is, in general, a porous material; thus its BET specific surface
should be noted that the measured relative mass change was very low area is significantly higher than Blaine surface area, which can be un-
(as it could be expected in a fired material, such as ceramic), thus the derstood as the outer surface area of particles. The porosity of particles
obtained TG curve was not smooth. Nevertheless, two mass losses were and pore size distribution, as measured by the BJH method based on
observed at 420 °C and 700 °C. They were not related to illite which was nitrogen adsorption, is presented in Fig. 4. The main portion of pores in
the only thermally unstable phase (Hulan et al., 2017) detected by XRD the ceramic powders had a diameter in the range of 0.08–1 μm; the L

Fig. 1. XRD patterns of RCC precursors H and L. (Mi - mica (muscovite, illite); Mc – microcline; A – albite; Mu – mullite; Q – quartz; D – diopside; Ak – akermanite; H
– hematite; An – anhydrite). A detail of Mica basal line splitting (muscovite and illite) is sketched in the insert.

84
M. Keppert et al. Applied Clay Science 161 (2018) 82–89

0.0 0.008

0.007
-0.2
H

pore volume increment (cm3/g)

0.006
-0.4 L
relative mass change (%)

0.005
-0.6
0.004

-0.8 0.003
H
0.002
-1.0
L 0.001
-1.2
0.000
0.001 0.01 0.1 1
-1.4
0 200 400 600 800 1000 pore diameter ( m)

temperature (°C) Fig. 4. Pore size distribution of H and L ceramic precursors obtained by BJH
method.
Fig. 2. Thermogravimetry of H and L ceramic powders.

3.2. Rheology of fresh geopolymer pastes

100
The mixtures were subjected to rheological measurements to de-
cumulative particle size distribution (%)

termine the effect of the dose of the activating solution on a workability

80 of the fresh pastes. Rheology was also used to compare the effect of
H different amorphous phase contents in individual precursors on the
L consistency, viscosity, and overall flow behavior of the fresh pastes.
60 Also the presence of CaO could dramatically change the chemistry of
the system and the interaction forces between particles, which can take
effect in rheology and overall workability of the mixtures.
40 The results of rheology experiments were expressed as shear rate vs.
shear stress (flow curves) and the Herschel-Bulkley model (1) was ap-
plied to downward flow curves to fit the experimental data and to de-
scribe the pastes rheological behaviour (Mezger, 2014):
20
τ = τ0 + kγ̇n (1)

where τ is the shear stress, γ̇ is the shear rate, τ0 corresponds to the

0 yield stress, k is the consistency coefficient, and n is the fluidity index
0.1
Fig. 3. Particle size distribution of H and L ceramic powders.
Table 5
Fundamental parameters of H and L ceramic precursors.
ρ (g/cm3) d50 (um) Sg
g)
H 2.628 42 0.2
L 2.654 41 0.2
precursor was found to be somewhat more porous; this can be attrib-
uted to the higher CaO content in the L ceramics (Sobott et al., 2014).
When the BJH results were compared with those obtained by mercury
intrusion porosimetry (MIP) of the compact brick body of similar
ceramics blocks (Pavlík et al., 2014), one observed a lower porosity
indicated by BJH. It was partly due to the absence of lightening pores in
powdered ceramics and partly due to overestimation of ink-bottle pores
volume by MIP. The porosity indicated by BJH can be understood as the
volume of pores present in ceramics without any lightening additive
(e.g. saw dust, cellulose fibers from papermaking sludge, cut straw, etc.)
which form, due to burning out, pores larger than 1 μm, what is the
highest pore diameter indicated by the BJH method.
1 10 100 1000
which characterizes shear-thinning (n < 1) or shear-thickening
particle size ( m) (n > 1).
The shape of flow curves (Figs. 5 and 6) indicated thixotropic
character of precursor-activator dispersions and exhibited a pseudo-
plastic (shear-thinning) behavior. The yield stress of pastes decreased
with the increasing dosage of alkali activator (Table 6); whence it fol-
lows that the pastes became more fluid. This was expected because the
Blaine (m2/ Sg BET (m2/ VBJH (cm3/ PBJH (%) higher silicate solution content provoked the solubility of the precursor
g) g) particles. The yield stresses of pastes were very low and even lower than
the yield stresses of common cement pastes (Banfill, 2003). This sug-
3.9 0.030 7.9
3.7 0.049 13.0 gests that the cohesion of the aluminosilicate gel produced of the
geopolymeric reaction was much weaker than the attraction between
CSH in the early stages. Colloidal interactions between precursor grains
suspended in the alkaline silicate solution were negligible and only low
energy interactions between grains in the presence of an interstitial gel
were at the origin of the small yield stress of the dispersions. The ele-
vated yield stress values could point to some colloidal interactions in
the L system due to the higher content of calcium. On the other hand,
the consistency coefficients of the pastes were markedly higher than of
cement pastes or metakaolin-based geopolymer pastes, despite of bigger
particle sizes of red-clay ceramics precursors than of cement or meta-
kaolin (Aboulayt et al., 2014). The activator solutions used for the red-
clay paste formulations are Newtonian fluids but with a viscosity which
is 10 to 100 times higher than the viscosity of water. As a consequence
the viscous dissipation in these fresh pastes was expected to be 10 to
100 times higher than that measured in a standard cement paste. The

