Full text of DOI:10.1111/j.1151-2916.2000.tb01659.x

J. Am. Ceram. Soc., 83 [12] 2907–12 (2000)
journal
Synthesis of Hexacelsian Barium Aluminosilicate by a
Solid-State Process
Kuo-Tong Lee* and Pranesh B. Aswath*
Department of Mechanical and Aerospace Engineering and Materials Science and Engineering Program,
University of Texas, Arlington, Texas 76019
Synthesis of hexacelsian barium aluminosilicate (BaAl Si O
8
change and the temperature of the hexagonal-to-orthorhombic
transformation in BAS are decreased, making it feasible to use
hexacelsian BAS as a matrix material in composites.
2
2
or BAS) from BaCO , Al O , and amorphous SiO was
3
2
3
2
examined. BaCO3 can react with SiO2 and Al O to form
2
3
barium silicates (Ba SiO or B S, BaSiO or BS, and BaSi O
With low processing costs and ease of operation, solid-state
reaction for producing ceramic components is a common industrial
method. Ceramists who employ solid-state reaction methods may
take advantage of information available in phase diagrams. How-
ever, the usefulness of the available phase diagram for the
2
4
2
3
2
5
or BS ) and barium aluminate (BaAl O or BA). It is shown
2
2
4
that there are two competitive reactions leading to the forma-
tion of hexacelsian BAS. One involves a reaction between BS2
and Al O and the other involves a reaction between BA and
2
3
1
SiO . In experiments with the model BS –Al O and BA–SiO
BaO–Al O –SiO system is limited, especially for subsolidus
2
2
2
3
2
2
3
2
1
5
systems it is shown that the reaction between BS and Al O is
equilibria. Therefore, Planz et al. made extensive studies of the
formation of hexacelsian from powder mixtures containing
BaCO , ␥-Al O , and amorphous SiO . The purpose of the present
2
2
3
much faster than the reaction between BA and SiO . However,
2
in the BAS system, Al O suppresses the reactions which form
2
3
3
2
3
2
BS2 and instead reacts with B S and BS to form BA. The
paper is to determine the reaction path for the formation of
2
kinetics of hexacelsian BAS formation are greatly enhanced
when BS is made separately and fired with Al O to yield
BAS.
hexacelsian from ␥-BaCO (witherite), ␣-Al O , and amorphous
SiO2.
3
2 3
2
2
3
II. Experimental Procedure
Samples were prepared from reagent-grade powders of BaCO3
I. Introduction
aAl Si O (BAS) is the only ternary compound in the BaO–
2
2 8
(Mallinckrodt AR having 99% purity), SiO2
(dried powder of
Nyacol 2034DI colloidal silica having 20 ␮m particle size), and
Al
1
B
Al O –SiO system. The polymorphism and crystal struc-
2 3 2
tures of BAS, a member of the feldspar family, has been widely
2O3 (Baikowski SM-8 having 99.99% purity and 0.15 ␮m
2
–5
studied by mineralogists. In recent years considerable attention
has been devoted to BAS by ceramists because of its potential
particle size). These powders in appropriate ratios were well mixed
and calcined in ambient stagnant air at various temperatures to
form the desired compounds. The furnace (chamber dimensions:
17.8 cm ϫ 12.7 cm ϫ 25.4 cm) was stabilized at each temperature
for 30 min or more before introducing the powders. Reactants in
the form of loose powder with approximately 0.4 g weight were
6
–9
application as a matrix in ceramic-matrix composites.
BAS exists primarily in three different polymorphs: celsian
monoclinic), hexacelsian (hexagonal), and ␣-hexacelsian (orthor-
hombic). Below 1590°C celsian is the stable phase and hexacelsian
(
is stable from 1590°C to the melting point at approximately
placed within an Al O boat of ϳ10 mm depth. The samples were
2
3
5
1
760°C. However, the hexacelsian-to-celsian transformation is
held isothermally for periods varying from 4 to 24 h and then
removed from the furnace for air-quenching. No reaction was
extremely sluggish and hexacelsian exists as a metastable phase
below 1590°C. When celsian is synthesized by any of the
found to occur between the Al
O boat and the reactants.
3
2
1
0,11
following methods—solid-state reaction,
crystallization of a
In order to understand the mechanisms of formation of BAS
from its precursors, several experiments were designed. These
include model calcination experiments where binary systems like
1
2
13
glass preform, sol–gel method, or the oxidation of solid
metallic precursors (SMP) —hexacelsian always appears at first
as a metastable phase before the formation of celsian. In addition
1
4
BaCO3–
SiO , BaCO –Al O , Al O –SiO , and BaSiO –BaCO
2
3
2
3
2
3
2
3
3
when used as a matrix in Si N -bearing composites, it has been
and the ternary BaCO3–Al
3
4
2
O
3
–SiO system were examined. Re-
2
shown that the formation of celsian is suppressed and hexacelsian
actions were conducted for various combinations of temperature
and time to provide insight into the formation of BAS.
