Hydratable Alumina-Bonded Suspensions: Evolution of Microstructure and Physical Properties During First Heating

Abstract

Hydratable alumina (HA) is a calcium-free and high-refractoriness binder for alumina-based suspensions. Although recent studies have improved its dispersion, mixing and drying behaviours, a drawback related to its loss of strength between 250 and 900 °C remains unexplored. Pores generated after decomposition of HA curing products are usually an explanation for the effect; however, no experimental result has supported this hypothesis so far. This study investigated the effects of thermal treatment (120−1500 °C) upon the microstructure and physical properties of calcined alumina suspensions containing different amounts of HA (10−40 vol.-%). Porosity, compression strength and flexural elastic modulus measurements, thermal linear variation and thermogravimetric analysis were compared with scanning electron microscopy and X-ray diffraction results. The average matrix particle size and amount of HA in the formulation play major roles in the types of curing products that are formed. The strength reduction observed during first heating was not directly associated with the increase in porosity.

Full text

024 2017
R. Salomão*, M. A. Kawamura, A.D.V. Souza, J. Sakihama
Hydratable Alumina-Bonded Suspensions: Evolution of Microstructure
and Physical Properties During First Heating
ABSTRACT
AUTHOR
Hydratable alumina (HA) is a calcium-free and high-refractoriness binder
for alumina-based suspensions. Although recent studies have improved
Dr. Rafael Salomão received his Ph.D. in Materials Sci-
ence and Engineering from the Federal University of São
Carlos (Brazil, 2005). Since 2010 he has been associate
professor and researcher in the Materials Engineering
Department (SMM) of São Carlos School of Engineering
(EESC), at the University of São Paulo (USP, Brazil). The
main research lines of Prof. Salomão’s Group (Integrated Solutions in
Manufacture and Ceramic Materials, SIMMaC) are advanced refractory
ceramics, microporous ceramics for thermal insulation and ceramic-poly-
mer composites. He has authored/co-authored over 80 technical papers.
E-Mail: rsalomao@sc.usp.br
KEYWORDS: hydratable alumina, aqueous castable suspensions, microstructure evolution, porosity
 Refractories Manual (2017)
1 of 15
Hydratable Alumina-Bonded Suspensions: Evolution of
Microstructure and Physical Properties During First Heating
R. Salomão*, M. A. Kawamura, A.D.V. Souza, J. Sakihama
University of São Paulo, Materials Engineering Depart.,
São Carlos School of Engineering,
Avenida Trabalhador São-carlense
400 São Carlos, SP - Brazil
* Corresponding author, Salomão, R.: rsalomao@sc.usp.br
Dr. Rafael Salomão received his Ph.D. in Materials Science and Engineering from the Federal
University of São Carlos (Brazil, 2005). Since 2010 he has been associate professor and
researcher in the Materials Engineering Department (SMM) of São Carlos School of Engineering
(EESC), at the University of São Paulo (USP, Brazil). The main research lines of Prof. Salomão’s
Group (Integrated Solutions in Manufacture and Ceramic Materials, SIMMaC) are advanced
refractory ceramics, microporous ceramics for thermal insulation and ceramic-polymer composites.
He has authored/co-authored over 80 technical papers. E-Mail: rsalomao@sc.usp.br
Abstract
Hydratable alumina (HA) is a calcium-free and high-refractoriness binder for alumina-based
suspensions. Although recent studies have improved its dispersion, mixing and drying behaviours,
a drawback related to its loss of strength between 250 and 900 °C remains unexplored. Pores
generated after decomposition of HA curing products are usually an explanation for the effect;
however, no experimental result has supported this hypothesis so far. This study investigated the
effects of thermal treatment (120−1500 °C) upon the microstructure and physical properties of
calcined alumina suspensions containing different amounts of HA (10−40 vol.-%). Porosity,
compression strength and flexural elastic modulus measurements, thermal linear variation and
thermogravimetric analysis were compared with scanning electron microscopy and X-ray diffraction
results. The average matrix particle size and amount of HA in the formulation play major roles in
the types of curing products that are formed. The strength reduction observed during first heating
was not directly associated with the increase in porosity.
1. Introduction
1.1 Transition aluminas and rho-alumina (ρ-Al2O3) production
Transition aluminas are obtained through partial calcination of alumin-
ium hydroxide (Al(OH)3) or boehmite (β-AlOOH) and are intermediate
compounds between those precursors and alumina’s most stable phase
(α-Al2O3) (Fig. 1) [1−4]. Depending on the type of Al(OH)3 precursor
employed, its particles’ characteristics (mainly size distribution, surface
area and morphology) and process variables (heating rate, maximum
temperature achieved and atmosphere), dierent transition aluminas can
be produced [5−12]. These materials show an interesting combination of
useful properties in many technological elds. A particular type of transi-
tion alumina known as ρ-Al2O3 (Al2O3nH2O, 0,2 ≤ n ≤ 1) or hydratable
alumina (HA) is used as a hydraulic binder to improve the refractoriness
and corrosion resistance of refractory castables and other types of ceramic
materials shaped by aqueous processing [1, 2,13−19].
In the production of HA, coarse Al(OH)3 aggregates (gibbsite, α-Al(OH)3,
or bayerite, β-Al(OH)3) larger than 100 μm are roughly ground to ap-
proximately 10−50 μm [2, 5, 6, 13]. This initial step is important for ad-
justment of the particle size distribution in the nal product. The milled
particles partially preserve the hexagonal plate-like shape of the Al(OH)3
crystals obtained by chemical precipitation. However, they also exhib-
it some irregular surfaces produced during comminution (Figs. 2a−b)
[20]. The particles are then ash-calcined in an ascending ow of hot
air (typical temperatures range from 700−1200 °C). High heating rates
(greater than 150 Ks-1) establish a micro-hydrothermal condition inside
each Al(OH)3 particle prior to the start of dehydroxilation [8, 12, 21, 22].
University of São Paulo, Materials Engineering Depart., São Carlos School of Engineering,
Avenida Trabalhador São -carlense, 400 São Carlos, SP - Brazil
* Corresponding author, Salomão, R.: rsalomao@sc.usp.br
its dispersion, mixing and drying behaviours, a drawback related to its
loss of strength between 250 and 900 °C remains unexplored. Pores
generated after decomposition of HA curing products are usually an ex-
planation for the eect; however, no experimental result has supported
this hypothesis so far. This study investigated the eects of thermal treat-
ment (120−1500 °C) upon the microstructure and physical properties of
calcined alumina suspensions containing dierent amounts of HA (10−
40 vol.-%). Porosity, compression strength and exural elastic modulus
measurements, thermal linear variation and thermogravimetric analysis
were compared with scanning electron microscopy and X-ray diraction
results. The average matrix particle size and amount of HA in the formula-
tion play major roles in the types of curing products that are formed. The
strength reduction observed during rst heating was not directly associ-
ated with the increase in porosity.
Withdrawal of a large amount of pressurized water vapour generates
micro-cracks all over the particles, which increases their specic surface
area (from approximately 7 up to 80−150 m2g-1, Table 1), solid density
(from 2,4 to 2,8 gcm-3) and pore volume content (from less than 0.001
to 2.15 cm3g-1). The resulting particles show a hybrid aspect, mixing
elements of the original shape of the Al(OH)3 crystals with brous and
layered patterns originated during their volumetric shrinkage (Figs. 2c−f)
[13]. Similar physical changes in Al(OH)3 particles can also be observed in
conventional calcination processes for the production of α-Al2O3 [4, 7, 11,
12]. However, because much lower heating rates are employed (0,05−0,5
Kmin-1), dierent crystalline phases are obtained. Whereas conventional
dehydroxilation of Al(OH)3 produces a sequence of crystalline transition
aluminas, its ash-calcination results in a defect-rich structure comprised
of amorphous and meta-stable ρ-Al2O3 (Fig. 1).
1.2 Rho-alumina setting mechanism
Unlike other phases, ρ-Al2O3 shows signicant solubility in water and
re-hydroxylates to form compounds such as boehmite (β-AlOOH) and
bayerite (β-Al(OH)3) [5, 6, 13−17]. Due to this unique ability, HA is add-

