Effect of solvents on the preparation of lithium aluminate by sol-gel method

Article abstract of DOI:10.1016/j.materresbull.2003.10.022

γ-Lithium aluminate was prepared by sol-gel method using lithium methoxide and aluminum-sec-butoxide precursors in i-propanol, n- and tert-butanol. Clear gels could be obtained due to the addition of ethylacetoacetate and the dried solids were calcined at 550 and 900°C. The resulting solids were characterized by X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FT-IR), scanning electron microscopy (SEM), thermo-gravimetric analysis/differential thermal analysis (TGA/DTA). γ-Lithium aluminate with the highest purity was obtained with t-butanol solvent and LiAl₅O₈ was the second major phase.

Full text of DOI:10.1016/j.materresbull.2003.10.022

Materials Research Bulletin 39 (2004) 715–724
Effect of solvents on the preparation of lithium aluminate
by sol–gel method
*
Faruk Oksuzomer, S. Naci Koc, Ismail Boz , M. Ali Gurkaynak
Department of Chemical Engineering, Faculty of Engineering, Istanbul University, Avcilar, 34160 Istanbul, Turkey
Received 9 August 2002; received in revised form 16 January 2003; accepted 26 October 2003
Abstract
g-Lithium aluminate was prepared by sol–gel method using lithium methoxide and aluminum-sec-butoxide
precursors in i-propanol, n- and tert-butanol. Clear gels could be obtained due to the addition of ethylacetoacetate
and the dried solids were calcined at 550 and 900 8C. The resulting solids were characterized by X-ray diffraction
(
XRD), Fourier transform infrared spectroscopy (FT-IR), scanning electron microscopy (SEM), thermo-
gravimetric analysis/differential thermal analysis (TGA/DTA). g-Lithium aluminate with the highest purity was
obtained with t-butanol solvent and LiAl O was the second major phase.
5
8
#
2004 Elsevier Ltd. All rights reserved.
Keywords: A. Oxides; B. Sol–gel chemistry; C. X-ray diffraction
1
. Introduction
Lithium aluminate, g-LiAlO , has gained attention for its potential use as irradiation blanket for
2
nuclear fusion reactors. It can also be used as ceramic carrier material in molten carbonate fuel cell due
to its high mechanical and thermal stability [1,2]. These potential usages make the synthesis of the
highest purity g-LiAlO an important issue. The synthesis of pure g-LiAlO by a modified sol–gel
2
2
method is an alternative route to the conventional solid-state reaction [3].
Lithium aluminate has three allotropic forms and they are named a-, b-, and g-LiAlO with
2
hexagonal, monoclinic and tetragonal structures, respectively. g-LiAlO is the most stable form and a-,
2
b-LiAlO phases transform to g-LiAlO at elevated temperatures [3]. In the synthesis of g-LiAlO ,
2
2
2
calcination above 1000 8C is considered to be ‘‘high temperature calcination’’ and below this threshold
it is regarded to be ‘‘low temperature calcination’’ [9]. Preparation of lithium aluminate by the
conventional solid-state fusion method, which requires the high calcination temperatures necessary for
*
Corresponding author. Tel.: þ90-212-473-70-70x17665; fax: þ90-212-591-19-97.
E-mail address: ismailb@istanbul.edu.tr (I. Boz).
0025-5408/$ – see front matter # 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.materresbull.2003.10.022