7
,10,11
exists as a metastable phase below 1590°C.
The metastable
hexacelsian undergoes a phase transformation to the orthorhombic
structure at approximately 300°C accompanied by a large volume
Phases present in the calcined powders were identified at room
temperature by X-ray diffraction using a Phillips PW 1729
diffractometer equipped with CuK␣ radiation. The intensities of
the X-ray diffraction peaks were also used to characterize the
extent of reaction which is summarized in Table I. The notation
used to identify the variety of compounds in this study is BaCO3
2
change which is undesirable for composite applications. Bandyo-
7
padhyay et al. found that with an increasing amount of Si N in
3
4
Si N -reinforced BAS matrix composites, the extent of volume
3
4
(
W), BaO (B), SiO (S), Al O (A), 2BaO⅐SiO (B S), BaO⅐SiO
2 2 3 2 2 2
(
BS), 2BaO⅐3SiO2 (B S ), BaO⅐2SiO (BS ), BaO⅐Al O (BA),
2
3
2
2
2 3
R. J. Kerans—contributing editor
Al O ⅐SiO (AS), Al O ⅐2SiO (AS ), 3Al O ⅐2SiO (A S ), and
2
3
2
2
3
2
2
2
3
2
3 2
hexacelsian BaO⅐Al O ⅐2SiO (H or BAS).
2
3
2
Manuscript No. 191490. Received September 25, 1996; approved August 5, 1998.
This paper was presented under the title “Formation Mechanism of Hexacelsian
III. Results and Discussion
3 2 3 2
Barium Aluminosilicate from Mixtures of BaCO , Al O and SiO ” at the 98th
Annual Meeting of the American Ceramic Society, Indianapolis, IN, April 14–17,
(1) Absence of Mullite During Synthesis
According to the accepted phase diagram for the BaO–Al O –
SiO system shown in Fig. 1, mullite (A S ) is the only binary
2 3 2
1
996 (Paper No. B-145-96, Reactive Processing of Composites Symposium).
Supported by the National Science Foundation under Contract No. MSS-9108891.
2
3
*Member, American Ceramic Society.
2
907

2
908
Journal of the American Ceramic Society—Lee and Aswath
Vol. 83, No. 12
Table I. Summary of the Results of X-ray Diffraction Analysis
Heat treatment
conditions
BaOSiO
2
ϩ 0.4 wt%
†
Al
2
O
3
ϩ 2SiO
2
BaCO
3
ϩ Al
2
O
3
BaCO
3
ϩ SiO
2
BaCO
3
ϩ 2SiO
2
BaCO
3
3 2 3 2
BaCO ϩ Al O ϩ 2SiO
‡
‡
‡
6
7
7
8
8
9
9
00°C, 6 h No reaction No reaction
W
W
No reaction
W, A
00°C, 6 h No reaction No reaction
50°C, 6 h No reaction No reaction
00°C, 6 h No reaction No reaction
W 2, B S
W 2, B S
W 2, B S
W 2, A, B S
2
2
2
2
W 2, B S 1 W 2, B S 1
W 2, B S 1 W 2, A, B S 1
2
2
2
2
W 2, B S 1 B S 1, BS
W 2, A, B S 1
2
2
2
50°C, 6 h No reaction W 2, A 2, BA
W 2, B S 1 B S 1, BS 1
W 2, A, B S 1
2
2
2
00°C, 6 h No reaction W 2, A 2, BA 1 B S 1, BS
BS , B S 1, BS 1
A, B S, BS, BA
2
2
2
2
50°C, 6 h No reaction W 2, A 2, BA 1 B S, BS
BS 1, B S 2, BS 1
A 2, B S 2, BS, BA 1
2
2
2
2
1
1
1
1
1
000°C, 6 h No reaction BA 1
B S 2, BS 1 BS 1, B S 2, BS 1
A 2, B S 2, BS 1, BA 1
2
2
2
2
050°C, 6 h No reaction No further rxn
B S 2, BS 1 BS 1, B S 2, BS 1
H, B S 2, BS 1, BA 1
2
2
2
2
100°C, 6 h
150°C, 8 h
200°C, 6 h
†
B S 2, BS 1 BS 1, B S 2, BS
2
2
2
B S 2, BS 1
2
No further rxn BS , BS
H, BA
2
‡
2 3 2 2 3 2 2
Al O ϩ 2SiO means the raw material was prepared from Al O and SiO with 1:2 molar ratio. Amorphous SiO cannot be identified by X-ray diffraction.