85
M. Keppert et al. Applied Clay Science 161 (2018) 82–89

1600 Table 7
Fundamental physical properties of prepared geopolymers.
1400 ρ (kg/m3) ρbulk (kg/m3) P (%)

1200 H1 2632 1865 29.1

L1 2577 1628 36.8
H2 2643 1834 30.6
shear stress (Pa)

1000 L2 2491 1644 34.0

H3 2636 1806 31.5
L3 2501 1640 34.4
800

600 crystalline content were very reluctant to flow, especially if the grains
had a plate shape. However, the most viscous paste displayed the lowest
400 H1 fluidity index; it means that this paste was the most thixotropic.
H2
200
H3 3.3. Properties and composition of hardened geopolymers

0
0 20 40 60 80 100 The fundamental physical properties of synthesized geopolymers
are shown in Table 7. The H precursor was providing (with the same
shear rate (s-1)
activating system) a more compact product with higher matrix density
Fig. 5. Flow curves of H fresh pastes. Solid symbols – downward, open symbols and bulk density and lower porosity, as compared to L. The pore size
– upward. distribution of both material groups was different. The geopolymer
samples prepared from the H powder (Fig. 7) contained a narrow pore
1600 system with mean pore diameter of about 1 μm, while the pore system
of the L samples (Fig. 8) was broader. The mean pore diameter was
1400 lower in the L samples, while their threshold diameter was somewhat
higher. Luna-Galiano et al. (2016) performed porosimetry measure-
1200 ments in geopolymers prepared from FA (CaO content 4%) and mix-
tures of FA and GGBS (CaO content 46%). The pore system of samples
shear stress (Pa)

1000 containing just FA was unimodal with a maximum at 1 um, while the
adding of GGBS (i.e., CaO) caused bimodality of the pore system – a
800 new maximum at 0.01 um appeared due to CSH/CAH formation in
GGBS containing samples. The difference in pore size distribution be-
600 tween H and L samples thus can be attributed to higher content of CaO
in the L precursor, causing more intensive parallel CSH and CAH for-
400 mation.
The phase composition of prepared samples (Figs. 9 and 10, Table 8)
L1
200 reflects composition of ceramic precursors H and L; concerning the
L2
individual minerals, calcite is present in L-based geopolymers only due
L3
0 to the higher content of CaO in its precursor. Moreover, mullite was
0 20 40 60 80 100 found in L samples due to its presence in precursor L. Nevertheless, the
shear rate (s-1) information about chemical composition of prepared geopolymers is
hidden in the amorphous portion of prepared materials. Generally,
Fig. 6. Flow curves of L fresh pastes. Solid symbols – downward, open symbols
samples made from L contain more amorphous matter, as well as the L
– upward.