7
16
F. Oksuzomer et al. / Materials Research Bulletin 39 (2004) 715–724
the solid-state reactions of lithium salt and aluminate, was reported in earlier works [4–7]. Recently,
g-LiAlO was also prepared by fusion and sol–gel methods [8], and it was found that the structural,
2
morphological, and thermal properties of lithium aluminate depended strongly on the preparation method.
The effect of heat treatment on the structure of the lithium aluminate was also studied [9] and it was found
that after calcinations above 750 8C although both g-LiAlO and LiAl O phases existed when it was
2
5 8
prepared by sol precipitation, but pure g-LiAlO could be obtained at this temperature with sol–gel
2
method [9]. Thus, we focused on sol–gel method to improve the purity of g-LiAlO . In this study, the
2
effects of employment of different alcohols, i-propanol, n-butanol and t-butanol from lithium methoxide
and aluminum-sec-butoxide in the presence of ligand aid, ethylacetoacetate, were investigated.
2
. Experimental
2
.1. Preparation of lithium aluminate by sol–gel method
The reagents used in this study were products of Merck AG and they were used without further
purification. Equimolar amounts of lithium methoxide and aluminum-sec-butoxide were utilized as
precursors, and synthesis grade solvents, such as, i-propanol (i-Pr), i-butanol (i-But) and t-butanol
(t-But), were used in preparations. Ethylacetoacetate (Eaa) was also used in all mixtures as ligand aid to
improve gelation and prevent early precipitation [10,11].
Lithium-methoxide (0.05 mol), Al-sec-butoxide (0.05 mol) and ethylacetoacetate (0.025 mol) were
mixed with 50 ml of alcohol solvent. Li:Al:Eaa molar ratios of 1:1:0.5, respectively, were kept
constant. 50 ml of water was added to start gelation. A scheme for the sol–gel preparation is illustrated
in Fig. 1. Clear gels were obtained in 2 days at room temperature and resulting gels were dried at
Fig. 1. The preparation scheme for g-lithium aluminate by sol–gel method.

F. Oksuzomer et al. / Materials Research Bulletin 39 (2004) 715–724
717
120 8C under air flow for 8 h. The obtained powders were calcined at 550 and 900 8C. The powders
were fluffy and easily crushable fine particles.
2
.2. Characterization
X-ray diffractograms were obtained in a Shimadzu X-ray diffractometer (XRD), using
˚
(
l ¼ 1:5404 A) Cu Ka radiation. TGA analysis was performed in a Shimadzu 50 instrument. A
heating rate of 10 8C/min up to 1100 8C was applied under air flow of 100 ml/min. DTA analysis was
performed in a Linseis instrument (L62) and a-Al O was used as reference. Samples in DTA analysis
2
3
were heated from room temperature to 1200 8C with a heating rate of 10 8C/min under air flow of
00 ml/min. Infrared spectra were obtained in a Mattson 1000 Fourier transform infrared
1
spectrophotometer. KBr discs were used in the transmission mode. Morphology of calcined samples
was determined by a JEOL JSM 5600 scanning electron microscope at 10 kV.
3
. Results and discussion
Crystal phases of the dried powders calcined at 550 8C, prepared with different solvents, are shown
in Fig. 2. The predominant phases were LiAl (OH) ꢀH O (JCPDS/31-704) and LiOHꢀH O (JCPDS/
2
7
2
2
2
5-486). Peaks of these two phases were comparatively sharper in t-butanol than those for the other
two solvents. g-LiAlO was not yet present in large quantities at this low calcination temperature.
2
X-ray diffractograms of dried powders calcined at 900 8C are shown in Fig. 3. In the diffractograms,
the predominant phase was g-LiAlO (JCPDS/38-1464). LiAl O (JCPDS/38-1425) was also present
2
5 8
as second phase. g-LiAlO peaks were sharper and more intense in t-butanol solvent than those for
2
the other two solvents. It is worth mentioning that in all solvent systems, minor amounts of
LiAl (OH) ꢀH O remained at 900 8C. Kwon and Park [3] also reported that LiAl (OH) ꢀH O did not
2
7
2
2
7
2
disappear even at 1000 8C. In the gels obtained with these solvent systems, g-LiAlO was
2
Fig. 2. XRD diagrams of samples calcined at 550 8C (&, LiAl (OH) ꢀH O; ^, LiOHꢀH O).
2
7
2
2

7
18
F. Oksuzomer et al. / Materials Research Bulletin 39 (2004) 715–724
Fig. 3. XRD diagrams of samples calcined at 900 8C (*, g-LiAlO
2
; ~, LiAl
5
O
8
; &, LiAl
2
(OH)
7
2
ꢀH O).
Fig. 4. FT-IR spectra of air-dried samples at 120 8C.