compound in the Al O –SiO subsystem and could form as an
(2) Formation of B S from 700°C to 850°C
2
2
3
2
equilibrium phase if the overall composition of a precursor is BaO
In the BaCO3–Al O –SiO system no reaction occurs below
700°C and, between 700° and 850°C, there is a decrease in BaCO3
2
3
2
1
6
deficient in the BaCO –Al O –SiO system. Planz et al. inves-
3
2
3
2
tigated the solid-state reactions involved in the BaO–Al O –SiO
coupled with an increase in B S. Based on thermodynamic data
2
3
2
2
system. They observed the formation of mullite above 1000°C
shown in Fig. 2, pure BaCO does not decompose into BaO and
3
1
2
only in the Al O ⅐2SiO subsystem. Bansal et al. prepared
CO
with SiO
2
until approximately 900°C. However, when BaCO
3
is mixed
2
3
2
samples with composition (in wt%) 10BaO⅐38Al O ⅐51SiO ⅐
2
, the BaCO reacts with SiO and forms various silicates
3
2
2
3
2
at temperatures beginning as low as 700°C. The absence of BaO as
an intermediate phase may be due to the slow rate of BaCO3
1
MoO , inside the primary mullite field. After melting at
3
ϳ2000°C followed by quenching and annealing at 1150°C for 1 or
0 h, the samples were identified as mullite and they found
substantial amounts of amorphous phase remaining. Pavlova et
decomposition relative to the rates of formation of BS and B S or
1
2
due to the direct reaction between BaCO and SiO . The following
3
2
1
7
are three possible reactions for the formation of B S:
2
al. using kaolin, clay, and carbonates as starting materials,
concluded that hexacelsian is formed by BaO diffusion in the
BaCO ϩ SiO ϭ BS ϩ CO
(1)
(2)
(3)
3
2
2
1
4
metakaolinite residue (AS ). Schmutzler et al. also found the
2
existence of mullite before the formation of hexacelsian at 1150°C.
In their experiments pure metallic powders of Ba, Al, and Si were
mixed and oxidized between 300° and 1200°C in an environment
of oxygen. However, in the current experiments, as seen in Table
BS ϩ BaCO
3
ϭ B
2
S ϩ CO
2
2
BaCO ϩ SiO ϭ B S ϩ 2CO
3
2
2
2
Two possible approaches may be used to explain the formation
I, no Al O –SiO compound was observed in the Al O ⅐2SiO
2
3
2
2
3
2
of B S:
2
system as well as in the BaO–Al O –SiO system at all tempera-
2
3
2
(i) Two-step process to form B S with formation of BS as
intermediate step: According to thermodynamic data
2
1
9
tures up to 1200°C. The absence of mullite, below 1200°C, in the
ex-
Al O ⅐2SiO system may be because ␣-Al O used in this study is
pressed in terms of the partial pressure of CO (PCO2) as a function
2
3
2
2
3
2
more stable than ␥-Al O used in Planz’s studies. In the BaO–
of temperature for various reactions in Fig. 2, reaction (1) is more
2
3
2
0
Al O –SiO system, this absence of mullite is consistent with the
favorable than reaction (3). Keler and Glushkova studied the
solid-state reactions between BaCO and crystalline SiO on
2
3
2
1
1
15
18
works of Bandyopadhyay et al., Planz et al., and Zhang et al.
3
2
continuously increasing the temperature. They observed that the
reaction, for all initial compositions, starts with the formation of
Fig. 2. Calculated CO2 pressures as a function of temperature for the
Ϫ4
indicated reactions. In dry air, the partial pressure of CO is 3 ϫ 10 atm
2
and log [PCO2] ϭ Ϫ3.5. The four reactions of interest are (a) BaCO3
ϩ
SiO2 ϭ BS ϩ CO , (b) 2BaCO ϩ SiO2 ϭ B S ϩ 2CO , (c) BS ϩ
BaCO ϭ B S ϩ CO , and (d) BaCO ϭ BaO ϩ CO .
3 2 2 3 2
2
3
2
2
1
Fig. 1. Phase diagram for the BaO–Al O –SiO system.
2
3
2

December 2000
Synthesis of Hexacelsian Barium Aluminosilicate by a Solid-State Process
2909
BS at 700°C, but the amount of BS is insignificant up to 800°C. To
obtain the desired silicates at low temperatures, they investigated
2
1
the formation of barium silicates from BaO and crystalline SiO2.
Again, it was observed that BS and even the more acid silicates,
such as BS and B S , form first at temperatures between 300° and
2
2 3
4
00°C, independent of the ratio of the reactants. In our experi-
ments the absence of BS at temperatures between 700° and 750°C
may be due to the lower reactivity of the amorphous SiO used
2
(while the activity of SiO2 is 1 in both cases the reactivity is
dependent on other factors such as crystallinity), which results in
the slower rate of reaction (1) as compared to reaction (2).
