Reactions in eucryptite-based lithium aluminum silicates

Article abstract of DOI:10.1111/j.1551-2916.2010.03618.x

The reaction sequence to synthesize β-eucryptite, LiAlSiO₄, from the raw ingredients SiO₂, Li₂CO₃, and Al₂O₃ was studied using thermal analysis and X-ray diffraction techniques. Reactions were examined by heating the raw ingredients as two-component and three-component mixtures in air, then cooling for phase analysis. In some cases, cyclic heating was performed to ensure a complete reaction. It was found that a complex sequence of reactions involving several intermediate phases occurs. The single oxides (SiO₂ and Al ₂O₃) react with Li₂CO₃ to form the binary oxides (Li₂Si₂O₅, Li₂SiO ₃, and LiAlO₂); SiO₂ reacts with Li ₂CO₃ before Al₂O₃. Subsequently, the binary oxides form ternary oxides (LiAlSi₂O₆ and LiAlSiO₄). Finally, the ternary oxides form the equilibrium product, β-eucryptite (LiAlSiO₄). These reactions are discussed in the context of the thermodynamic properties of the various compounds.

Full text of DOI:10.1111/j.1551-2916.2010.03618.x

J. Am. Ceram. Soc., 93 [6] 1591–1596 (2010)
DOI: 10.1111/j.1551-2916.2010.03618.x
r 2010 The American Ceramic Society
ournal
J
Reactions in Eucryptite-Based Lithium Aluminum Silicates
Timothy Jochum and Ivar Reimanisw
Colorado Center for Advanced Ceramics, Metallurgical and Materials Engineering Department, Colorado School of
Mines, Golden, Colorado 80401
The reaction sequence to synthesize b-eucryptite, LiAlSiO4,
The mechanism by which this reaction occurs has not been
established. In air, the least stable of the three reactants in Re-
action (1) is Li CO . Its decomposition in air is, as described by:
from the raw ingredients SiO , Li CO , and Al O was
2
2
3
2
3
studied using thermal analysis and X-ray diffraction techniques.
Reactions were examined by heating the raw ingredients as two-
component and three-component mixtures in air, then cooling
for phase analysis. In some cases, cyclic heating was performed
to ensure a complete reaction. It was found that a complex se-
quence of reactions involving several intermediate phases occurs.
The single oxides (SiO and Al O ) react with Li CO to form
2
3
Li2CO3 ! Li2O þ CO2ðgÞ (2)
Reaction (2) is thermodynamically favorable when the concen-
tration of CO decreases below a critical value that depends on the
2
temperature. For example, using the thermodynamic properties
2
2
3
2
3
9
for the compounds in Reaction (2), at 7271C, the critical partial
the binary oxides (Li Si O , Li SiO , and LiAlO ); SiO reacts
2
2
5
2
3
2
2
with Li CO before Al O . Subsequently, the binary oxides
z
pressure of CO in air is about 0.13 torr. Our calculations with the
program HSC Chemistry 5.1 (Outokumpo Research Oy, Pori,
Finland) on the relative thermodynamic stability of Li CO in
2
2
3
2
3
form ternary oxides (LiAlSi O6 and LiAlSiO ). Finally,
2
4
2
3
the ternary oxides form the equilibrium product, b-eucryptite
LiAlSiO ). These reactions are discussed in the context of the
thermodynamic properties of the various compounds.
flowing air (B1 L/min) predict that decomposition begins to occur
at about 6001C, well below the melting temperature of Li CO
in the three-
component mixture described in Reaction (1) decomposes at rel-
atively low temperatures and produces Li O, which is highly re-
active with Al and SiO , as described in more detail below.
It is well established that Al and SiO do not react with
each other for the temperatures and times used to calcine
(
4
2
3
1
0
m 2 3
(T 57231C ). Thus, it is expected that the Li CO
I. Introduction
2
2
O
3
2
4
b-EUCRYPTITE, LiAlSiO , has interest in a variety of applications
O
2 3
2
ranging from low coefficient of thermal expansion (CTE) ma-
1
–3
terials to superionic conductors. This material has a hexago-
nal crystal structure and belongs to the space group, P6 22.