F. Oksuzomer et al. / Materials Research Bulletin 39 (2004) 715–724
719
Fig. 5. FT-IR spectra of samples calcined at 550 8C.
predominant phase. LiAl O8 and LiAl (OH) ꢀH O phases were detected in appreciably less
5
2
7
2
quantities in t-butanol than those in the other two solvents. Furthermore, better crystallinity was
obtained at 900 8C with t-butanol.
ꢁ
1
Fig. 4 shows the infrared spectra of samples dried at 120 8C. Peaks at 550 and 780 cm are due to
the vibrations of AlO and the LiO þ AlO combination, respectively, [9,12,13] and 1370, 1425 and
6
6
6
ꢁ
1
small shoulder at 1475 cm are all due to Al-sec-butoxide [14]. Band at 1630 cm is due to hydroxyl
ꢁ1
ꢁ
1
formation. Peaks at 875 and 1303 cm are due to Eaa ligand formation [14]. Infrared spectra of
ꢁ1
samples calcined at 550 8C are shown in Fig. 5. Bands at 810, 650 and 550 cm indicate the presence
of g-LiAlO and LiAl O phases. Although g-LiAlO and LiAl O phases were not detected at this low
2
5
8
2
5 8
calcination temperature, FTIR analysis clearly shows their presence. Bands showing the presence of
organic residues are not visible at this temperature. This is in line with the findings from TG analysis
since weight loss due to the removal of organics was finished before 550 8C. Fig. 6 depicts infrared
spectra of samples calcined at 900 8C. Infrared spectra of g-LiAlO and LiAl O phases were well
2
5 8
ꢁ
1
ꢁ1
defined in samples with strong bands at 810, 650 and 550 cm and weak bands at 520 and 450 cm
ꢁ
1
for the samples calcined at 900 8C. Bands appearing at 610 and 680 cm refer to LiAl O phase
5
8
ꢁ
1
[
9,12,13]. Peaks at 810 and 650, 550 cm , characteristic of g-LiAlO phase, were sharper for samples
2
prepared with t-butanol solvent than those with other two solvents. This was an indication of that higher

7
20
F. Oksuzomer et al. / Materials Research Bulletin 39 (2004) 715–724
Fig. 6. FT-IR spectra of samples calcined at 900 8C.
purity g-LiAlO was produced when t-butanol was used as solvent. This result was also in line with the
2
XRD results of this study.
The SEM images of samples calcined at 900 8C were shown in Fig. 7. None of the particles showed
edges and well-defined crystals of g-LiAlO but the sample prepared with t-butanol was partially
2
faceted, although irregularly shaped, possibly corresponding to tetragonal g-LiAlO , as reported by
2
Valenzuela et al. [8]. With i-propanol and i-butanol, the samples showed more rounded corners. Also,
the samples showed a distribution of particle sizes.
Thermal behavior of the gels prepared with different solvents was studied by TGA and DTA. TGA and
DTA curves are, respectively, shown in Figs. 8 and 9. Slight weight losses observed at about 100 8C in
Fig. 8 are attributed to the dehydration of samples. There is no corresponding endothermic peak at the
same temperatures in DTA curves given in Fig. 9. Two endothermic peaks seen in the DTA curves were
attributed to the reactions between components as well as organic residue pyrolysis. There were two
successive weight losses observed around 200–255 8C and 255–550 8C in TGA curves as shown in Fig. 8.
Temperatures of two successive weight losses in TG analysis did not coincide with the temperatures of
major endothermic peaks observed in DTA. In all alcohol systems, weight losses stopped before 550 8C.
Overall weight losses were around 55% for alcohol systems. One interesting feature of the DTA curves

F. Oksuzomer et al. / Materials Research Bulletin 39 (2004) 715–724
721
Fig. 7. The SEM images of i-Pr, i-But, and t-But samples calcined at 900 8C.