This approach is also very similar to the one proposed to explain
2
2
the formation of B T (Ba TiO ) from BaCO and TiO . In that
2
2
4
3
2
reaction, B T (orthotitanate) only occurs by a reaction between
2
BaCO and BT (BaTiO ) and not directly by a reaction between
3
3
Fig. 3. X-ray diffraction patterns showing the formation of B S from BS
2
BaCO3 and TiO . In defense of this approach to explain the
and BaCO : (a) BS with a little B S made from the mixture of BaCO and
2
3
2
3
formation of B S, several experiments were designed. Figure 3(a)
SiO in the molar ratio 1:1 calcined at 1150°C for 8 h, (b) unchanged BS
after exposed at 850°C for 4 h, and (c) the formation of B S from the
2
2
2
indicates the production of BS. When BaCO and SiO are mixed
3
2
mixture of the almost pure BS with BaCO (approximately 2:1 molar ratio)
in a 1:1 molar ratio and calcined at 1150°C for 8 h, they yield BS
3
fired at 750°C for 4 h; (ء) B S peaks, (●) BS peaks, and (W) BaCO peaks.
with little B S. This BS powder does not decompose on further
2
3
2
calcination at 850°C for 4 h, shown in Fig. 3(b). BS is stable as
predicted by thermodynamic data shown in Fig. 2. However, when
BS is mixed with BaCO3 (approximately 2:1 molar ratio) and
and Al
O
to form BA and SiO
, which is reaction (6); subse-
2
3
2
calcined at 750°C for 4 h, B S is formed. This suggests that, if BS
quently the BA reacts further with BS to form B
Al by reaction (5). However, these two steps are thermody-
namically unfavorable. Further, if BS cannot be formed via
reaction (4) due to the presence of Al , what is the reaction that
can contribute to the formation of BS at 900°C in the BaCO
Al –SiO system? It is worth to noting in Table I that the
2
S and release
2
forms first, the formation of BS is an intermediate step in the
O
2 3
2
3
formation of B S. In addition, Lee et al., using a diffusion couple
2
of two pellets, one made of BaCO and the other made of SiO ,
O
2 3
3
2
concluded that barium silicates form within the SiO pellet after
–
3
2
diffusion of BaO through the interface between them. It is
proposed that for all molar ratio’s of BaCO and SiO , BS forms
2
O
3
2
formations of BA and BS begin at the same firing condition of
900°C for 6 h. This suggests that we may relate the formation of
BS to that of BA, just like reaction (5) or a combination of Eqs. (5)
and (6).
In the binary BaCO
tures above 850°C and 3BaO⅐Al
up to 1050°C, as shown in Table I. These results are in agreement
with the work of Perier-Camby et al. There is an absence of
intermediate-phase B A because, like CO , the oxygen in air
suppresses the formation of B A. However, in the BaCO3–
3
2
first and acts as a nucleus for formation of B S. This reaction then
2
continues until all of the BaCO3 is consumed. Experimental
evidence shown in Table I indicates the presence of B S and no
2
BS. The above hypothesis is feasible only if the kinetics for
reaction (2) are much faster than reaction (1) and all BS is
consumed as soon as it is formed.
3
–Al
2
O
3
system, BA is formed at tempera-
O
(noted B A) is not observed
2
3
3
2
4
(ii) Direct reaction between BaCO3 and SiO2 yielding B2S
with no BS formation: This reaction yields B S as shown in
3
2
2
2
4
reaction (3). This reaction is thermodynamically less favorable
when compared to reaction (1), but is still possible if the kinetics
for reaction (1) are lower than that for reaction (3).