Polycrystalline b-eucryptite exhibits a slightly negative average
11,12
b-eucryptite,
involve Li CO
research by Smirnov et al. describes the physical behavior of
and thus the first reactions expected to occur
or its decomposition product, Li O. Previous
4
2
3
2
1
3
ꢁ
6
CTE, resulting from the thermal anisotropy (a 5 8.6 ꢀ 10
a
Li CO with regard to the decomposition toward Li O. The
2
3
2
ꢁ
1
, a
ꢁ6
ꢁ1 4,5
1C
c
5 ꢁ18.4ꢀ 10 1C
) between the two main crystal-
2 3 2
Li CO –Li O system forms a eutectic at approximately 7051C
lographic directions of the hexagonal a-axis and c-axis. Com-
posites with near-zero CTE may be designed by combining
positive CTE materials with b-eucryptite. Aside from the un-
usual thermal properties, b-eucryptite also undergoes a pressure-
induced phase transformation at ambient temperatures. The
transformation has been observed by in situ synchrotron
X-ray diffraction (XRD) as well as in situ Raman spectros-
2
and 13 mol% of Li O. Thus, a reactive liquid is expected to form
slightly below the Li CO melting point when the three reactants
2
3
in Eq. (1) are heated. The formation of lithium aluminates and
However, it is
14,15
lithium silicates has been studied previously.
not well understood what sequence of reactions are present for
the three-component mixture, as shown in Reaction (1).
The present study is motivated by the desire to determine and
ultimately control the calcination process in the synthesis of
b-eucryptite. The three different two-component raw mixtures
comprising the reactants in Reaction (1) were prepared and
characterized with thermal analysis and XRD, with the goal of
better understanding how the two-component reactions may in-
fluence reactions in the three-component raw system.
6
,7
copy. This unique behavior motivates fundamental studies to
understand better the material structure–property relationships.
Processing of b-eucryptite and its composites is typically
achieved by either a glass ceramic technique, whereby eucryptite
3
crystals are formed within a glass matrix during ceramming, or by
5,6
powder synthesis and subsequent powder metallurgy. In the
latter case, control over stoichiometry, purity, and powder char-
acteristics such as morphology, size, and size distribution is im-
portant as these factors will influence the sintering behavior as well
as the final properties. Such a control may only be achieved if the
sequence of reactions during the powder synthesis is understood.
II. Experimental Procedure
(
1) Powder Preparation
Raw powders were mixed with a specific composition target of
LiAlSiO . As provided by Reaction (1), the ratio of Li CO ,
5
,6,8
Previous studies
the appropriate ratios of Li
about 12001C according to the following reaction:
have shown that b-eucryptite forms when
2
CO , Al , and SiO are heated to
3
O
2 3
2
4
2
3
Al O , and SiO is [1:1:2], respectively. The mass ratio corre-
2
2
3
sponding to this molar ratio is [1:1.38:1.63]. Lithium carbonate
(
2 3
Li CO —Alfa Aesar, Ward Hill, MA), aluminum oxide
Li2CO3 þ Al2O3 þ 2SiO2 ! 2LiAlSiO4 þ CO2ðgÞ
M. Rigaud—contributing editor
(1)
(
Al O —CoorsTek, Golden, CO), and silicon oxide (SiO —
2
2
3
Sigma Aldrich, St. Louis, MO) were weighed and then mixed
to achieve the ratio [1:1:2]. Because the Sigma Aldrich SiO
2
y
2
contains excess H O, the mass ratio used to weigh the powders
z
In ambient air in Golden, Colorado, where the present experiments were performed,
. For comparison, the average
concentration in the Earth’s atmosphere in 2009 is about 387 ppm (source: http://
www.esrl.noaa.gov/gmd/ccgg/trends/).
yIt was determined in separate thermogravimetric experiments conducted by the authors
that the Sigma Aldrich contains about 9% by mass H O which is bonded to SiO ; all of the
excess H O leaves as a vapor when the specimen is heated to 4001C, and thus the H O is
thought not to participate in any higher temperature reactions.
this corresponds to about 200 parts per million (ppm) CO
CO
2
2
Manuscript No. 26940. Received October 9, 2009; approved December 14, 2009.