0.24
Table 6
The rheological parameters of pastes.
0.20 H1
τ0 (Pa) k (Pa s) n (-)
H2
H1 12.08 30.9 0.82
pore volume (cm3/g)

H2 3.43 24.3 0.94 0.16 H3

H3 0.74 25.9 0.90
L1 10.62 9.1 0.91
L2 6.05 45.2 0.75 0.12
L3 2.55 17.3 0.94

0.08
differences in viscosities of metakaolin-based geopolymer pastes and
the red-clay ceramic pastes might be due to the higher amount of Ca in
the red-clay precursors. The presence of divalent cations could result in 0.04
attractive ion correlation forces, whereas monovalent cations only in-
duced repulsive forces. One might then expect the colloidal interactions
0.00
between grains to be enhanced as well (Favier et al., 2014). The L2 0.001 0.01 0.1 1 10 100
system was the most viscous but generally the pastes made from the H
pore diameter ( m)
precursor exhibited a higher consistency coefficient. This was due to the
different content of the amorphous phase, where mixtures with high Fig. 7. Pore size distribution of geopolymer prepared from H precursor.

86
M. Keppert et al.
0.24
0.20 L1
L2
pore volume (cm3/g)
0.16 L3
0.12
0.08
0.04
0.00
0.001 0.01 0.1 1 10
pore diameter ( m)
Fig. 8. Pore size distribution of geopolymer prepared from L precursor.
precursor does.
According to (Davidovits, 1991), the formula of geopolymer is
nM2O·Al2O3·xSiO2·yH2O where M is Na, K, or Li, n ranges between 1 and
1.6, y between 3 and 7 and x is higher than 2. Since the precursors H
and L contain, compared to e.g. metakaolin, high amount of CaO, it was
taken into account as well; the prepared material can contain, besides
geopolymer, also calcium-silicate-hydrates (Li et al., 2010). The com-
position of amorphous portion was calculated as difference between
known amount of raw materials (Table 2) and results of quantitative
XRD phase analysis of prepared materials (Table 8). The portion of raw
materials, which is not present in minerals detected by XRD in H/L 1–3
materials, must be present in amorphous matter in H/L 1–3, i.e. in the
synthesized geopolymer. The stoichiometric coefficients of components
– fixed to Al2O3 to be 1 – according to above given formula were cal-
culated (Table 9). The content of Fe2O3, MgO, TiO2 and SO3 was not
taken into account. The highest compressive strength (Table 9) reached
sample L1; its stoichiometric coefficients indicated that favorable – with
respect to strength – is lower content of alkalis and SiO2 and higher
content of Al2O3.
The compressive strength of H samples was relatively independent
on the activator dosage, while the L samples were more sensitive, and
one could observe that a higher activator content reduced the com-
pressive strength. The flexural strength generally obeyed the same
Fig. 9. XRD patterns of geopolymers prepared from H precursor. (Mi - mica (muscovite, illite); Mc – microcline; A – albite; Q – quartz; D – diopside; Ak – akermanite;
H – hematite).
Applied Clay Science 161 (2018) 82–89
trend (higher dosage of activator – lower strength) but here the H
samples were more sensitive to materials' composition.
4. Discussion
The tested red-clay-ceramic powders differed highly from the con-
ventional precursors used for geopolymer synthesis, such as MK, FA,
and GGBS. The differences were found in both chemical and physical
properties. The most important difference lied in the phase composi-
tion; while conventional aluminosilicate precursors are dominantly
amorphous or glassy, the RCC contained a significant amount of crys-
talline minerals (quartz, feldspars, muscovite, etc.) which might be also
partially alkali-activated but most of their matter remains unaffected
and acts as filler in geopolymers (Zaharaki et al., 2016). The mor-
phology of RCC was closer to metakaolin (both RCC and MK are pro-
ducts of clay dehydroxylation) than to glassy (“high temperature”) slag
100 and fly ash. The similarity with MK was reflected in measurable BET
specific surface area indicating certain porosity of the RCC powders
(the FA and GGBS particles have glassy, non-porous surface). The
measured values (Table 5) were lower than those typical for MK
(Table 1) due to the presence of crystalline minerals in RCC, which
“dilute” the amorphous matter.
The obtained values of geopolymer strength do not allow us ar-
ticulating a simple conclusion, which kind of precursor is better and
why. The mechanical properties (usually compressive strength) of
geopolymers are frequently discussed with respect to SiO2/Al2O3,
Al2O3/Na2O and other ratios in the starting mixture of precursor and
activator (Reddy et al., 2016; Yun-Ming et al., 2016; Lahoti et al., 2017;
Lee et al., 2017). Because of the high content of crystalline minerals in
ceramic precursors H and L (Table 4), just its amorphous portion (and
liquid activator) were taken into account when the above given
“composition ratios” were calculated. The resulting values were used as
the independent variable for experimentally obtained compressive
strength (Fig. 11, open symbols). The solid symbols in the graph re-
present the composition of geopolymers calculated by help of XRD
(Table 9). Points of both series are mostly close to each other what
indicates the equivalency of both applied approaches to estimate the
geopolymer composition. The H and L data formed, roughly said, lines
which indicates the trends in the system. The SiO2/Al2O3 ratio in pre-
pared geopolymers varied between 2.8 and 3.6; the lower seemed to
provide higher strength. The favorable Al2O3/(Na2O + K2O) ratio was
found to be 1, what was in accordance with the theoretical Davidovits
geopolymer formula cited above.
Lahoti et al. (2017) analyzed statistically the factors influencing the
87
M. Keppert et al. Applied Clay Science 161 (2018) 82–89