3
(3) Formation of BS and BA from 850° to 1000°C
In the BaCO –Al O –SiO system all of the BaCO is con-
3
2
3
2
3
sumed to yield B S by 900°C. Between 850° and 1000°C, the
2
following reactions are proposed:
B S ϩ SiO ϭ 2BS
(4)
(5)
(6)
2
2
B S ϩ Al O ϭ BA ϩ BS
2
2
3
BS ϩ Al O ϭ BA ϩ SiO
2
2
3
A series of experiments were performed with BaCO and SiO
3
2
mixtures to study the formation of BS. Figure 4(a) shows pure B2S
made from a mixture of BaCO and SiO in molar stoichiometric
3
2
ratio (2:1) fired at 1100°C for 8 h. B S remains stable when held
2
at 850°C for 4 h, as shown in Fig. 4(b). When B S is mixed with
2
SiO in a molar ratio of 1:3 and fired at 730°C for 4 h, there is no
2
evidence of BS formation as shown in Fig. 4(c), despite the fact
that the standard free energy change for this reaction is negative at
all temperatures as shown in Fig. 5. Above 850°C, BS starts to
form as shown in Fig. 4(d) and Table I. In addition, when BS is
mixed with Al O (in approximately 1:1 molar ratio) and fired at
Fig. 4. X-ray diffraction patterns showing that the formation of BS from
2
3
B S and SiO begins at 850°C and is retarded by the presence of Al O : (a)
9
00°C for 4 h, the Al O is not consumed in the reaction, but BS
2
2
2 3
2
3
pure B S made from the mixture of BaCO and SiO mixed in stoichio-
decomposes to yield B S and SiO as shown in Fig. 4(e). The
2
3
2
2
2
metric ratio and fired at 1100°C for 8 h, (b) unchanged B S after being held
2
peaks of SiO2 are absent because of its possible amorphous
structure. This sequence suggests that the presence of Al O in the
at 850°C for 4 h, (c) no reaction occurred in the mixture of B S and SiO
2
2
2
3
in the molar ratio 1:3 after firing at 750°C for 4 h, (d) the formation of BS
under the same conditions as (c) but fired at 850°C for 4 h, and (e) the
decomposition of BS in the mixture of the almost pure BS with Al O
BaCO –Al O –SiO system will retard the formation of BS via
3
2
3
2
reaction (4) at 900°C. The suppression could be due to either
thermostructural effects or thermochemical effects. If this is a
thermochemical effect, we can treat it as a reaction between BS
2
3
(approximately 1:1 molar ratio) fired at 900°C for 4 h; (●) BS peaks, and
(A) Al O peaks.
2
3

2
910
Journal of the American Ceramic Society—Lee and Aswath
Vol. 83, No. 12
thermodynamically favorable. Thermodynamic calculation for re-
action (6) at 1000°C indicates that for the case where the activities
of BS, Al O , and BA are 1, the presence of SiO with an activity
2
3
2
less than 0.0587 can stabilize the system and allow for reaction (6)
to proceed with the formation of BA. Therefore, reaction (6) is
corrected as below:
BS ϩ Al O ϭ BA ϩ ͑SiO ͒
(7)
2
3
2
where (SiO ) refers to silica dissolved in a silicate phase with a
2
modest silica activity.
Comparison of Fig. 4(e) with Figs. 6(b) and (c) indicates that
Al O plays an important role in the kinetics of formation of BAS.
2
3
In all three cases, almost pure BS with Al O (approximately 1:1
2
3
molar ratio) were used and fired at 900°C for 4 h, 1000°C for 4 h,
and 1100°C for 24 h, respectively. As discussed earlier, Al O can
2
3
retard reaction (4), slowing down the reaction between B S and
2
SiO2 until Al O itself reacts with B S. At a temperature of
Fig. 5. Standard Gibbs free energy change as a function of temperature
for the indicated reactions: (a) BaCO ϩ Al O ϭ BA ϩ CO , (b) BS ϩ
2
3
2
1
000°C, Al O may be reacting with BS to form BA and a
2 3
3
2
3
2
Al O ϭ BA ϩ SiO , (c) B S ϩ Al O ϭ BS ϩ BA, (d) B S ϩ SiO ϭ
silica-bearing phase with a modest silica activity. Figure 6(c)
shows that BS and Al O react to form BA and hexacelsian BAS
when fired at 1100°C for 24 h. The overall reaction can be broken
down into reactions (7) and (8) or into reactions (7) and (9). The
formation of hexacelsian BAS is indirect evidence of the existence
2
3
2
2
2
3
2
3
2
2
BS , (e) BS ϩ SiO ϭ BS , (f) 2BS ϩ SiO ϭ B S , and (g) B S ϩ
2 2 2 2 2 3 2
2 3
SiO ϭ 2BS.
2
of SiO in the product formed by reaction (7).
Al O –SiO system the formation of BA occurs above 900°C after
2
2
3
2
all of the BaCO has been consumed to yield B S. Hence, in the
3
2
BS ϩ ͑SiO ͒ ϩ Al O ϭ hexacelsian BAS
(8)
(9)
2
2
3
BaCO –Al O –SiO system the BA is formed by a reaction
3
2
3
2
between B S and Al O , not between BaCO and Al O . A control
BA ϩ 2͑SiO ͒ ϭ hexacelsian BAS
2
2
2
3
3
2
3
experiment was conducted where a mixture of B S and Al O in a
2
2 3
molar ratio of 1:2 was calcined at 900°C for 4 h yielding BA and
BS by reaction (5), as shown in Fig. 6(a). Table I shows, from 900°
to 950°C in the BaCO –Al O –SiO system, there is no change in
In all of the experiments Al
the B S and BS remaining at 1050°C, as shown in Table I, indicate
that there must be a small amount of Al present in the sample
may be dissolved
2 3
O was exhausted by 1050°C. But,
2
O
2 3
3
2
3
2
the amount of BS related to an increase in BA and decreases in
B S and Al O . This may be because the rate for reaction (5) is
to further react with them. The undetected Al
O
2 3
in the phases of barium silicates.