This work was supported by the National Science Foundation Ceramics Program Grant
DMR-0746086 (TJ) and the U.S. Department of Energy’s Office of Basic Energy Sciences
Grant #DE-FG02-07ER46397 (IR).
#
2
2
2
2
wAuthor to whom correspondence should be addressed. e-mail: reimanis@mines.edu
1
591

1
592
Journal of the American Ceramic Society—Jochum and Reimanis
Table 1. List of Prepared Samples
Vol. 93, No. 6
Material/mixture
Molar ratio
Li CO3
2
Al O3
2
Pure
Pure
SiO2
Li CO –Al O
3
Pure (plus water)
[1:1]
2
3
2
Li CO –SiO
2
[1:2]
[1:2]
2
3
Al O –SiO
2
2
3
Li CO –Al O –SiO
2
[1:1:2]
[1:1:2]
2
3
2
3
Li O–Al O –SiO
2
2
2
3
was [1:1.38:1.85]. The particle sizes of the raw materials are:
Li CO , B1 mm; Al , B1 mm; and SiO , 50–70 mm. For the
2
3
O
2 3
2
present study, two-component mixtures were studied to isolate
specific reactions. Three two-component raw mixtures and one
three-component raw mixture were prepared, approximately 30
g of each, as shown in Table I. One three-component raw mix-
Fig. 1. The differential scanning calorimetry/thermogravimetric analy-
sis data from lithium carbonate exhibiting a sharp melting peak and a
2
broad CO loss doublet, mass loss becomes significant just before the
ture sample was prepared using lithium oxide (Li
2
O—Alfa
doublet at B9001C.
Aesar) rather than Li CO . Each powder mixture was blended
2
3
by ball milling with aluminum oxide media for 24 h to ensure
proper dispersion before analysis. Scanning electron microscopy
with energy-dispersive spectroscopy was performed on the
mixed powders to ensure intimate mixing.
that exhibited significant mass loss was Li CO . The DSC data
2
3
from the two-component raw mixtures are shown in Figs. 3–5.
The two-component raw mixtures that contained Li CO not
only exhibited peaks corresponding to the melting of Li CO
2
3
2
3
,
(2) Differential Scanning Calorimetry (DSC) and XRD
Analysis
but also showed subsequent chemical reactions. For example,
Li CO –Al O leads to three DSC peaks at approximately 800,
2
3
2
3
DSC was used to determine the initial energy profiles of each
raw powder, two-component raw mixtures, and the three-com-
ponent raw mixtures. A Netzsch STA 409C (Netzsch Instru-
ments Inc., Burlington, MA) was used to gather DSC and
thermogravimetric analysis (TGA) data concurrently. High-
2 3 2
950, and 11001C, and the Li CO –SiO mixture exhibits endo-
thermic peaks at approximately 10001 and 10501C. The dashed
lines on these diagrams indicate the temperatures from which
specimens were cooled with the goal of performing XRD to es-
tablish the phases present. The DSC data from the two-compo-
purity aluminum oxide crucibles (purity level 99.7% Al
2 3
O )
nent raw mixture of Al
12001C (Fig. 5). It is apparent from Figs. 3 and 4 that the mass
loss in two-component raw mixtures containing Li CO occurs
at a substantially lower temperature. For example, when Li CO
is heated with Al , the mass loss is significant at about 8001C,
and when it is combined with SiO , the mass loss occurs below
8001C. In contrast, Li CO alone exhibited significant mass loss
2 3 2
O and SiO revealed no reaction up to
obtained from Netzsch Instruments Inc. were used to contain
the specimens. Platinum crucibles, also obtained from Netzsch
Instruments Inc., were also used in some cases to confirm that
the aluminum oxide crucibles did not contaminate the speci-
mens. The heating profile for each run began at 201C and was
heated to 12001C at 101C/min. All experiments were performed
in air that flowed through the apparatus at a rate of about
2
3
2
3
O
2 3
2
2
3
at temperatures above 9001C.