025REVIEW PAPERS
Fig.1 • Phase changes that occur during thermal treatment of different types of Al(OH)3
1
Table 1 · Characteristics of the tested raw materials.
Physico-chemical
properties
Raw materials
Calcined alumina c Hydratable alumina (HA)
a Fine (FA) b Coarse (CA)
d Composition / mass-% α-Al2O3: 99.8; Na2O: 0,12; Fe2O3:
0.01; SiO2: 0.04; CaO: 0.03
α-Al2O3: 99.7; Na2O: 0.2; Fe2O3:
0.03; SiO2: 0.34; CaO: 0.03
α-Al2O3: 99.4; Na2O: 0.05; Fe2O3:
0.02; MgO: 0.07; SiO2: 0.03; CaO: 0.02
e Particle size
/ D50 / D90 / μm and
f morphology
0.5 / 2.3 1.5 / 9.6 7.6 / 31
g Solid density ρ / gcm-3 3.99 3.97 2.72
h Specific surface area / m2g-1 8.2 1.3 95
h Pore content / cm3g-1 <1 <1 215
i LOI / mass-%, 1000 ºC <0.1 <0.1 8.2
15 of 15
Table 1: Characteristics of the tested raw materials.
Physico-
chemical
properties
Raw materials
Calcined alumina c Hydratable alumina (HA)
a
F
ine (FA) b Coarse (CA)
dComposition /
mass-%
-Al2O3: 99.8; Na2O: 0,12;
Fe2O3: 0.01; SiO2: 0.04; CaO:
0.03
-Al2O3: 99.7; Na2O: 0.2;
Fe2O3: 0.03; SiO2: 0.34; CaO:
0.03
-Al2O3: 99.4; Na2O: 0.05;
Fe2O3: 0.02; MgO: 0.07; SiO2:
0.03; CaO: 0.02
e Particle size
/ D50 / D90 /

m and
f morphology
0.5 / 2.3
3 m3 m
1.5 / 9.6
3 m3 m
7.6 / 31
gSolid density

/ g∙cm-3 3.99 3.97 2.72
h Specific
surface area /
m2∙g-1
8.2 1.3 95
hPore content
/ cm3∙g-1 <1 <1 215
iLOI / mass-
%, 1000 º
C
<0.1 <0.1 8.2
a) A1000 SG (Almatis, USA). b) E-SY 1000 (Almatis, USA). c) Alphabond 300 (Almatis, USA). d) X-ray dispersive spectroscopy (EDX 720,
Shimadzu, Japan. After calcination at 1000 °C for 5 h). e) DT-1202 (Dispersion Technology Inc., USA). f) FEG-SEM (Inspect F50, FEI,
Netherlands). g) Helium pycnometer method (Ultrapyc 1200e, Quantachrome Instruments, USA). h) N2 adsorption (Nova 1200e, Quantachrome
Instruments, USA, ASTM C 1069-09 standard). i) Thermogravimetry (TGA-Q50, TA Instruments, 25-1000 ºC, 5 K∙min-1 heating rate and synthetic
air atmosphere).
Table 2: Tested compositions
Raw materials Sample identification
90FA_10HA 80FA_20HA 60FA_40HA 90CA_10HA 80CA_20HA 60CA_40HA
50 vol.-% solids /
vol.-% / mass-%
Hydratable
alumina (HA) 5.00 / 5.59 10.00 /11.49 20.00 / 24.27 5.00 / 5.61 10.00 / 11.53 20.00 / 24.34
Fine calcined
alumina (FA) 45.00 / 73.84 40.00 / 67.40 30.00 / 53.41 0.00 / 0.00 0.00 / 0.00 0.00 / 0.00
Coarse
calcined
alumina (CA)
0.00 / 0.00 0.00 / 0.00 0.00 / 0.00 45.00 / 73.75 40.00 / 67.29 30.00 / 53.59
Distilled water
/ vol.-% / mass-% 50.00 / 20.56 50.00 / 21.11 50.00 / 22.31 50.00 / 20.64 50.00 / 21.19 50.00 / 22.37
a Dispersant
/ mass-%. dry basis 0.16 0.19 0.22 0.04 0.14 0.19
HA content after
drying / vol.-% / mass-
%
10.0 / 7.0 20.0 / 14.5 40.0 / 31.2 10.0 / 4.8 20.0 / 10.2 40.0 / 23.3
a) FS20 Cas
t
ament. BASF. Germany
15 of 15
Table 1: Characteristics of the tested raw materials.
Physico-
chemical
properties
Raw materials
Calcined alumina c Hydratable alumina (HA)
a
F
ine (FA) b Coarse (CA)
dComposition /
mass-%
-Al2O3: 99.8; Na2O: 0,12;
Fe2O3: 0.01; SiO2: 0.04; CaO:
0.03
-Al2O3: 99.7; Na2O: 0.2;
Fe2O3: 0.03; SiO2: 0.34; CaO:
0.03
-Al2O3: 99.4; Na2O: 0.05;
Fe2O3: 0.02; MgO: 0.07; SiO2:
0.03; CaO: 0.02
e Particle size
/ D50 / D90 /

m and
f morphology
0.5 / 2.3
3 m3 m
1.5 / 9.6
3 m3 m
7.6 / 31
gSolid density

/ g∙cm-3 3.99 3.97 2.72
h Specific
surface area /
m2∙g-1
8.2 1.3 95
hPore content
/ cm3∙g-1 <1 <1 215
iLOI / mass-
%, 1000 º
C
<0.1 <0.1 8.2
a) A1000 SG (Almatis, USA). b) E-SY 1000 (Almatis, USA). c) Alphabond 300 (Almatis, USA). d) X-ray dispersive spectroscopy (EDX 720,
Shimadzu, Japan. After calcination at 1000 °C for 5 h). e) DT-1202 (Dispersion Technology Inc., USA). f) FEG-SEM (Inspect F50, FEI,
Netherlands). g) Helium pycnometer method (Ultrapyc 1200e, Quantachrome Instruments, USA). h) N2 adsorption (Nova 1200e, Quantachrome
Instruments, USA, ASTM C 1069-09 standard). i) Thermogravimetry (TGA-Q50, TA Instruments, 25-1000 ºC, 5 K∙min-1 heating rate and synthetic
air atmosphere).
Table 2: Tested compositions
Raw materials Sample identification
90FA_10HA 80FA_20HA 60FA_40HA 90CA_10HA 80CA_20HA 60CA_40HA
50 vol.-% solids /
vol.-% / mass-%
Hydratable
alumina (HA) 5.00 / 5.59 10.00 /11.49 20.00 / 24.27 5.00 / 5.61 10.00 / 11.53 20.00 / 24.34
Fine calcined
alumina (FA) 45.00 / 73.84 40.00 / 67.40 30.00 / 53.41 0.00 / 0.00 0.00 / 0.00 0.00 / 0.00
Coarse
calcined
alumina (CA)
0.00 / 0.00 0.00 / 0.00 0.00 / 0.00 45.00 / 73.75 40.00 / 67.29 30.00 / 53.59
Distilled water
/ vol.-% / mass-% 50.00 / 20.56 50.00 / 21.11 50.00 / 22.31 50.00 / 20.64 50.00 / 21.19 50.00 / 22.37
a Dispersant
/ mass-%. dry basis 0.16 0.19 0.22 0.04 0.14 0.19
HA content after
drying / vol.-% / mass-
%
10.0 / 7.0 20.0 / 14.5 40.0 / 31.2 10.0 / 4.8 20.0 / 10.2 40.0 / 23.3
a) FS20 Cas
t
ament. BASF. Germany
15 of 15
Table 1: Characteristics of the tested raw materials.
Physico-
chemical
properties
Raw materials
Calcined alumina c Hydratable alumina (HA)
a
F
ine (FA) b Coarse (CA)
dComposition /
mass-%
-Al2O3: 99.8; Na2O: 0,12;
Fe2O3: 0.01; SiO2: 0.04; CaO:
0.03
-Al2O3: 99.7; Na2O: 0.2;
Fe2O3: 0.03; SiO2: 0.34; CaO:
0.03
-Al2O3: 99.4; Na2O: 0.05;
Fe2O3: 0.02; MgO: 0.07; SiO2:
0.03; CaO: 0.02
e Particle size
/ D50 / D90 /