2
2 3
faster than the dissociation of BS. In another control experiment
almost pure BS mixed with Al O (approximately 1:1 molar ratio)
2
3
(4) Formation of BS and Hexacelsian between 1000° and
2
was fired at 1000°C for 4 h. From the XRD spectrum shown in Fig.
(b) it appears all of the BS is reacted, yielding B S and BA, and
1
200°C
6
2
At the final stage, above 1000°C, the possible paths for
hexacelsian formation are as follows:
the unreacted Al O remains in the mixture. This could be
2
3
explained as the BS dissociation to yield B S and SiO concurrent
2
2
with the reaction of BS with Al O to yield BA and SiO2
2BS ϩ SiO ϭ B S
2 3
(10)
(11)
(12)
(13)
2
3
2
(reactions (4) and (6), respectively). In reality, at 1000°C the
suppressive effect of Al O on reaction (4) in the real BaCO –
B
2
S
3
ϩ SiO ϭ 2BS
2 2
2
3
3
Al O –SiO system will be largely alleviated by the consumption
2
3
2
BS ϩ Al O ϭ H
2
2
3
of Al O via reaction (5). That is the reason why in Table I the
2
3
amount of BS starts to increase at 1000°C. Furthermore, as shown
in Fig. 5 the standard Gibbs free energy change for reaction (6) is
positive for all temperatures calculated while reaction (5) is more
BA ϩ 2SiO ϭ H
2
According to the phase diagram shown in Fig. 1, there are four
phases—the B S , B S , B S , and BS —which are more acidic
2
3
5
8
3
5
2
silicates than BS. However, except BS the other three phases do
2
not appear in Table I. Using 5 ␮m particles of quartz instead of
amorphous SiO , we observed the formation of B S and BS ,
2
2
3
2
after B S and BS were formed, in both BaCO ⅐2SiO and
2
3
2
BaO–Al O –SiO systems. Because B5S8 and B3S5 are still
2
3
2
absent, the reaction path for the formation of hexacelsian discussed
in this section does not include these two intermediate phases.
Based on thermodynamic data in Fig. 5, reactions (10) and (11)
can produce B S and eventually BS even at lower temperatures
2
3
2
in the presence of BS. In the BaCO ⅐2SiO system BS forms at
3
2
2
9
00°C and the amount of BS2 increases with an increase in
temperature with a concurrent increase in BS and a decrease in
B S. The absence of B S in Table I may possibly be because
2
2 3
reaction (11) is faster than reaction (10). When BaCO and 5 ␮m
3
particles of quartz were mixed in a 1:2 molar ratio and calcined at
1
150°C for 4 h, pure BS2 is produced as shown in Fig. 7(a).
Fig. 6. X-ray diffraction patterns showing the formation of BA from B2S
Because quartz has a finer particle size, a choice of quartz over
amorphous SiO makes it possible to make pure BS . However, in
the BaO–Al O –SiO system the presence of Al O retards the
formation of BS . Eventually as the Al O is consumed to form
BA, reactions (10) and (11) recover to form BS above 1000°C.
or BS reacting with Al O : (a) the formation of BA from the mixture of
2
3
2
2
B S and Al O in molar ratio 1:2 calcined at 900°C for 4 h, (b) the
2
2 3
2
3
2
2 3
formation of BA from the mixture of the almost pure BS with Al O
2
3
(approximately 1:1 molar ratio) fired at 1000°C for 4 h, and (c) the
2
2 3
formation of BA and hexacelsian under the same conditions as (b), but
2
fired at 1100°C for 24 h; (ء) B S peaks, (●) BS peaks, (ϩ) BA peaks, (A)
Controlled experiments were performed to examine the role of
BS in BAS formation. When pure BS is mixed with Al O in a
2
Al O peaks, (H) hexacelsian BAS peaks, and (?) unknown peaks.
2
3
2
2
2 3

December 2000
Synthesis of Hexacelsian Barium Aluminosilicate by a Solid-State Process
2911
Fig. 7. X-ray diffraction patterns showing the formation of hexacelsian
by BS or BA: (a) pure BS made from the mixture of BaCO and quartz
2
2
3
(
SiO ) in molar stoichiometric ratio calcined at 1150°C for 4 h, (b) the
2
formation of hexacelsian from the mixture of the pure BS and Al O in the
2
2 3
molar ratio 1:1 fired at 1200°C for 6 h, (c) the formation of hexacelsian
from the mixture of BS, Al O , and SiO in the molar ratio 1:1:1 fired at
2
3
2
1
200°C for 6 h, (d) the formation of hexacelsian from the mixture of the
pure BA and SiO in the molar ratio 1:2 fired at 1200°C for 6 h, and (e) the
2
formation of hexacelsian BAS from the mixture of BaCO , Al O , and
3
2 3
SiO in the molar ratio 1:1:2 fired at 1200°C for 6 h; (?) unknown peaks,
2
(ϩ) BA, (H) hexacelsian BAS, and (h) holder.