1
liter/minute. Because of the small specimen size required for
DSC of the three-component mixture of raw materials ex-
hibited multiple peaks, similar to the two-component raw mix-
the DSC, no soak at temperature is required. The raw mixtures
were thermally cycled either twice or thrice to ensure a complete
reaction during analysis.
Data gathered from the DSC instrument were used to gen-
erate a set of furnace treatments for each of the raw mixtures.
The heat treatment procedure for each sample followed the
same format. Furnace conditions consisted of a ramp rate of
51C/min to the target temperature with a soak for 3 h followed
by a furnace cool. Sufficient powder was used so that enough
material was available for accurate XRD analysis (2–5 g). The
XRD analysis was completed using a Philips X’Pert Pro
Diffractometer (PANalytical Inc., Westborough, MA) with a
˚
copper source, with l 5 1.54 A, and a nickel filter. The detector
used for all of the XRD analyses was an X’Celerator (PANa-
lytical Inc.). The XRD data were characterized using both peak
and profile data and a reference International Centre for
Diffraction Data (ICDD) database.
III. Results
(1) DSC Characterization
DSC of the raw powders is shown in Figs. 1 and 2. Both Al
O
2 3
and SiO are stable across the temperature range examined. The
2
Al O exhibited no mass loss and the SiO exhibited a minor
2
2
3
Fig. 2. The differential scanning calorimetry data from Al
2
SiO with no significant peaks. The thermogravimetric analysis data
were omitted because a significant mass loss was not measured in either
sample.
2 3
O and
mass loss due to water. However, Li
2
CO
3
demonstrated a sharp
melting peak and a set of peaks, which correspond to the loss of
1
CO , consistent with previous studies. The only raw material
3
2

June 2010
Reactions in Eucryptite
1593
Fig. 3. The differential scanning calorimetry/thermogravimetric analy-
sis data from the two-component mixture of Li CO :Al showing the
2
3
2 3
O
major reaction peak at B7401C and the corresponding heat treatment
temperatures of 7801, 8501, 10001, and 11001C.
Fig. 5. The differential scanning calorimetry data from the two-com-
ponent mixture of Al O :SiO with no reaction peaks. The thermogravi-
2 3 2
metric analysis data were omitted because they revealed only a minor
mass loss due to water.
tures, shown in Fig. 6. There are two main endothermic peaks;
the first occurred at B7001C and the second at B10001C. The
three-component raw mixture that contained Li
similar endothermic peak at B10001C. However, the Li
ture revealed an exothermic peak at B6001C in contrast to the
endothermic peak observed in the Li CO -containing three-
component mixture (Fig. 7). The mass loss for the Li O mix-
2
O contained a
Li SiO
2
3
and Li
2
Si
2
O
5
. Before the first endothermic peak seen in
SiO were found in the heat
2
O mix-
the DSC, trace amounts of Li
2
3
treatment at 5851C. Based on differences in peak heights in
Fig. 9, it is deduced that the heat treatment at 11001C led to
differences in the relative quantity of phases present from
the heat treatments at 9501 and 10101C. However, the qualita-
tive phase analysis is identical to the previous heat treatments.
The difference in peak heights is likely due to the melt/solidifi-
cation of the powder during this heat treatment, which may lead
to preferential ordering of the material. The XRD data from the
two-component mixture of Al O and SiO after heat treatment
2
3
2
ture is negligible and attributed to the bonded water within the
SiO powder. The dashed lines in Fig. 6 pertain to the temper-
2
atures used to heat treat the powder mixture. Table II provides a
comprehensive list of each temperature used for the heat treat-
ment processes of the two-component mixtures. The three-com-
ponent mixture heat treatments are described in Table III.