m and
f morphology
0.5 / 2.3
3 m3 m
1.5 / 9.6
3 m3 m
7.6 / 31
gSolid density

/ g∙cm-3 3.99 3.97 2.72
h Specific
surface area /
m2∙g-1
8.2 1.3 95
hPore content
/ cm3∙g-1 <1 <1 215
iLOI / mass-
%, 1000 º
C
<0.1 <0.1 8.2
a) A1000 SG (Almatis, USA). b) E-SY 1000 (Almatis, USA). c) Alphabond 300 (Almatis, USA). d) X-ray dispersive spectroscopy (EDX 720,
Shimadzu, Japan. After calcination at 1000 °C for 5 h). e) DT-1202 (Dispersion Technology Inc., USA). f) FEG-SEM (Inspect F50, FEI,
Netherlands). g) Helium pycnometer method (Ultrapyc 1200e, Quantachrome Instruments, USA). h) N2 adsorption (Nova 1200e, Quantachrome
Instruments, USA, ASTM C 1069-09 standard). i) Thermogravimetry (TGA-Q50, TA Instruments, 25-1000 ºC, 5 K∙min-1 heating rate and synthetic
air atmosphere).
Table 2: Tested compositions
Raw materials Sample identification
90FA_10HA 80FA_20HA 60FA_40HA 90CA_10HA 80CA_20HA 60CA_40HA
50 vol.-% solids /
vol.-% / mass-%
Hydratable
alumina (HA) 5.00 / 5.59 10.00 /11.49 20.00 / 24.27 5.00 / 5.61 10.00 / 11.53 20.00 / 24.34
Fine calcined
alumina (FA) 45.00 / 73.84 40.00 / 67.40 30.00 / 53.41 0.00 / 0.00 0.00 / 0.00 0.00 / 0.00
Coarse
calcined
alumina (CA)
0.00 / 0.00 0.00 / 0.00 0.00 / 0.00 45.00 / 73.75 40.00 / 67.29 30.00 / 53.59
Distilled water
/ vol.-% / mass-% 50.00 / 20.56 50.00 / 21.11 50.00 / 22.31 50.00 / 20.64 50.00 / 21.19 50.00 / 22.37
a Dispersant
/ mass-%. dry basis 0.16 0.19 0.22 0.04 0.14 0.19
HA content after
drying / vol.-% / mass-
%
10.0 / 7.0 20.0 / 14.5 40.0 / 31.2 10.0 / 4.8 20.0 / 10.2 40.0 / 23.3
a) FS20 Cas
t
ament. BASF. Germany
ed to castable formulations in 5−40 vol.-% amounts. During the mixing
and curing steps of cast parts, it reacts with water and assures enough
strength for demoulding and drying [13, 18, 19, 23−25]. The initial step
of the setting mechanism is wetting of particle surfaces [13− 15, 26, 27].
Because of its high specic surface area and large mesopore content,
HA-containing formulations usually require greater water amounts and
longer and more intensive mixing processes than other hydraulic binders
[18, 24]. The rst contact between HA and liquid water is followed by a
signicant heat release. An analogous eect can be observed during the
wetting of other high specic surface area powders. When air adsorbed at
a particle’s surface and trapped inside its mesopores is replaced by water,
large and thermodynamically more stable gas bubbles are formed. Excess
free energy is then released from the system as evolved heat [26, 27].
A later release of heat occurs when the particles at the wet surfaces be-
gin to rehydroxylate and dissolve (Equation 1):
(1) Al2O3.nH2O (0,2 ≤ n ≤ 1) + H2O → Al(OH)4- + H2O
(2) Al(OH)4- + H2O → β-AlOOH
After a short period of induction, a high concentration of Al(OH)4- ions
saturates the water around the solid particles, which causes precipitation
of pseudo-boehmite (Equation 2) as a bulky coat of amorphous gel at their
surfaces.
This viscous gel permeates the spaces amongst the particle suspen-
sion, hampers particle movement and sets the system [13−15]. Previous
a) A1000 SG (Almatis, USA). b) E-SY 1000 (A lmatis, USA). c) Alphab ond 300 (Almatis, US A). d) X-ray dispersi ve spectro scopy (EDX 720, Shimad zu, Japan. Afte r calcinatio n at 1000 °C for 5 h). e) DT-1202 (Dispersion
Technology I nc., USA). f) FEG-SEM (In spect F50, FEI, Net herlands). g) Helium pyc nometer metho d (Ultrapyc 1200e, Qu antachrome Ins truments, US A). h) N2 adsorption (Nova 1200e, Quantachrome Instruments,
USA, ASTM C 1069 -09 standar d). i) Thermogravim etry (TGA-Q50, TA Ins truments, 25-1000 º C, 5 Kmin-1 heating rate and synthetic air atmosphere).

(a)
(c)
(e)
(b)
(d)
(f)
026 2017
Fig.2 • a−b) Al(OH)3 precursor of hydratable alumina, c) tested hydratable alu-
mina particles, (d−f) larger magnification of the region highlighted in c)
Fig.3 • Schematic view of the thermogravimetric equipment employed to assess
the dehydroxylation behaviour of green samples
Fig.4 • XRD patterns for the compositions containing 60 vol.-% of calcined alumi-
na (coarse or fine) and 40 vol.-% of hydratable alumina after thermal treatment
at different temperatures. Key to symbols: Ο = Corindon (α-Al2O3, JCPDS file 46-
1212),  = Boehmite (β-AlOOH, JCPDS file 76-1871)
2 3
4
studies have shown that the extent and intensity of both events can be
strongly aected by the mixing temperature [15−17], curing atmosphere
[18,19] and presence of other raw materials and additives [24, 26−31].
1.3 First heating of HA-bonded castables
Because of the technological importance of HA-bonded suspensions,
many recent studies have aimed at improving their dispersion, mixing
and rheological behaviours [18, 26, 28, 31].They have also attempted to
understand how the HA curing process aects other raw materials [24, 26,
27, 29, 30] and increases the safety of rst heat-up [18, 24, 31−35]. How-
ever, a drawback related to signicant loss of strength and rigidity that
occurs after drying (at approximately 250 °C) and before the early stages
of sintering (at 900−1000 °C) remains unexplored [12, 13, 30, 36−39].
During early thermal treatment, large volume components (weighing
from hundreds of kilograms to several tons) can fail under compression
from their own weight. Such an eect can be dangerous, particularly for
thermal insulating linings, which are intrinsically low-strength materials
due to their high levels of porosity. There is consensus in the literature that
this strength reduction is related to the pores generated after decompo-
sition of HA-hydroxylated curing products [23, 25, 36, 37, 40]. However,
to the best of our knowledge, no experimental result or observation by
high-resolution microscopy has supported this hypothesis.
The present study investigated the eects of thermal treatment
(120−1500 °C) on the microstructure and physical properties of calcined
alumina (ne or coarse grades) castable suspensions containing varied
amounts of hydratable alumina (10−40 vol.-%). Measurements of physi-
cal properties (total porosity, compression strength, exural elastic mod-
ulus, and thermal linear variation) and thermogravimetric analysis were
compared with comprehensive scanning electron microscopy investiga-
tion and X-ray diraction analysis.