Fig. 8. Reaction scheme of the solid-state reactions involved in the
BaCO ⅐Al O ⅐2SiO system. The numbers inside the parentheses are the
3
2
3
2
reaction numbers used in the text. The reactions are located on this chart at
the temperatures at which they occur.
1
:1 molar ratio and fired at 1200°C for 6 h, BAS is formed as
shown in Fig. 7(b). The formation of hexacelsian BAS by reaction
8) is incomplete and the reaction rate is slightly slower when a
mixture of pure BS, SiO , and Al O in a 1:1:1 molar ratio are
(
the reaction BA ϩ 2SiO
BaCO , Al , and SiO
followed by reaction with Al
reaction to form BAS.
ϭ H unavoidable when a mixture of
is used. Presintering BaCO and 2SiO
enhances the kinetics of the
2
2 3
2
fired at 1200°C for 6 h, as seen from a comparison of Figs. 7(b)
and (c). The slower kinetics in the latter case are attributed to the
multistep reactions (10), (11), and (12) for the formation of
3
2
O
3
2
3
2
O
2 3
hexacelsian BAS. When pure BA and SiO in the molar ratio 1:2
(iv) The summary of reaction path for the formation of
2
are fired at 1200°C for 6 h, BAS is formed by reaction (13).
However, the kinetics of the reaction are very slow, as evidenced
by the large amount of residual BA left behind as shown in Fig.
hexacelsian from ␥-BaCO , ␣-Al O , and amorphous SiO is
3 2 3 2
shown as Fig. 8.
1
5
7
(d). Planz et al. also reported the rate of formation of hexacel-
Acknowledgments
sian from BA and SiO is slower than that from BS and Al O .
2
2
2 3
When the starting materials are BaCO , Al O , and SiO , the
X-ray facilities provided by the Geology Department at the University of Texas at
Arlington are gratefully acknowledged. Raw materials were supplied by the Loral
Vought Systems. Also, the authors thank Dr. Sandhage for his valuable advice and Dr.
Rangarajan and Dr. Murali Hanabe for their help.
3
2
3
2
formation of BA is inevitable, as shown in Table I, and makes
reactions (13) or (9) inevitable. After the mixture of BaCO3,
Al O , and SiO in a molar ratio 1:1:2 was fired at 1200°C for 6 h,
2
3
2
Fig. 7(e) indicates the presence of BAS and large amounts of
residual BA in the sample. Therefore, it can be concluded that the
formation of BA at an intermediate step slows the kinetics of
formation of hexacelsian BAS. Given a choice, prefiring BaCO3
and SiO to yield BS followed by reaction with Al O would
References
1
C. H. Drummond III, “Glass Formation and Crystallization in High-Temperature
Glass-Ceramics and Si N ,” J. Non-Cryst. Solids, 123, 114–28 (1990).
B. Yoshiki and K. Matsumoto, “High-Temperature Modification of Barium
Feldspar,” J. Am. Ceram. Soc., 34 [9] 283–86 (1951).
3
4
2
2
2
2 3
yield hexacelsian BAS faster.
3
2 2 8
Y. Kakeuchi, “A Detailed Investigation of the Structure of Hexagonal BaAl Si O
with Reference to its ␣–␤ Inversion,” Mineral. J., 2 [5] 311–32 (1958).
4
R. E. Newnham and H. D. Megaw, “The Crystal Structure of Celsian (Barium
IV. Conclusions
i) There are no Al O –SiO compounds formed during the
Feldspar),” Acta Crystallogr., 13, 303–12 (1960).
H. C. Lin and W. R. Foster, “Studies in the System BaO–Al O –SiO I. The
2 3 2
5
(
Polymorphism of Celsian,” Am. Mineral., 53, 134–44 (1968).
2
3
2
6
D. W. Freitag, “In Situ Processed Si
3 4
N Whiskers in the System Barium
synthesis of BAS.
3 4
Aluminosilicate–Si N
,” Mater. Sci. Eng. A, [195] 197–205 (1995).
A. Bandyopadhyay, P. B. Aswath, W. Porter, and O. B. Cavin, “The Low
Temperature Hexagonal to Orthorhombic Transformation in Si N Reinforced BAS
(
ii) Al O plays an important role at the intermediate stages of
7
2
3
the reaction by forming BA at temperatures above 850°C. The
3
4
presence of Al O retards the reaction B S ϩ SiO ϭ 2BS
Matrix Composites,” J. Mater. Res., 10 [5] 1256–63 (1995).