2
3
2
(
to the starting constituents, consistent with published phase
not shown here) revealed only two phases, which are identical
1
1,12
(2) XRD Characterization
diagrams.
Included in Table II is the phase information
XRD was performed on each specimen that was heat treated to
the temperatures indicated by the dashed lines in Figs. 4–6, and
these data are shown in Figs. 8–10. The diffraction data for the
Li CO –Al O mixture confirmed the existence of the equimolar
gathered from the XRD analysis for each mixture. The two-
component mixtures containing Li CO :SiO and Li CO :Al O
contained a small fraction of the raw materials, SiO and Al O
respectively, even after exposure to 12001C and were excluded
from Table III.
2
3
2
2
3
2
3
3
2
2
2
3
2
3
compound LiAlO
and 11001C show LiAlO as a pure phase. The two-component
2
mixture of Li CO –SiO produced a mixture of two phases,
2
above 8501C. The heat treatments at 10001
2
The XRD data gathered from each of the heat treatment
temperatures for the Li CO -containing three-component
2
3
2
3
Fig. 4. The differential scanning calorimetry/thermogravimetric analy-
sis data from the two-component mixture of Li CO :SiO showing a
major reaction peak at B7301C and the corresponding heat treatments
at 7551, 9501, 10101, 11001C. The significant peak at 10501C was deter-
mined to be the melting of one of the species present.
Fig. 6. The differential scanning calorimetry/thermogravimetric analy-
sis data from the three-component mixture of Li CO :Al :SiO show-
ing major reaction peaks at B7001 and B10001C. The corresponding
heat treatment temperatures are shown here and listed in Table III.
2
3
2
2
3
O
2 3
2

1
594
Journal of the American Ceramic Society—Jochum and Reimanis
Vol. 93, No. 6
Table II. Two Component Raw Mixture Heat Treatments
Heat Treatment
Phases
Mixture
temperatures (1C)
identified
Li CO –Al O
3
780, 850, 1000, 1100
755, 950, 1010, 1100
None
LiAlO2
Li SiO , Li Si O
—
2
3
2
Li CO –SiO
3
2
2
2
3
2
2
5
Al O –SiO
2
2
3
calculated free energy of reaction at 7001C
Li2CO3 ! Li2O þ CO2ðgÞ DG ¼ 18:07 kcal
Li2O þ 2SiO2 ! Li2Si2O5 DG ¼ ꢁ35:16 kcal
Li2O þ SiO2 ! Li2SiO3 DG ¼ ꢁ33:83 kcal
Li2O þ Al2O3 ! 2LiAlO2 DG ¼ ꢁ29:02 kcal
(6)
(7)
(8)
(9)
Fig. 7. The differential scanning calorimetry/thermogravimetric analy-
sis data from the three-component mixture of Li O:Al :SiO exhibit-
2
O
2 3
2
ing an exothermic 6001C peak and an endothermic 10001C peak and no
significant mass loss.
2 3
The decomposition of Li CO must result in the formation of
a reactive lithium-containing species, which is likely Li O (Re-
2
action (2)). This decomposition to Li O and CO correlates with
2
mixture expose multiple compounds that appear and disappear
during the calcination process. The major phases in the heat
treatments at higher temperatures are the compounds contain-
2
the specimen mass loss (shown in Figs. 1, 3, 4, and 6). Similarly,
the reactions indicated by Reactions (2)–(5) are also accompa-
nied by the loss of CO (g). By comparing the DSC data from
ing all three cations (Li, Al, Si): b-spodumene (LiAlSi
2 6
O ) and
2
2 3
the three-component raw mixture containing Li CO to the
b-eucryptite. Table III shows the transformation from raw ma-
terials to the existence of ternary oxide species, and ultimately
pure LiAlSiO4.