027REVIEW PAPERS
Fig.5 • Thermogravimetric analysis of humid green samples (after 24 hrs of cu-
ring): a) mass loss W, b−c) mass loss rate dW/dt
52. Materials and Methods
Calcined alumina (FA: ne or CA: coarse), hydratable alumina (HA) and
dispersant (Table 1) were dry-mixed (Table 2) and added to distilled water
(50 vol.-% of solids) in a paddler mixer (PowerVisc, IKA, Germany) oper-
ating at 1000 rpm for 5 minutes. The homogenized suspensions were cast
under vibration as cylinders (70 mm length x 16 mm diameter, 16 mm
length x 16 mm diameter, 38 mm length x 38 mm diameter, and 8 mm
length x 6 mm diameter) and kept for 24 hrs at 60 ± 2 °C in closed asks
having humidity close to 100 %. The samples were then transferred to a
ventilated atmosphere at 60 ± 2 °C for 24 h followed by another 24 h at
120 °C. These curing and drying conditions maximised the binding eect
of hydratable alumina and reduced the likelihood of explosive spalling
during rst heat-up [18, 24, 32, 35].
A contact dilatometer (DIL 402 C, Netzsch, Germany) assessed the linear
thermal variation of the dried green compositions (8 x 6 mm samples)
during the rst heating up to 1500 °C (5 Kmin-1 heating rate).
Thermogravimetric equipment developed by the authors’ research
group (Fig. 3) evaluated the drying and dehydroxylation behaviour of hu-
mid green samples. It contained an electronic mass recorder connected
to a sample placed inside an electric furnace and a register that contin-
uously monitors variations of mass and temperature inside the chamber
of the furnace and at the centre of the samples. In comparison to typical
instruments of thermogravimetric analysis (TGA), the equipment has two
important novelties: a) ability to use large samples of up to 500 g and size
10 cm x 10 cm x 10 cm, and b) the possibility of testing both monolithic
and powdered samples (analytical TGA usually has a limitation of approx-
imately 1 g of powdered material). Several recent studies have employed
similar equipment for analysis of the drying behaviour of refractory
castables [18, 24, 29, 31, 35, 36, 40].
During the thermogravimetric tests, humid green samples (cylinders
of 38 mm x 38 mm and approximately 200 g) were heated up to 120 °C
(0,5 Kmin-1) and kept at that temperature for 17 hrs. Afterwards, the
temperature was increased to 800 °C (2 Km-1). During thermal treatment,
a computer recorded the samples’ mass and surface temperature every
10 s (the curves in Fig 5 are based on a sequence of experimental points).
This choice of thermal program highlighted the amount of water released
in each drying stage [29, 35, 36, 41]. Up to 120 °C, the mass loss is re-
Table 2 · Tested compositions
Hydratable
alumina (HA) 5.00 / 5.59 10.00 /11.49 20.00 / 24.27 5.00 / 5.61 10.00 / 11.53 20.00 / 24.34
Fine calcined
alumina (FA) 45.00 / 73.84 40.00 / 67.40 30.00 / 53.41 0.00 / 0.00 0.00 / 0.00 0.00 / 0.00
Coarse calcined
alumina (CA) 0.00 / 0.00 0.00 / 0.00 0.00 / 0.00 45.00 / 73.75 40.00 / 67.29 30.00 / 53.59
Distilled water /
vol.-% / mass-% 50.00 / 20.56 50.00 / 21.11 50.00 / 22.31 50.00 / 20.64 50.00 / 21.19 50.00 / 22.37
a Dispersant /
mass-%. dry basis 0.16 0.19 0.22 0.04 0.14 0.19
HA content after drying /
vol.-% / mass-% 10.0 / 7.0 20.0 / 14.5 40.0 / 31.2 10.0 / 4.8 20.0 / 10.2 40.0 / 23.3
Raw materials Sample identication
90FA_10HA 80FA_20HA 60FA_40HA 90CA_10HA 80CA_20HA 60CA_40HA
a) FS20 Castam ent. BASF. Germany
50 vol.-% solids /
vol.-% / mass-%

028 2017
lated mostly to evaporative withdrawal of non-chemically bonded water
(or free water). This rst stage of heating provides useful information on
the drying behaviour of the compositions (for instance, how long samples
must be exposed to heat for the removal of most of the free water and
reduction of the likelihood of explosive spalling). From 120 to 800 °C, the
remaining water leaves the system by ebullition (as pressurized water va-
pour) and decomposition of hydroxylated compounds.
Mass loss (W, mass-%, ranging from 0 to 100 %) and mass loss rate
(dW/dt, mass-%h-1), for a particular time ti, was calculated by Equations
3 and 4, respectively.
where Mi is the instantaneous mass recorded at time ti during the heat-
ing of the samples, M0 is the initial mass and MF is the nal one (the i+10
and i-10 index refer to points collected 10 s after and before the analyzed
time i, respectively). These calculated parameters are particularly useful
for comparisons amongst formulations containing the same total water
amount. Other studies have described equivalent methods and phenom-
ena for similar systems [18, 28, 29, 31, 34, 35]. The reactions occurring
between HA and water during curing were studied by means of mass loss
and mass loss rate versus temperature curves. Information on the kinet-
ics and products formed was related to other physical properties, such as
mechanical strength and porosity.
Green and red cast samples of 70 mm x 16 mm were measured (L:
length in cm and D: diameter in cm), weighed (M in g) and then ground
(DParticle < 100 μm) and characterised by helium pycnometry (solid den-
sity, ρSolid in gcm-3). Their geometric total porosity (TP in %) was calcu-
lated using Equation 5:
The ground samples were investigated by X-ray diraction (Rotaex RV
200B, Rigaku-Denki Corp., Japan; CuKα radiation, 10 ° to 70 ° --range,
2 °min-1 scan rate) for identication of phases.
The Young’s modulus (E, GPa) of green and red samples was measured
by the impulse excitation of vibration technique (Sonelastic, ATCP, Brazil)
according to the ASTME 1876-01 standard ("Standard Test Method for Dy-
namic Young’s Modulus, Shear Modulus, and Poisson’s Ratio by Impulse
Excitation of Vibration"). Samples were then sliced into cylinders of equal
length and diameter and tested for mechanical strength under diametric
compression (adapted from the ASTM C133-97 (2008) standard, "Stan-
dard Test Methods for Cold Crushing Strength and Modulus of Rupture of
Refractories") in an MTS 810 TestStar II tensile tester, at a 2 Ns-1 loading
rate. Freshly demoulded humid green samples were also sliced and tested
for diametric compression strength (total porosity and Young’s modulus
measurements could not be taken for these samples). Five dierent in-
stances of each type of sample were tested to determine porosity, Young’s
modulus, and compression strength.
Fractured cross sections of the samples were examined by eld emission
scanning electron microscopy (FEG-SEM, Inspect F50, FEI, Netherlands).
(5)
5 of 15
computer recorded the samples’ mass and surface temperature every 10 s (the curves in Fig 5 are
based on a sequence of experimental points). This choice of thermal program highlighted the
amount of water released in each drying stage [29, 35, 36, 41]. Up to 120 °C, the mass loss is
related mostly to evaporative withdrawal of non-chemically bonded water (or free water). This first
stage of heating provides useful information on the drying behaviour of the compositions (for
instance, how long samples must be exposed to heat for the removal of most of the free water and
reduction of the likelihood of explosive spalling). From 120 to 800 °C, the remaining water leaves
the system by ebullition (as pressurized water vapour) and decomposition of hydroxylated
compounds.
Mass loss (W, mass-%, ranging from 0 to 100 %) and mass loss rate (dW/dt, mass-%∙h-1), for a
particular time ti, was calculated by Equations 3 and 4, respectively.