Fig. 10. XRD patterns of geopolymers prepared from L precursor. (Mi - mica (muscovite, illite); Mc – microcline; A – albite; Q – quartz; D – diopside; Ak – akermanite;
H – hematite; C – calcite).

Table 8 40
Phase composition of prepared geopolymers.
H1 H2 H3 L1 L2 L3

Amorphous – 40.1 45.1 44.7 52.1 47.3 56.2 compressive strength (MPa) 35
portion
Quartz SiO2 17.1 16.4 17.0 16.1 15.9 14.3
Hematite Fe2O3 2.8 0.8 1.2 1.3 0.9 1.2
Albite NaAlSi3O8 11.7 11.2 11.0 6.2 7.1 5.8
Microcline KAlSi3O8 4.8 4.4 4.4 6.1 7.4 5.5 30
Muscovite KAl2(AlSi3O10)(OH)2 10.6 10.2 10.3 4.0 5.3 3.2
Illite K0.65Al2(Al0.65Si3.35O10) 3.7 3.3 2.5 1.7 2.4 1.7
(OH)2
Diopside CaMgSi2O6 4.8 4.6 5.0 2.2 2.7 2.2
Akermanite Ca2MgSi2O7 4.4 4.1 4.2 2.4 2.5 2.6 25
Mullite Al6Si2O13 – – – 2.1 3.3 2.4
Calcite CaCO3 – – – 5.7 4.5 4.9

20
Table 9 0 1 2 3 4
Composition (expressed as stoichiometric coefficients of components) of pre-
molar ratio
pared geopolymers and their strength. Best performing material is italicized.
Al2O3/(Na2O+K2O) SiO2/Al2O3
H1 H2 H3 L1 L2 L3
Fig. 11. Influence of molar ratios SiO2/Al2O3 and Al2O3/(Na2O + K2O) in
Stoichiometric coefficient Na2O 1.1 1.2 1.4 0.8 1.1 1.0 geopolymers on compressive strength. Solid symbols – values according to XRD
K2O 0.0 0.0 0.0 0.1 0.1 0.1 (Table 9); open symbols – values based on composition of precursor and acti-
CaO 0.6 0.6 0.5 0.9 1.3 0.9 vator.
Al2O3 1.0 1.0 1.0 1.0 1.0 1.0
SiO2 3.4 3.8 3.6 2.8 2.9 3.4
H2O 2.5 2.4 2.9 3.4 4.5 3.3 recommendations for MK-based systems and the data in Fig. 11 in-
Strength (MPa) Comp. 32.5 33.1 30.3 38.8 27.5 23.3 dicated that the proposed approach based on the quantification of just
Bending 9.5 7.6 4.1 9.4 9.5 6.9
the amorphous portion of ceramic precursors was correct.