three-component raw mixture with Li O, it is concluded that
2
the endothermic reactions (Reactions (3)–(5)) are favored over
the exothermic Li O-based reactions (Reactions (6)–(8)). How-
2
ever, the net free energy for the decomposition of Li
2
CO
O reactions (Reactions
6)–(8)) are very similar to the values for the direct Li CO
3
(
(
Reactions (2)) and the subsequent Li
2
IV. Discussion
2
3
The experiments presented above reveal that a complex set of
intermediate reactions occurs during the calcination described
by Reaction (1). These are discussed next in the context of the
thermodynamic properties of the various phases.
reaction (Reactions (3)–(4)) making them indistinguishable ex-
perimentally. Despite the difference in mechanism between the
two synthesis routes, both sets of reactions involve the release of
CO (g) and produce identical intermediate compounds.
2
The intermediate species formed during the synthesis of
b-eucryptite are not the only possible reaction products. Lith-
2 3
During heating, all mixtures containing Li CO exhibit an
endothermic peak at about 7301C, which corresponds to the
1
melting of Li CO . It is further evident that at about the same
0
2
3
5 8 5 4
ium aluminate species such as LiAl O and Li AlO are
temperature (i.e., B7301C), a chemical reaction occurs when
Li CO is present. Based on the DSC data from the different
two-component mixtures (Figs. 3 and 4), the product of the
thermodynamically favored as well (DG 5 ꢁ9.66 and 5 ꢁ9.26
kcal, respectively). These species were not found during the
XRD analysis and were thus eliminated as intermediate com-
pounds. The absence of these products suggests that the lithium
compounds (i.e. Li CO or Li O) react with the SiO completely
2
3
reaction that occurs at B7301C is either LiAlO or Li SiO ,
2
2
3
depending on whether Li CO
2
3
is present with Al
O
2 3
or SiO
2
.
2
3
2
2
XRD analysis of the heat treatments from the three-component
mixtures shows that the reaction with SiO occurs at a lower
before the reaction with alumina begins. It is recognized that
local equilibrium conditions may vary, leading to the formation
of compounds not thermodynamically predicted considering the
overall composition.
2
2 3
temperature. In fact, relatively small amounts of Li SiO are
even detected in the three-component mixture after a heat treat-
ment of only 5851C. In this same temperature range where
Li CO melts and chemically reacts, a significant mass loss is
In the present experiments, the aluminum-containing
2
3
2 6
lithium silicates, b-spodumene (LiAlSi O ), and b-eucryptite
2
observed. The mass loss is likely due to the release of CO , which
(
mation corresponds to the second endothermic peak in the
LiAlSiO ), only formed above about 10001C. Their for-
4
might occur by several mechanisms, as described below. In con-
trast, pure Li CO decomposes at a significantly higher temper-
ature. Figure 11 reproduces the TGA data for pure Li CO and
2
3
three-component raw mixture DSC data (B10001C, Fig. 6).
2
3
the various mixtures. In the light of these observations, the fol-
lowing are the possible reactions. The free energies of formation
were calculated using HSC Chemistry 5.1 software at 7001C,
and are indicated
Table III. Heat Treatments—Li CO Three Component
3
2
Raw Mixture
Temperature
1C)
Li2CO3 þ 2SiO2 ! Li2Si2O5 þ CO2ðgÞ
(
Phases Detected
(3)
(4)
(5)
DG ¼ ꢁ17:10 kcal
RT
585
750
Li CO , SiO , Al O
2 3 2 2 3
Li CO , SiO , Al O , Li SiO
2 3 2 2 3 2
3
2
2
Li2CO3 þ SiO2 ! Li2SiO3 þ CO2ðgÞ
DG ¼ ꢁ15:75 kcal
SiO , Al O , Li SiO , LiAlO
2 2 3 2 3
SiO , Al O , Li SiO , LiAlO
2 2 3 2 3
8
9
00
50
SiO , Al O , Li SiO , Li Si O , LiAlO
2
2
2
3
2
3
2
2
5
Li2CO3 þ Al2O3 ! 2LiAlO2 þ CO2ðgÞ
DG ¼ ꢁ11:13 kcal
1
050
SiO , Al O , Li SiO , Li Si O , LiAlSi O ,
2 2 3 2 3 2 2 5 2 6
LiAlSiO4
SiO , Al O , Li SiO , LiAlSi O , LiAlSiO
4
1
1
100
2
2
3
2
3
2
6
The following reactions are also considered in the context of
Reaction (2), which has been reproduced again along with the
100 (15 h) LiAlSiO4

June 2010
Reactions in Eucryptite
1595
°
°
°
°
Fig. 8. The comparison of Li
tion data showing the formation of only LiAlO
Al (J).