F0
i0
iMM
MM
%100W 

 (3)



10i10i
10i10i
itt
WW
dt
dW

 

 (4)
where Mi is the instantaneous mass recorded at time ti during the heating of the samples, M0 is the
initial mass and MF is the final one (the i+10 and i-10 index refer to points collected 10 s after and
before the analyzed time i, respectively). These calculated parameters are particularly useful for
comparisons amongst formulations containing the same total water amount. Other studies have
described equivalent methods and phenomena for similar systems [18, 28, 29, 31, 34, 35]. The
reactions occurring between HA and water during curing were studied by means of mass loss and
mass loss rate versus temperature curves. Information on the kinetics and products formed was
related to other physical properties, such as mechanical strength and porosity.
Green and fired cast samples of 70 mm  16 mm were measured (L: length in cm and D: diameter
in cm), weighed (M in g) and then ground (DParticle < 100 m) and characterised by helium
pycnometry (solid density, Solid in g∙cm-3). Their geometric total porosity (TP in %) was calculated
using Equation 5:













Solid
2LD
M4
1%100TP
(5)
The ground samples were investigated by X-ray diffraction (Rotaflex RV 200B, Rigaku-Denki
Corp., Japan; CuKα radiation, 10 ° to 70 ° --range, 2 °∙min-1 scan rate) for identification of phases.
The Young’s modulus (E, GPa) of green and fired samples was measured by the impulse
excitation of vibration technique (Sonelastic, ATCP, Brazil) according to the ASTME 1876-01
(3)
(4)
5 of 15
computer recorded the samples’ mass and surface temperature every 10 s (the curves in Fig 5 are
based on a sequence of experimental points). This choice of thermal program highlighted the
amount of water released in each drying stage [29, 35, 36, 41]. Up to 120 °C, the mass loss is
related mostly to evaporative withdrawal of non-chemically bonded water (or free water). This first
stage of heating provides useful information on the drying behaviour of the compositions (for
instance, how long samples must be exposed to heat for the removal of most of the free water and
reduction of the likelihood of explosive spalling). From 120 to 800 °C, the remaining water leaves
the system by ebullition (as pressurized water vapour) and decomposition of hydroxylated
compounds.
Mass loss (W, mass-%, ranging from 0 to 100 %) and mass loss rate (dW/dt, mass-%∙h-1), for a
particular time ti, was calculated by Equations 3 and 4, respectively.


F0
i0
iMM
MM
%100W 


(3)



10i10i
10i10i
itt
WW
dt
dW

 

 (4)
where Mi is the instantaneous mass recorded at time ti during the heating of the samples, M0 is the
initial mass and MF is the final one (the i+10 and i-10 index refer to points collected 10 s after and
before the analyzed time i, respectively). These calculated parameters are particularly useful for
comparisons amongst formulations containing the same total water amount. Other studies have
described equivalent methods and phenomena for similar systems [18, 28, 29, 31, 34, 35]. The
reactions occurring between HA and water during curing were studied by means of mass loss and
mass loss rate versus temperature curves. Information on the kinetics and products formed was
related to other physical properties, such as mechanical strength and porosity.
Green and fired cast samples of 70 mm  16 mm were measured (L: length in cm and D: diameter
in cm), weighed (M in g) and then ground (DParticle < 100 m) and characterised by helium
pycnometry (solid density, Solid in g∙cm-3). Their geometric total porosity (TP in %) was calculated
using Equation 5:













Solid
2LD
M4
1%100TP (5)
The ground samples were investigated by X-ray diffraction (Rotaflex RV 200B, Rigaku-Denki
Corp., Japan; CuKα radiation, 10 ° to 70 ° --range, 2 °∙min-1 scan rate) for identification of phases.
The Young’s modulus (E, GPa) of green and fired samples was measured by the impulse
excitation of vibration technique (Sonelastic, ATCP, Brazil) according to the ASTME 1876-01
5 of 15
computer recorded the samples’ mass and surface temperature every 10 s (the curves in Fig 5 are
based on a sequence of experimental points). This choice of thermal program highlighted the
amount of water released in each drying stage [29, 35, 36, 41]. Up to 120 °C, the mass loss is
related mostly to evaporative withdrawal of non-chemically bonded water (or free water). This first
stage of heating provides useful information on the drying behaviour of the compositions (for
instance, how long samples must be exposed to heat for the removal of most of the free water and
reduction of the likelihood of explosive spalling). From 120 to 800 °C, the remaining water leaves
the system by ebullition (as pressurized water vapour) and decomposition of hydroxylated
compounds.
Mass loss (W, mass-%, ranging from 0 to 100 %) and mass loss rate (dW/dt, mass-%∙h-1), for a
particular time ti, was calculated by Equations 3 and 4, respectively.


F0
i0
iMM
MM
%100W 

 (3)



10i10i
10i10i
itt
WW
dt
dW

 


(4)
where Mi is the instantaneous mass recorded at time ti during the heating of the samples, M0 is the
initial mass and MF is the final one (the i+10 and i-10 index refer to points collected 10 s after and
before the analyzed time i, respectively). These calculated parameters are particularly useful for
comparisons amongst formulations containing the same total water amount. Other studies have
described equivalent methods and phenomena for similar systems [18, 28, 29, 31, 34, 35]. The
reactions occurring between HA and water during curing were studied by means of mass loss and
mass loss rate versus temperature curves. Information on the kinetics and products formed was
related to other physical properties, such as mechanical strength and porosity.
Green and fired cast samples of 70 mm  16 mm were measured (L: length in cm and D: diameter
in cm), weighed (M in g) and then ground (DParticle < 100 m) and characterised by helium
pycnometry (solid density, Solid in g∙cm-3). Their geometric total porosity (TP in %) was calculated
using Equation 5:













Solid
2LD
M4
1%100TP (5)
The ground samples were investigated by X-ray diffraction (Rotaflex RV 200B, Rigaku-Denki
Corp., Japan; CuKα radiation, 10 ° to 70 ° --range, 2 °∙min-1 scan rate) for identification of phases.
The Young’s modulus (E, GPa) of green and fired samples was measured by the impulse
excitation of vibration technique (Sonelastic, ATCP, Brazil) according to the ASTME 1876-01
3. Results and discussion
3.1 Green samples (humid and dried at 120 °C)
Knowledge about the hydroxylation products formed during the curing
of HA and how they can be aected by matrix particles is mandatory for
understanding the way a suspension’s physical properties change during
rst heat-up. Therefore, the authors evaluated three aspects of the HA hy-
droxylation products formed in green samples, namely: a) chemical com-
position and crystalline structure, b) amount generated, and c) particle
morphology.
After the drying process for both the FA and CA systems, the gel respon-
sible for hardening the structure precipitated as poorly crystalline boeh-
mite particles (Figs. 4a−d), unlike other studies that reported traces of
bayerite and gibbsite [13−17]. This divergence occurred because dierent
time-temperature conditions were employed and the HA hydroxylation in
our study was conducted in the presence of calcined alumina (the systems
in other literature reports employed only HA and water). Thermogravi-
metric analysis carried out on humid samples after 24 h of curing (Fig. 5)
revealed that the ratio between the amount of free water (released up to
120 °C) and chemically-bonded water (released above 120 °C) was similar
for all compositions containing the same amount of HA but dierent cal-
cined alumina grades. Therefore, the average size of the calcined alumina
particles did not aect the type and quantity of hydroxylation products
formed when samples with equal HA amount were tested.
On the other hand, the morphologies of the HA hydroxylation products
were signicantly dierent for the dierent types of calcined alumina
in the samples. Whereas the precipitated boehmite of the coarse alu-
mina matrix sample formed a continuous coating of stacked thin plates
(Fig. 6a), the precipitate of ne alumina showed discrete particles whose
average size ranged from roughly 20 to 50 nm (Fig. 6b). This behaviour can
be understood through analysis of the physical characteristics of the cal-
cined alumina particles. Both grades have similar chemical composition
and crystalline structure (Figs. 4a−b), but considerably dierent levels of
6
Fig.6 • Green samples (dried at 120 °C) of the compositions containing 20 vol.-%
hydratable alumina and 80 vol.-% calcined alumina: a) coarse, b) fine
(a)
(b)