7
22
F. Oksuzomer et al. / Materials Research Bulletin 39 (2004) 715–724
Fig. 8. TGA curves of samples prepared with different solvents.
on changing the solvent from t-butanol to i-butanol and to i-propanol, is the shift in the first and second
endothermic peaks (ꢂ270–355 8C and 408–432 8C) to comparatively higher temperatures as shown in
Fig. 9. Similar shifts in the onset and end temperatures of weight losses in TGA curves were observed.
Furthermore, in all samples there was a small endothermic peak near 700 8C in DTA curves. Since no
preceding a or b-lithium alumina could be detected from X-ray diffractograms of samples calcined at
550 8C (see Fig. 2), this peak should not be due to a or b-lithium aluminas’ phase change to g-LiAlO2
and it is more reasonable to consider this endothermic peak as an indication of direct formation of
g-LiAlO phase. In line with our findings, Kwon and Park [3] reported that this peak was related to the
2
direct formation of gamma lithium aluminate in the sol–gel method and lithium salt and alumina
precursor systems. They confirmed g-LiAlO phase formation from X-ray diffractograms of samples
2
obtained after calcinations near 700 8C. On the other hand, Turner et al. [15] reported no such peak.
Kwon and Park [3] reported that when butoxide–butoxide was mixed with methanol or ethanol
instead of butanol due to alcohol–alkoxide exchange, the dried gel did not disintegrate easily to give
fine homogeneous powders, whereas butoxide–butoxide system led itself to fine homogeneous powders.
However, in this study, with a mixed system (lithium methoxide and aluminum-sec-butoxide) and
various alcohols as solvents, highly pure and crystalline gamma lithium aluminate was produced. With
the help of ethylacetoacetate as ligand aid, homogeneous gels were produced. It can be argued that
alcohol–alkoxide exchange reaction occurred during gelation resulted in poorer dispersion with
i-butanol and i-propanol and consequently this led to less pure crystals. Thus, it is important to avoid this
type of exchanges. It has been shown here that, with the help of ligand aid such as ethylacetoacetate,
poor distribution of sol–gel precursors can be avoided.

F. Oksuzomer et al. / Materials Research Bulletin 39 (2004) 715–724
723
Fig. 9. DTA curves of samples prepared with different solvents.
4
. Conclusions
This work showed that the crystallinity and purity of g-LiAlO was depended on the solvent used.
Ligand aid such as ethylacetoacetate led to the better gelation of reagents by the sol–gel method and
2
hence higher purity g-LiAlO was produced. Major intermediates leading to the formation of g-LiAlO
2
2
were LiOHꢀH O and LiAl (OH) ꢀH O. Lithium methoxide–aluminum-sec-butoxide and t-butanol
2
2
7
2
system resulted in g-LiAlO of the highest crystallinity among the solvents studied. Some minor
2
amounts of LiAl O were also present in all solvents.
5
8
References
[
[
[
[
[
1] J.N. Singh, J.T. Dusek, J.W. Sim, Ceram. Bull. 60 (1981) 629.
2] W. Heyi, Z. Junke, L. Yangming, J. Nucl. Mater. 208 (1994) 195.
3] S.-W. Kwon, S.-B. Park, J. Nucl. Mater. 246 (1997) 131.
4] J. Becerril, P. Bosch, S. Bulbulian, J. Nucl. Mater. 185 (1991) 304.
5] K. Kinoshita, J.W. Sim, J.P. Ackerman, Mater. Res. Bull 13 (1978) 445.

7
24
F. Oksuzomer et al. / Materials Research Bulletin 39 (2004) 715–724
[
[
[
[
6] J.P. Kopasz, C.A. Seils, C.E. Jhonson, J. Nucl. Mater. 191 (1992) 231.
7] R.C. Anderson, W.J. Thomson, Adv. Ceram. 27 (1994) 13.
8] M.A. Valenzuela, L. Tellez, P. Bosch, H. Balmori, Mater. Lett. 47 (2001) 252.
9] R.A. Riberio, G.G. Silva, N.D.S. Mohallem, J. Phys. Chem. Solids 62 (2001) 857.
[
[
[
[
[
[
10] C.J. Brinker, D.W. Scherer, Sol–gel Science, Academic Press, New York, 1990.
11] R.C. Mehrota, J. Non-Cryst. Solids 100 (1998) 1.
12] W. Gessner, D. Muller, Anorg. Allg. Chem. 505 (1983) 195.
13] P. Tarte, Spectrochim. Acta 23A (1967) 2127.
14] J. Parmantier, M.R. Plouet, S. Vilminot, Mater. Res. Bull. 33 (1998) 1717.
15] C.W. Turner, B.C. Clatworty, A.H.Y. Grin, Adv. Ceram. 25 (1996) 141.