2
CO
3
:Al
2
O
3
heat treatment X-ray diffrac-
2
(
ꢂ
) at the expense of
O
2 3
Fig. 10. The comparison of the three-component mixture, containing
Li CO , heat treatment X-ray diffraction data, the onset of binary and
ternary oxides are marked with ^ and , respectively. The uppermost
2
3
ꢂ
scan, labeled pure, is the result of a full calcination of this mixture.
The possible reactions to produce these include:
Li2SiO3 þ SiO2 þ Al2O3 ! 2LiAlSiO4
DG ¼ ꢁ14:87 kcal
(
10)
Li2Si2O5 þ Al2O3 ! 2LiAlSiO4 DG ¼ ꢁ12:85 kcal (11)
Li2SiO3 þ Al2O3 þ 3SiO2 ! 2LiAlSi2O6
(
12)
DG ¼ ꢁ12:14 kcal
The product of Reaction (11) may react with the product of
Reaction (8) to produce b-eucryptite according to the HSC cal-
culations:
LiAlO2 þ LiAlSi2O6 ! 2LiAlSiO4
(
13)
DG ¼ ꢁ10:37 kcal
Fig. 11. The comparison of thermogravimetric analysis data from
Li
responding to CO
tween Li CO and SiO
reaction between Li CO
2
CO
3
-containing mixtures showing the earlier onset of mass loss cor-
leaving the system. This illustrates the reaction be-
occurring at a lower temperature than the
and Al
2
2
3
2
2
3
2 3
O .
This final reaction is revealed in the DSC data between the
0501C and the 11001C heat treatments (Fig. 6). The progres-
1
sion of Reaction (12) was also confirmed by the XRD patterns
of samples heat treated at these two temperatures (Fig. 9).
The published and referenced procedure for synthesizing
5
,6,8
b-eucryptite by calcination calls for 15 h at 11001C.
This
soak time at 11001C allows any nonequilibrium products to
reach their equilibrium counterparts, and thus b-eucryptite as a
pure phase.
V. Conclusion
The present study illuminates the reaction sequence for the ca-
Fig. 9. The comparison of Li
tion data, which shows the formation of the two lithium silicate phases.
The phases are marked as follows: Li SiO (^), Li Si (.), and SiO
X). The difference in appearance of the 11001C scan is likely due to
melting of the sample.
2 3 2
CO :SiO heat treatment X-ray diffrac-
lcination of Li CO , SiO , and Al O to produce b-eucryptite.
2
3
2
2
3
By using DSC in combination with XRD, it is possible to draw
the following conclu‘sions. The reaction sequence for the given
set of raw materials (Li CO , SiO , and Al O ) is as follows: The
2
3
2
O
2 5
2
(
2
3
2
2
3

1
596
Journal of the American Ceramic Society—Jochum and Reimanis
Vol. 93, No. 6
6
J. Zhang, A. Celestian, J. B. Parise, H. Xu, and P. J. Heaney, ‘‘A New
), e-Eucryptite, and Thermal Expansion of a-
single oxides (SiO
binary oxides (Li Si
Li CO before Al O . Subsequently, the binary oxides form ter-
2
and Al
2
O
3
) react with Li
2
CO
3
to form the
Polymorph of Eucryptite (LiAlSiO
4
2
O , Li
2 5 2
SiO , and LiAlO ), SiO
3
2
2
reacts with
and e-Eucryptite at High Pressure,’’ Am. Miner., 87, 566–71 (2002).