029REVIEW PAPERS
7 8
Fig.7 • Green samples (dried at 120 °C) of the compositions containing a) 10, b)
20, and c) 40 vol.-% hydratable alumina (i.e. 90, 80 and 60 vol.-% fine calcined
alumina, respectively)
Fig.8 • Effect of thermal treatment (from 120 to 1500 °C) on the physical proper-
ties (diametric compression strength, flexural elastic modulus and total porosity)
of samples containing different amounts of HA (10, 20 and 40 vol.-%) and calcined
alumina grades (coarse or fine)
specic surface area (CA = 1,3 m2g-1; FA = 8,2 m2g-1, Table 1). Although
the same amount of hydroxylated product was formed, its precipitation
on a smaller surface area would result in a continuous thicker and wrin-
kled coating (Fig. 6a). Conversely, if precipitation occurs across a wider
area, the coating would most likely be discontinuous, as observed with
FA (Fig. 6b). In precipitation processes based on aqueous solutions, every
calcined alumina particle behaves as a nucleation site; consequently, the
smaller the matrix particles, the more numerous and ner the precipitate.
Boehmite-precipitated particles behaved as clipping or interlocking
points amongst the calcined alumina particles, which strengthened the
structure [13, 14]. Consequently, the larger the HA content in the formu-
lation, the more abundant the points (Fig. 7) and the stronger the struc-
ture (Figs. 8a−d). The dierences in compression strength and rigidity
levels for samples containing equal HA content, but dierent calcined
alumina matrices can be attributed to the eciency of these connecting
points and the average pore size observed after setting. Because ne
alumina-containing samples showed more ecient connections (due to
their larger surface area) and smaller inter-particle pores, they become
stronger than their counterparts made with coarse alumina.
In cast suspensions containing hydraulic binders, such as hydratable
alumina and calcium aluminate cement, a consolidation process occurs
during the curing step and in the presence of water. Consequently, the
total porosity of the structure after drying is similar to the volumetric
amount of water in the composition (50 vol.-%) [20, 38, 39, 42, 43]. For
both types of samples, greater HA content in the formulation (rising from
10 to 40 vol.-%) increased the compression strength of humid and dried
green samples and the rigidity of dried green samples by almost one or-
der of magnitude, but reduced the total porosity by less than 5 percent-
age points (Fig. 8). These results indicate that the HA setting mechanism
is more strongly based on the interlocking of matrix particles than on a
(a)
(b)
(c)
reduction of the structural porosity. They also explain why HA-bonded
castable structures usually show very low permeability [18, 32−35]. The
very ne precipitated boehmite particles clog the permeable paths locat-
ed at the inter-particle pores, increasing their tortuousness and reducing
the permeability of the structure. As a consequence, the higher the HA
content, the lower was the maximum drying rate observed at 120 °C
(Fig. 5b).
3.2 Samples treated at 300−900 °C
In comparison to the dry samples (24 hrs at 120 °C ), those treated at 300 °C
showed signicantly higher levels of compression strength, rigidity and
porosity (Fig. 8). The microstructure evolution (Figs. 9a−b and 10a−b)
and thermogravimetric analysis (Fig. 5c) suggest that this behaviour is re-
lated to withdrawal of a portion of residual free-water (which occurred up
to 200 °C) [18, 35]. At the same time, the interlocking points amongst the
calcined alumina particles become stier. Several studies have indicated
that plain boehmite decomposition starts at approximately 270−290 °C
and lasts up to 350−400 °C, depending on the heating rate, degree of
crystallisation and particle size [3, 7−12]. Because traces of low crystal-
linity boehmite still remain detectable by XRD after thermal treatment at
300 °C in the tested systems (Figs. 4e−f), the particles may be helping to
maintain the rigidity of the structure up to this temperature.
Above 300 °C , all compositions showed a signicant reduction in com-
pression strength and rigidity and a slight porosity increase (Fig. 8). At this
point, no detectable traces of boehmite remained in the samples (Figs.
4g−l) and volumetric shrinkage was observed at around 450 °C (Fig. 11,
magnied region). Two eects explain this behaviour. First, as reported
for similar systems, the conversion of boehmite to other transition phases
of alumina (such as η-Al2O3) is followed by a density increase [9−12,
44, 45] and a consequent volumetric contraction. Second, the connecting

030 2017
910
Fig.9 • Effect of thermal treatment (up to 900 °C) on the microstructure of com-
positions containing 20 vol.-% hydratable alumina and 80 vol.- % fine calcined
alumina
Fig.10 • Effect of thermal treatment (up to 1500 °C) on the microstructure of com-
positions containing 20 vol.- % hydratable alumina and 80 vol.- % coarse calcined
alumina
(a)
(b)
(c)
(d)
(e)
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
points amongst the calcined alumina particles begin to collapse at 500 °C
(Figs. 9c−d and 10c−d) and are completely destroyed by 900 °C (Figs. 9e
and 10e). Then the absence of bonding points amongst the matrix parti-
cles has large impact, weakening the structure as porosity is enhanced.
The thermal treatment temperature aected compositions of distinct
grades of calcined alumina dierently (Fig. 8). For samples of coarse
alumina and high HA content, the strength reduction was less intense
than for corresponding mixtures containing ne alumina. On the other
hand, porosity increased along with binder content, i.e. the greater the
HA amount, the more porous the samples. These results can be explained
by the formation of ne transition alumina particles resulting from boeh-
mite decomposition throughout the calcined alumina matrix. For sam-
ples of coarse matrix, the introduction of these ne particles increased
the driving force for sintering and reduced the average pore size [44, 45];
for the ne matrix, which was already very reactive, the introduction of a
large quantity of pores reduced its sinterability.
3.3 Samples treated at 1100−1500 °C
Above 900 °C , as sintering events began, all compositions showed huge
gains of strength and rigidity, became denser and showed signicant vol-
umetric shrinkage (Figs. 8 and 11). The ne transition alumina particles
formed after HA decomposition (up to 900 °C ) sintered with the calcined
alumina particles (Figs. 10f−h and 12a−c).
[45]. Comparison of samples within the same alumina grade showed
that dierent HA content exerted only minor inuence over physical
properties. On the other hand, signicant dierences were observed for
compositions with the same HA content but dierent calcined alumina
matrices. Fine calcined alumina-containing samples showed higher den-
sication (greater porosity reduction and linear thermal shrinkage) across
all temperatures. As discussed in the previous section, the greater driving
force for sintering originates from the grade’s smaller average particle size
and high surface area.
Comparisons of the values for each physical property with results of oth-
er studies on similar systems [38, 39, 40, 42, 45−48] highlight the lower
levels of strength and high porosity attained after sintering at 1500 °C.