strength of MK-based geopolymers and found Si/Al and Al/Na ratios to
be the most influential parameters; the optimum values were found to
be 1.7–2.2 (i.e. 3.4–4.4 when expressed as SiO2/Al2O3) and 0.8–1.2,
respectively. It should though be noted that the strengths obtained in
the MK-based geopolymer system (Lahoti et al., 2017) of optimum
composition were about double compared to the results obtained for
ceramic precursors in this paper. The attempt to rationalize the ex-
perimental data on RCC-based geopolymers presented in this paper
omitted: a) formation of calcium silicate hydrates which was likely, in
particular, in the materials based on L ceramic with a high CaO content
(Li et al., 2010) and b) likely partial dissolution (and activation) of
quartz and other minerals in the highly alkaline activator. Un-
fortunately, these phenomena were impossible to be quantified.
Nevertheless, the agreement between Lahoti et al. (2017)
5. Conclusions
Two red-clay ceramics (RCC) were activated by a varying amount of
Na2O/SiO2 activator with the silicate modulus of 1.4. The RCC pre-
cursors had the same granulometry but differed in chemical and phase
composition. The principal differences were the content of CaO and the
amorphous portion of precursor. Although the RCC differed from con-
ventional aluminosilicate precursors mainly in their higher crystal-
linity, even the low (< 30%) content of amorphous matter was suffi-
cient for geopolymer synthesis and both tested precursors enabled to
prepare geopolymers with satisfactory mechanical properties, which
were comparable to common cement pastes.
Due to their rheological behavior, the studied red-clay ceramics
pastes do not seem to be suitable as binding components in mortars for
adhesive applications where a high yield stress is needed or in