A. Bandyopadhyay, S. W. Quander, P. B. Aswath, D. W. Freitag, K. K.
2
3
2
2
8
(reaction (4)), which is one of the intermediate steps in the
Richardson, and D. L. Hunn, “Kinetics of In-Situ ␣ to ␤ Si N Transformation in a
3
4
formation of BS , until Al O largely reacts with B S at 1000°C.
2
2
3
2
Barium Aluminosilicate Matrix,” Scr. Metall. Mater., 32 [9] 1417–22 (1995).
In addition, Al O may react with BS to form BA at 1000°C.
9
2
3
K. K. Richardson, D. W. Freitag, and D. L. Hunn, “Barium Aluminosilicate
(
iii) There are two possible routes to the formation of hexacel-
Reinforced In Situ with Silicon Nitride,” J. Am. Ceram. Soc., 72 [10] 2662–68 (1995).
I. G. Talmy, D. A. Haught, and E. J. Wuchina, “Ceramic in the System BAS–SAS
Polymorphism, Processing and Properties,” Proc. Int. SAMPE Electron. Conf., 6th,
1
0
sian BAS. One reaction is BS2 ϩ Al O ϭ hexacelsian BAS
reaction (12)) and the other is BA ϩ 2SiO ϭ hexacelsian BAS
reaction (9)). The kinetics of the latter reaction is much lower than
2
3
(
(
2
6
87–98 (1992).
11
A. Bandyopadhyay and P. B. Aswath, “A Phase Transformation Study in
that of the former. The formation of BA is inevitable and makes
BAS–Si N System,” J. Mater. Res., 10 [12] 3143–48 (1995).
3 4

2
912
Journal of the American Ceramic Society—Lee and Aswath
Vol. 83, No. 12
1
2
N. P. Bansal and M. J. Hyatt, “Crystallization Kinetics of BaO–Al
Glasses,” J. Mater. Res., 4 [5] 1257–65 (1989).
2
O
3
–SiO
2
Oxidation of Solid-Metal-Bearing Precursors”; Paper No. B-146-96 presented at the
98th Annual Meeting of the American Ceramic Society, Indianapolis, IN, April 17,
1996.
1
3
C. Liu, S. Komarneni, and R. Roy, “Crystallization and Seeding Effect in
BaAl Si Gels,” J. Am. Ceram. Soc., 78 [9] 2521–26 (1995).
H. J. Schmutzler and K. H. Sandhage, “Transformation of Ba–Al–Si Precursors
to Celsian by High-Temperature Oxidation and Annealing,” Metall. Mater. Trans. B,
6, 135–50 (1995).
O
8
19
S. Fried, S. R. Ernst, and S. S. Wang (Eds.), Thermochemical Data of Pure
2
2
1
4
Substances. VCH, Weinheim, Germany, 1993.
20
E. K. Keler and V. B. Glushkova, “The Condition of Formation of Barium
2
Silicates,” J. Inorg. Chem. (USSR), 1, 2283–93 (1956).
1
5
J. E. Planz and H. Muller-Hesse, “Investigations on Solid-State Reactions in the
System BaO–Al –SiO . Part II: The Formation of Celsian by Solid-State Reac-
tions,” Ber. Dtsch. Keram. Ges., 40, 191–200 (1963).
21
V. B. Glushkova and E. K. Keler, “The Condition of Formation of Barium
2
O
3
2
Silicates, II. Reactions between Barium Peroxide and Silicon Dioxide,” J. Inorg.
Chem. (USSR), 2, 1001–1007 (1957).
1
6
J. E. Planz and H. Muller-Hesse, “Investigations on Solid-State Reactions in the
System BaO–Al O –SiO . Part I: Solid-State Reactions in the Subsystem BaO–
2
2
A. Beauger, J. C. Mutin, and J. C. Niepce, “Synthesis Reaction of Metatitanate
,” J. Mater. Sci., 18, 3543–50 (1983).
K.-T. Lee and P. B. Aswath, unpublished work.
2
3
2
BaTiO
3
Al
2
O
3
, BaO–SiO
2
and Al
L. P. Pavlova, I. T. Ivanov, and D. P. Lepkova, “Investigation of Phase-
Formation During Celsian Synthesis,” C. R. Acad. Bulg. Sci., 34 [2] 185–8 (1981).
2 3 2
O –SiO ,” Ber. Dtsch. Keram. Ges., 38, 440–50 (1961).
2
2
3
4
1
7
L. Perier-Camby and G. Thomas, “Solid Way Synthesis of Barium Aluminate.
1
8
X. D. Zhang, S. M. Allameh, K. H. Sandhage, and H. L. Fraser, “Fabrication and
Phase Evolution of High-Temperature Monolithic Celsian and Composites by the
Kinetic Study of the Formation of Intermediate Compounds,” Solid States Ionics,
63–65, 128–35 (1993).
Ⅺ