031REVIEW PAPERS
11 13
12
Fig.11 • Dilatometric analysis of samples containing a) coarse, and b) fine cal-
cined alumina and different amounts of hydratable alumina
Fig.13 • Large porous aggregate resulting from non-reacted rho-alumina par-
ticles formed after sintering (composition 90FA_10HA sintered at 1500 °C )
Fig.12 • Effect of thermal treatment (1100, 1300 and 1500 °C) on the microstruc-
ture of compositions containing 20 vol.- % hydratable alumina and 80 vol.-% fine
calcined alumina
(a)
(b)
For instance, under similar sintering conditions, ne calcined alumina
compacted by uniaxial pressing usually shows porosity levels below 5 %
and exural elastic modulus over 350 GPa. The highest value achieved in
this study was 267 GPa for the 90FA_10HA composition. This divergence
is explained by the presence of particles of non-hydratable phases of alu-
mina. During the production of rho-alumina, part of the Al(OH)3 precursor
is converted into other transition aluminas that do not react with water
(according to the HA producer, 1−3 mass-% κ-Al2O3 and η-Al2O3 can be
found in each batch). During thermal treatment, these particles evolve
into large porous clusters inserted in the sintered calcined alumina matrix
that reduce the structure’s strength and rigidity. Fig. 13 shows one of the
regions where a awed interface is clearly seen between the highly densi-
ed ne alumina matrix and the large porous aggregate.
4. Final remarks
The hydratable alumina (HA) setting mechanism starts with the hydrox-
ylation of rho-alumina and its further dissolution and the generation of
a pseudo-boehmite gel. During drying, the gel precipitates and forms
boehmite nano-units amongst calcined alumina particles. Both tested
systems showed formation of the same hydroxylation products (boehmite
nanoparticles) in very similar quantities. Nevertheless, the morphology
of the nano-units was signicantly aected by the average size of the
calcined alumina particles. In the formulation containing coarse calcined
alumina matrix (D50 of 10 μm), boehmite precipitates as a continuous
coating of stacked thin plates. The ne alumina-containing sample, on
the other hand, showed discrete boehmite particles whose average size
roughly ranged from 20−50 nm. This diversity in particle morpholo-
gy was attributed to the dierence in specic surface area between the
calcined alumina grades (precipitation occurring on low specic surface
area generated thicker layers of particles). Because the boehmite units re-
strain movement of matrix particles and harden the structure, greater HA

032 2017
amounts in the formulation result in stronger and less porous green sam-
ples. The bonding mechanism is more eective for ne particles, probably
due to the establishment of a larger number of connections, which pro-
duces higher levels of strength and rigidity.
During nitial heating above 300 °C , the boehmite connections begin
to collapse as they become denser and shrink due to the dehydroxilation
process. This collapse, rather than a slight porosity increase, was respon-
sible for the samples’ huge loss of strength and rigidity that lasted up to
900 °C. The intensity of variation in compression strength, exural elastic
modulus and total porosity was also aected by both the amount of add-
ed HA and the grade of calcined alumina employed in the compositions.
During sintering of coarse calcined alumina particles, ne HA products
enhanced the driving force for sintering and caused earlier densication
and strength gain. On the other hand, ne alumina−based compositions
are systems of intrinsically high reactivity, and therefore the HA particles
introduced extra porosity and reduced sinterability.
We can provide some recommendations regarding the use of HA in
castable systems based on the observed mechanisms and achieved re-
sults. Use of coarser particles is a more suitable choice for production of
porous structures (thermal insulating castables, for example) due to their
inherent densication and sintering diculty. In this case, HA content
larger than 10 % yields two important advantages. First, higher levels of
strength in green samples facilitates demoulding and manipulation; sec-
ond, because HA particles also behave as porogenic agents, higher levels
of porosity can be achieved. On the other hand, ne calcined alumina par-
ticles are employed for production of dense parts (for thermo-mechanical
purposes, for instance) in which very low porosity after sintering is a man-
datory condition. Furthermore, use of HA amounts greater than 10 % will
certainly reduce the nal quality of the product.
Acknowledgements
The authors acknowledge Brazilian Research Foundations FAPESP (2010-
19274-5) and CNPq (470981/2011-3 and 306036/2011-8) for supporting
this study and Almatis (Brazil, USA and Germany) for kindly supplying
samples of calcined and hydratable aluminas. We are also indebted to P.L.
Lorenzo, J.J. Bernardi and Prof. Dr. J.M.D.A. Rollo (SMM/EESC) for dilato-
metric analyses and mechanical tests and Wagner R. Correr (Electron
Microscopy Laboratory of the Center for Technology of Hybrid Materials,
CTMH) for SEM images.
References
[1]
[2]
[3]
[4]
[5]
[6]
Brown, J.F., Clark D., Elliott, W.W.: The thermal decomposition of the alumina tri-
hydrate gibbsite. J. Chem. Soc. (1953) 84−88
Gitzen, W.H.: Alumina as ceramic material. The American Ceramic Society, Westerville
(1970)
Burtin, P.: Inuence of surface area and additives on the thermal stability of transi-
tion alumina catalyst supports, II: Kinetic model and interpretation. Appl. Catalysis 34
(1987) [1−2] 239−254
Levin, I., Brandon, D.: Metastable alumina polymorphs: crystal structures and tran-
sition sequences. J. Am. Ceram. Soc. 81 (1998) [8] 1995−2012
Musselman, L.L.: Production processes, properties, and applications for alumi-
num-containing hydroxides. Alumina chemicals: Science and technology handbook
(1990) 75−92, ISBN: 978-0-916094-33-1
Goodboy, K.P., Downing, J.C.: Production processes, properties, and applications for
activated and catalytic aluminas. Alumina chemicals: Science and technology hand-
book (1990) 93−108, ISBN: 978-0-916094-33-1
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
Zhou, R.S., Snyder, R.L.: Structure and transformation mechanisms of the η, ϒ, and θ
transition aluminas. Acta Crystall. Section B: Structural Science 47 (1991) [5] 617−630
Santos, P.S., Santos, H.S., Toledo, S.P: Standard transition aluminas: Electron mi-
croscopy studies. Mater. Res. 3 (2000) [4] 104−114
Bhattacharya, I.N., Das, S.C., Mukherjee, P.S., Paul, S., Mitra, P.K.: Thermal
decomposition of precipitated ne aluminium trihydroxide, Scand. J. Metal. 33 (2004)
[4] 211−219
Coelho, A.C.V., Santos H.S.S., Kiyohara P.K., Marcos K.N.P., Santos P.S.S.: Sur-
face area, crystal morphology and characterization of transition alumina powders from
a new gibbsite precursor. Mater. Res. 10 (2007) [2] 183−189
Gan, B.K., Madsen, I.C., Hockridge, J.G.: In situ X-ray diraction of the transformation
of gibbsite to alpha-alumina through calcination: Eect of particle size and heating rate.
J. Appl. Crystall. 42 (2009) [4] 697−705
Souza, A.D.V., Arruda, C.C., Fernandes, L., Antunes, M.L.P., Kiyohara, P.K., Sa-
lomão, R.: Characterization of aluminum hydroxide (Al(OH)3) for its use as a porogenic
agent in castable ceramics. J. Europ. Ceram. Soc. 35 (2015) [2] 803−812
Hong, Y.: ρ-Alumina bonded castable refractories. Taikabutsu Overseas 9 (1988) [1]
35−38
Ma, W., Brown, P.W.: Mechanisms of reaction of hydratable aluminas. J. Am. Ceram.
Soc. 82 (1999) [2] 453−456
Mista, W., Wrzyszcz, J.: Rehydration of transition aluminas obtained by ash calcina-
tion of gibbsite. Thermochimica Acta 331 (1999) [1] 67−72
Vaidya, S.D., Thakkar, N.V.: Eect of temperature, pH and ageing time on hydration
of rho alumina by studying phase composition and surface properties of transition
alumina obtained after thermal dehydration. Mater. Letters 51 (2001) [4] 295−300
Vaidya, S.D., Thakkar, N.V.: Study of phase transformation during hydration of rho
alumina by combined loss of ignition and X-ray diraction technique. J. Phys. and
Chem. Solids 62 (2001) [5] 977−986
Salomão, R., Ismael, M.R., Pandolfelli, V.C.: Hydraulic binders for refractory
castables: Mixing, curing and drying. CFI 84 (2007) [9] 103−108
Nagaoka, T., Duran, C., Isobe, T., Hotta, Y., Watari, K.: Hydraulic alumina binder
for extrusion of alumina ceramics. J. Am. Ceram. Soc. 90 (2007) [12] 3998−4001
Souza, A.D.V., Salomão, R.: Evaluation of the porogenic behavior of aluminum hy-
droxide particles of dierent size distribution in castable high-alumina structures. J.
Europ. Ceram. Soc. 36 (2016) [3] 885−897
Pinto, U.A., Visconte, L.L.Y., Gallo J.B.: Flame retardancy in thermoplastic poly-
urethane elastomers with mica and aluminum trihydrate. Polymer degradation and
Stability 69 (2000) [3] 257−260
Santos, P.S., Coelho, A.C.V., Santos, H.S.S., Kiyohara, P.K.: Hydrothermal syn-
thesis of well-crystallized boehmite crystals of various shapes. Mater. Res. 12 (2009)
[4] 437−445.
Rebouillat, L., Rigadu, M.: Andalusite-based high alumina castables. J. Am. Ceram.
Soc. 85 (2002) [2] 373−378
Ismael, M.R., Salomão, R., Pandolfelli, V.C.: Refractory castables based on colloi-
dal silica and hydratable alumina. Am. Ceram. Soc. Bull. 86 (2007) [9] 58−61
Zhang, J., Jia, Q., Yan, S., Zhang, S., Liu, X.: Microstructure and properties of hy-
dratable alumina bonded bauxite–andalusite based castable. Ceram. Inter. 42 (2016)
[1] 310−316
Ye, G., Troczynski, T.: Hydration of hydratable alumina in the presence of various
forms of MgO. Ceram. Inter. 32 (2006) [3] 257−262
Ahari, K.G., Sharp, J.H., Lee, W.E.: Hydration of refractory oxides in castable bond
systems, I: Alumina, magnesia, and alumina–magnesia mixtures. J. Europ. Ceram. Soc.
22 (2002) [4] 495−503