T. Jochum, I. E. Reimanis, M. J. Lance, and E. R. Fuller Jr., ‘‘In Situ Raman
7
2
3
2
3
nary oxides (LiAlSi
ides form the equilibrium product, b-eucryptite (LiAlSiO ).
4
2
O
6
and LiAlSiO ). Finally, the ternary ox-
4
Indentation of b-Eucryptite: Characterization of the Pressure-Induced Phase
Transformation,’’ J. Am. Ceram. Soc., 92 [4] 857–63 (2009).
H. Xu, P. J. Heaney, D. M. Yates, R. B. Von Dreele, and M. A. Bourke,
8
‘
‘Structural Mechanisms Underlying Near-Zero Thermal Expansion in b-Eucryp-
tite: A Combined Synchrotron X-Ray and Neutron Rietveld Analysis,’’ J. Mater.
Res., 14, 3138–51 (1999).
I. Barin and G. Platzki, Thermochemical Data of Pure Substances. Wiley and
Acknowledgment
9
The authors would also like to acknowledge the help of Michael Sanders for his
expertise with the DSC and the resulting discussions, and Dr. Gerald Martins for
helpful discussions.
Sons Inc., Hoboken, NJ, 1997.
D. R. Lide (Ed.), CRC Handbook of Chemistry and Physics. CRC Press, Boca
Raton, FL, 2004.
10
11
J. F. MacDowell and G. H. Beall, ‘‘Immiscibility and Crystallization in
–SiO Glasses,’’ J. Am. Ceram. Soc., 52 [1] 17–25 (1969).
R. F. Davis and J. A. Pask, ‘‘Diffusion and Reaction Studies in the System
Al O –SiO ,’’ J. Am. Ceram. Soc., 55 [10] 525–31 (1972).
Al
2
O
2
3
2
References
1
1
G. Orcel, J. Phalippou, and L. L. Hench, ‘‘Processing and Structural Evolution
2
1
3
2
3
of (xLi
2
Oꢃ (1ꢁx)Na
U. v. Alpen, H. Schulz, G. H. Talat, and H. Bo
Cooperative Li-Diffusion in b-Eucryptite,’’ Solid State Commun., 23, 911–4 (1977).
2
O) ꢃ Al
2
O
3
ꢃ 2SiO
2
Gels,’’ J. Non-Cryst. Solid, 82, 301–6 (1986).
M. V. Smirnov, I. Ya. Lyubimtseva, L. A. Tsiovkina, and N. Yu., ‘‘Phase
2
¨
hm, ‘‘One Dimensional
Diagram of Lithium Carbonate-Lithium Oxide System,’’ Krasnov, Zh. Neorg.
Khim.—Russ. J. Inorg. Chem. (Engl. Transl.), 16 [1] 130–31 (1971).
L. P. Cook, E. R. Plante, R. S. Roth, and J. W. Hastie. ‘‘Phase Equilibria of
3
14
W. Ho
¨
land and G. Beall, Glass Ceramic Technology. the American Ceramic
Society, Westerville, OH, 2002.
W. W. Pillars and D. R. Peacor, ‘‘The Crystal Structure of Beta Eucryptite as a
Stored Chemical Energy Reactants’’; pp. 1–103. Annual Report to the Office of
Naval Research for the Period May 25, 1983—May 25, 1984, Contract No.
N00014-83-F-0117, Department of Navy, Office of Naval Research; Arlington,
Virginia, 1984.
4
Function of Temperature,’’ Am. Miner., 58, 681–90 (1973).
A. I. Lichtenstein, R. O. Jones, H. Xu, and P. J. Heaney, ‘‘Anisotropic Thermal
5
15
Expansion in the Silicate b-Eucryptite: A Neutron Diffraction and Density Func-
tional Study,’’ Phys. Rev. B, 58, 6219–23 (1998).
2 2
F. C. Kracek, ‘‘The Binary System Li O–SiO ,’’ J. Phys. Chem., 34 [12] 2641–
50 (1930).
&