88
M. Keppert et al.
applications requiring fast flows, such as pumping or spraying, because
of their high viscosity. On the other hand, they could gain ground in
self-compacting or self-leveling applications where a low yield stress
allows the material to flow and spread under its own weight while the
high viscosity improves the stability of the coarse inclusions during
casting.
The obtained values of compressive strength were discussed with
respect to SiO2/Al2O3 and Al2O3/(Na2O + K2O) ratios in the geopo-
lymer mix. Since the ceramic precursors contained a significant amount
of crystalline minerals, the knowledge of both content and composition
of the amorphous portion was taken into account. The SiO2/Al2O3 and
Al2O3/(Na2O + K2O) ratios were calculated on the basis of the liquid
activator and the amorphous portion of the precursor as well as by help
of XRD analysis; the highest strength was obtained for their values of
2.8 and 1, respectively. This finding is in a good agreement with the
theoretical formula of geopolymer materials and with the results ob-
tained for metakaolin-based geopolymers by other investigators.
Therefore, the consideration of just amorphous portion of ceramics as
geopolymer forming matter is appropriate. The higher content of CaO
in the L precursor caused a broader pore size distribution and higher
porosity of resulting geopolymers due to the expectable formation of
calcium silicate hydrates.
Acknowledgement
The research was supported by the Czech Science Foundation, under
project No GA16-02862S.
References
Aboulayt, A., Riahi, M., Anis, S., Touhami, M.O., Moussa, R., 2014. Rheological behavior
of a fresh geopolymer based on metakaolin: effect of the introduction of calcium
carbonate. Int. J. Innov. App. Stud. 7, 1170–1177.
Banfill, P.F.G., 2003. The rheology of fresh cement and concrete - a review. In:
Proceedings of 11th International Cement Chemistry Congress, Durban, South Africa,
pp. 50–62.
Bergmann, J., Friedel, P., Kleeberg, R., 1998. BGMN – a new fundamental parameters
based Rietveld program for laboratory X-ray sources, its use in quantitative analysis
and structure investigations. In: CPD Newsl. vol. 20. Commission of Powder
Diffraction, International Union of Crystallography, pp. 5–8.
Bergmann, J., Monecke, T., Kleeberg, R., 2001. Alternative algorithm for the correction of
preferred orientation in Rietveld analysis. J. Appl. Crystallogr. 34, 16–19.
Belmokhtar, N., Ammari, M., Brigui, J., Ben allal, L., 2017. Comparison of the micro-
structure and the compressive strength of two geopolymers derived from metakaolin
and an industrial sludge. Constr. Build. Mater. 146 (621-269).
Borges, P.H.R., Bhutta, A., Bavuzo, L.T., Banthia, N., 2017. Effect of SiO2/Al2O3 molar
ratio on mechanical behavior and capillary sorption of MK-based alkali-activated
composites reinforced with PVA fibers. Mater. Struct. 50, 148.
Brus, J., Abbrent, S., Kobera, L., Urbanova, M., Cuba, P., 2016. Advances in Al-27 MAS
NMR studies of Geopolymers. Annu. Rep. NMR Spectrosc. 88, 79–147.
Cultrone, G., Sebastian, E., Elert, K., de la Torre, M.J., Cazalla, O., Rodriguez-Navarro, C.,
2004. Influence of mineralogy and firing temperature on the porosity of bricks. J.
Eur. Ceram. Soc. 24, 547–564.
Davidovits, J., 1991. Geopolymers – inorganic polymeric new materials. J. Therm. Anal.
37, 1633–1656.
Favier, A., Hot, J., Habert, G., Roussel, N., Lacaillerie, J., 2014. Flow properties of MK-
based geopolymer pastes. A comparative study with standard Portland cement pastes.
Soft Matter. 10, 1134–1141.
Fiala, L., Rovnaník, P., Černý, R., 2017. Investigation of the Joule's effect in electrically
enhanced alkali-activated aluminosilicates. Cement Wapno. Beton 22, 201–210.
Font, O., Moreno, N., Querol, X., Izquierdo, M., Alvarez, E., Diez, S., Elvira, J., Antenucci,
D., Nugteren, H., Plana, F., Lopez, A., Coca, P., Pena, F.G., 2010. X-ray powder dif-
fraction-based method for the determination of the glass content and mineralogy of
coal (co)-combustion fly ashes. Fuel 89, 2971–2976.
Hosan, A., Haque, S., Saikh, F., 2016. Compressive behaviour of sodium and potassium
activators synthetized fly ash geopolymer at elevated temperatures: a comparative
study. J. Build. Eng. 8, 123–130.
89
Applied Clay Science 161 (2018) 82–89
Hradil, D., Hradilová, J., Bezdička, P., Serendan, C., 2017. Late Gothic/early Renaissance
gilding technology and the traditional poliment material “Armenian bole”: Truly red
clay, or rather bauxite? Appl. Clay Sci. 135, 271–281.
Hu, N.J., Bernsmeier, D., Grathoff, G.H., Warr, L.N., 2017. The influence of alkali acti-
vator type, curing temperature and gibbsite on the geopolymerization of an inter-
stratified illite-smectite rich clay from Friedland. Appl. Clay Sci. 135, 386–393.
Hulan, T., Trnik, A., Kaljuvee, T., Uibu, M., Stubna, I., Kallavus, U., Traksmaa, R., 2017.