033REVIEW PAPERS
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]
Oliveira, I.R., Pandolfelli, V.C.: Castable matrix, additives, and their role on hydrau-
lic binder hydration. Ceram. Inter. 35 (2009) [4] 1453−1460
Salomão, R., Pandolfelli, V.C.: The role of hydraulic binders on magnesia containing
refractory castables: Calcium aluminate cement and hydratable alumina. Ceram. Inter.
35 (2009) [8] 3117−3124
Braulio, M.A.L., Bittencourt, L.R.M., Pandolfelli, V.C.: Selection of binders for
in situ spinel refractory castables. J. Europ. Ceram. Soc. 29 (2009) [13] 2727−2735
Oliveira, I.R., Pandolfelli, V.C.: Does a tiny amount of dispersant make any change
to refractory castable properties? Ceram. Inter. 36 (2010) [1] 79−85
Ribeiro, C., Innocentini, M.D.M., Pandolfelli, V.C.: Permeability behavior during
drying of refractory castables based on calcium-free alumina binders. J. Am. Ceram. Soc.
84 (2001) [1] 248−250
Innocentini, M.D.M., Pardo, A.R.F., Pandolfelli, V.C., Menegazzo, B.A., Bitten-
court, L.R.M., Rettore, R.P.: Permeability of high-alumina refractory based on various
hydraulic binders. J. Am. Ceram. Soc. 85 (2002) [6] 1517−1521
Salomão, R., Cardoso, F.A., Innocentini, M.D.M., Pandolfelli, V.C., Bittencourt,
L.R.M.: Eect of polymeric bers on refractory castables permeability. Am. Ceram. Soc.
Bul. 82 (2003) [4] 51−56
Cardoso, F.A., Innocentini, M.D.M., Miranda, M.F.S., Valenzuela, F.A.O., Pan-
dolfelli, V.C.: Drying behavior of hydratable alumina-bonded refractory castables. J.
Europ. Ceram. Soc. 24 (2004) [5] 797−802
Salomão, R., Pandolfelli, V.C.: Magnesia sinter hydration-dehydration behavior in
refractory castables. Ceram. Inter. 34 (2008) [8] 1829−1834.
Luz, A.P., Neto A.S., Santos, T., Medeiros, J., Pandolfelli, V.C.: Mullite-based re-
fractory castable engineering for the petrochemical industry. Ceram. Inter. 39 (2013)
[8] 9063−9070
Souza, A.D.V., Sousa, L.L., Fernandes, L., Cardoso, P.H.L., Salomão, R.:
Al2O3-Al(OH)3-Based castable porous structures. J. Europ. Ceram. Soc. 35 (2015) [6]
1943−1954
Salomão, R., Souza, A.D.V., Cardoso, P.H.L.: A comparison of Al(OH)3 and Mg(OH)2
as inorganic porogenic agents for alumina. InterCeram: Inter. Ceram. Rev. 64 (2015)
[4] 193−194
Salomão, R., Villas-Boas, M.O.C., Pandolfelli, V.C.: Porous alumina-spinel ceram-
ics for high temperature applications. Ceram. Inter. 37 (2011) [7] 1393−1399
Innocentini, M.D.M., Miranda, M.F.S., C ardoso, F.A., Pandolfelli, V.C.: Vaporiza-
tion processes and pressure builtup during dewatering of dense refractory castables. J.
Am. Ceram. Soc. 86 (2003) [9] 1500−1503
Sousa, L.L., Souza, A.D.V., Fernandes, L., Arantes, V.L., Salomão, R.: Develop-
ment of densication-resistant castable porous structures from in situ mullite. Ceram.
Inter. 41 (2015) [8] 9443−9454
Sousa, L.L., Salomão, R., Arantes, V.L.: Development and characterization of po-
rous moldable refractory structures of the alumina-mullite−quartz system. Ceram.
Inter. 43 (2017) [1B] 1362−1370
Deng, Z.Y., Fukasawa, T., Ando, M.: High-surface-area alumina ceramics fabricated
by the decomposition of Al(OH)3. J. Am. Ceram. Soc. 84 (2001) [3] 485−491
Kwon, S., Messing, G.L.: Sintering of mixtures of seeded boehmite and ultrane al-
pha-alumina. J. Am. Ceram. Soc. 83 (2000) [1] 82−88
Deng, Y., Fukasawa, T., Ando, M.: Microstructure and mechanical properties of po-
rous alumina ceramics fabricated by the decomposition of aluminum hydroxide. J. Am.
Ceram. Soc. 84 (2001) [11] 2638−2644
Oliveira, I.R., Leite, V.M.C., Lima, M.P.V.P., Salomão, R.: Production of porous
ceramic material using dierent sources of alumina and calcia. Revista Matéria 20
(2015) [3] 739−746
Salomão, R., Fernandes, L.: Porous co-continuous mullite structures obtained from
sintered aluminum hydroxide and synthetic amorphous silica. J. Europ. Ceram. Soc. 37
(2017) [8] 2849−2856
Received: 06.07.2017