The study of firing of a ceramic body made from illite and fluidized bed combustion
fly ash. J. Therm. Anal. Calor. 127, 79–89.
ICSD (Inorganic Crystal Structure Database), https://icsd.fiz-karlsruhe.de/search/, (ac-
cessed December 2017)
Kovarik, T., Rieger, D., Kadlec, J., Krenek, T., Kullova, L., Pola, M., Belsky, P., France, P.,
Riha, J., 2017. Thermomechanical properties of particle-reinforced geopolymer
composite with various aggregate gradation of fine ceramic filler. Constr. Build.
Mater. 143, 599–606.
Lahoti, M., Narang, P., Tana, K.H., Yang, E.-H., 2017. Mix design factors and strength
prediction of metakaolin-based geopolymer. Ceram. Int. 43, 11433–11441.
Lee, B., Kim, G., Kim, R., Cho, B., Lee, S., Chon, C.-M., 2017. Strength development
properties of geopolymer paste and mortar with respect to amorphous Si/Al ratio of
fly ash. Constr. Build. Mater. 151, 512–519.
Lemougna, P.N., Wang, K., Tang, Q., Melo, U.C., Cui, X., 2016. Recent developments on
inorganic polymers synthesis and applications. Ceram. Int. 42, 15142–15159.
Li, C., Sun, H., Li, L., 2010. A review: the comparison between alkali-activated slag (Si
+Ca) and metakaolin (Si+Al) cements. Cem. Concr. Res. 40, 1341–1349.
Luna-Galiano, Y., Fernández-Pereira, C., Izquierdo, M., 2016. Contributions to the study
of porosity in fly ash-based geopolymers. Relationship between degree of reaction,
porosity and compressive strength. Mater. Constr. e098, 66.
MacKenzie, K.J.D., Brew, D.R.M., Fletcher, R.A., Vagana, A., 2007. Formation of alumi-
nosilicate geopolymers from 1:1 layer-lattice minerals pre-treated by various
methods: a comparative study. J. Mater. Sci. 42, 4667–4674.
Madsen, I.C., Scarlett, N.V.Y., 2008. Chapter 11 – quantitative phase analysis. In:
Dinnebier, R.E., Billinge, S.J.L. (Eds.), Powder Diffraction Theory and Practice. vol.
2008. RSC Publishing, Cambridge, pp. 298–331.
Mezger, T.G., 2014. The Rheology Handbook, 4th ed. Vincentz Network, Hannover,
Germany.
Mohammed, S., 2017. Processing, effect and reactivity assessment of artificial pozzolans
obtained from clays and clay wastes: a review. Constr. Build. Mater. 140, 10–19.
Pacheco-Torgal, F., Labrincha, J., Leonelli, C., Palomo, A., Chindaprasit, P. (Eds.), 2014.
Handbook of Alkali-Activated Cements, Mortars and Concretes, 1st ed. Woodhead,
Cambridge, pp. 777–806.
Pavlík, Z., Fiala, L., Jerman, M., Vejmelková, E., Pavlíková, M., Keppert, M., Černý, R.,
2014. Theoretical and experimental analysis of moisture-dependent thermal con-
ductivity of lightweight ceramic bricks. Int. J. Thermophys. 35, 1912–1921.
Peys, A., Rahier, H., Pontikes, Y., 2016. Potassium-rich biomass ashes as activators in
metakaolin-based inorganic polymers. Appl. Clay Sci. 119, 401–409.
Poole, A.B., Sims, I., 2016. Concrete Petrography: A Handbook of Investigative
Techniques, 2nd ed. CRC Press, Boca Raton.
Provis, J.L., 2017. Alkali-activated materials. Cem. Concr. Res. http://dx.doi.org/10.
1016/j.cemconres.2017.02.009.
Rashad, A.M., 2013. Metakaolin as cementitious material: history, scours, production and
composition – a comprehensive overview. Constr. Build. Mater. 41, 303–318.
Reddy, M.S., Dinakar, P., Rao, B.H., 2016. A review of the influence of source material's
oxide composition on the compressive strength of geopolymer concrete. Microporous
Mesoporous Mater. 234, 12–23.
Sabeur, H., Platret, G., Vincent, J., 2016. Composition and microstructural changes in an
aged cement pastes upon two heating–cooling regimes, as studied by thermal analysis
and X-ray diffraction. J. Therm. Anal. Calor. 126, 1023–1043.
Sobott, R., Bente, K., Kittel, M., 2014. Comparative porosity measurements on ceramic
materials. Old Potters Almanack 21, 18–25.
Topçu, I.B., Toprak, M.U., Uygunoglu, T., 2014. Durability and microstructure char-
acteristics of alkali activated coal bottom ash geopolymer cement. J. Clean. Prod. 81,
211–217.
Ufer, K., Roth, G., Kleeberg, R., Stanjek, H., Dohrmann, R., Bergmann, J., 2004.
Description of X-ray powder pattern of turbostratically disordered layer structures
with a Rietveld compatible approach. Z. Kristallogr. 219, 519–527.
Ufer, K., Kleeberg, R., Bergmann, J., Curtius, H., Dohrmann, R., 2008. Refining real
structure parameters of disordered layer structures within the Rietveld method. Z.
Kristallogr. Suppl. (27), 151–158.
Williams, R.P., van Riessen, A., 2010. Determination of the reactive component of fly
ashes for geopolymer production using XRF and XRD. Fuel 89, 3683–3692.
Yun-Ming, L., Cheng-Yong, H., Al Bakri, M.M., Hussin, K., 2016. Structure and properties
of clay-based geopolymer cements: a review. Prog. Mater. Sci. 83, 595–629.
Zaharaki, D., Galetakis, M., Komnitsas, K., 2016. Valorization of construction and de-
molition (C&D) and industrial wastes through alkali activation. Constr. Build. Mater.
121, 686–693.