The influence of La2O3 on sol-gel derived mullite densification behavior
V. Mandic, E. Tkalcec, S. Kurajica
University of Zagreb, Faculty of Chemical Engineering and Technology, 19 Marulicev trg,
Zagreb, Croatia
Abstract
Monophasic mullite (3Al2O3·2SiO2) samples doped with 4.76, 9.09 and 13.04 mol% of
La2O3, were prepared by a sol-gel process. Prepared gels were dried, grinded, pressed into
pellets and sintered at temperatures from 1100-1600 °C for 4 h.
Phase formation and densification behavior has been investigated as a function of the
La2O3 content and sintering temperature. Mullite thermal evolution was analyzed by
Differential Thermal and Thermogravimetric Analysis (DTA/TGA) and X-ray diffraction
(XRD). The density of the sintered ceramics bodies was measured using Archimedes method.
With increasing lanthanum content, the DTA peak of mullite crystallization at about 980
ºC decreased in intensity, while the mullite crystallization was shifted to higher temperatures.
Neither La2O3 nor La related compounds were detected by the XRD analysis for all doping
levels, but beside mullite -alumina was determined. The intensity of -alumina diffraction
peaks increased with the La2O3 doping level. The mullite unit cell parameters suggest
predominant incorporation of La2O3 in the glassy phase, which quantity increased with the
doping level.
Keywords
Mullite, lanthanum, sol-gel, density
1. Introduction
Mullite ceramic is used widely as a structural and refractory ceramic material because of its
intrinsic high strength, low thermal expansion, good chemical stability and good creep
resistance at high temperatures. (Mazdiyazni 1972, Osendi 1996, Hong 1999) Since cell
morphology and porosity directly impact the material’s ability to perform desired functions in
a particular application, it is desirable to fabricate a ceramic with structure of regulated
microstructure and porosity. (Colombo 2002, Peng, 2000, Kim, 2004)
Mullite ceramics can be synthesized via various routes, including conventional ones that
use kaolinite as a precursor material (Yamuna 2002) as well as those which try to lower the
synthesis temperature like sol–gel processing (Ivankovic 2003) or reaction bonding. (Suttor
1997, She 2002, Kim 2002)
The only binary phase stable at ambient pressure in the system Al2O3–SiO2 is mullite
(ICDD 73-1253), to which has been assigned a compositional range between 3Al2O3:2SiO2
(3:2 mullite) and 3Al2O3:SiO2 (3:1 mullite).
The sol–gel methods are usually classified into two categories, based on the precursor
chemical homogeneity and the resultant crystallization behavior. In molecularly mixed
amorphous single-phase system, mullite is formed at about 980 ºC (McCarthy 1967), while
mullitization from diphasic mixtures with alumina-rich and silica-rich precursors occurs at
temperatures ranging from 1150 to 1350 °C, depending on the mixing scales. (Felsche 1969)
However, the large weight losses and shrinkages during drying and sintering had made
V. Mandic, E. Tkalcec, S. Kurajica “The influence of La2O3 on sol-gel derived mullite densification behaviour”
difficulties to produce large quantity of mullite materials using the sol–gel process.
(Mazdiyazni 1972) Furthermore, chemical routes usually use chemicals which are expensive
and difficult to handle.
Mullite formation is related not only to material processing but also to introduction of other
compounds, especially oxides. Our intention is to study the role of lanthanum oxide in
influencing mullitization behavior in the case of the sol–gel reaction process, in order to
provide useful information of mullite or mullite related materials. In the present paper, the
effect of La2O3 on mullite formation, thermal evolution, microstructure and densification
behavior, in single phase sol-gel derived mullite precursor, will be reported.
2. Experimental
Starting materials were (La(NO3)3·6H2O, p.a., Kemika, Croatia), (Al(NO3)3·9H2O, p.a.,
Kemika, Croatia), (Si(OC2H5)4 98%, Merck, Germany) and (C2H5OH 96%, Kemika, Croatia).
The appropriate amount of La(NO3)3·6H2O and Al(NO3)3·9H2O were mixed with
demineralised water in reactor. Si(OC2H5)4 was dissolved in C2H5OH separately. The
solution was stirred for 1 h and then added dropwise to the reactor. The mixture was stirred in
a closed reactor for 8 days at room temperature, during which synthesis took place. The molar
ratio of SiO2 and Al2O3 was held constant (2:3), while La2O3 was added in 4.76, 9.09 and
13.04 mol% respectively (Table 1). The samples hereafter are referred as LM0 (without
La2O3), LM1 (4.76 mol% La2O3), LM2 (9.09 mol% La2O3) and LM3 (13.04 mol% La2O3).
During synthesis the gelation occurred and after 8 days stirring at room temperature white
slurry has been obtained. The obtained gel was dried for 24 h by 150W IR lamp on 5 cm
indent from gel in order to remove excess solution.
The obtained samples were then calcined in box furnace in static air at a heating rate of 10
°C/h at 700 °C for 2 h in order to remove organic and nitrate phase decomposition residuals.
Yielding mass was subsequently grinded to form a fine powder by means of wet grinding in a
planetary mill with 200 rpm for 2 h and with propanol as a milling media. Remaining
propanol was removed by another calcification process at 700 °C for 2 h. Obtained mullite
powder was uniaxially pressed into Φ 22 mm pellets with 100MPa pressure. The pellets were
then sintered in static air at a heating rate of 10 °C/h at various temperatures between 1100–
1600 °C for 2 h.
Table 1: Composition of starting materials
Composition
mol%
Sample
LM0
LM1
LM2
LM3
SiO2
Al2O3
La2O3
40.00
38.10
36.36
34.78
60.00
57.14
54.55
52.17
0.00
4.76
9.09
13.04
The thermal behavior of powder precursor was characterized with Differential Thermal
Analysis (DTA) and Thermo-Gravimetric Analysis (TGA) using simultaneous DTA/TGA
analyzer Netzsch STA 409. For the thermal analysis approximately 50 mg of material were
V. Mandic, E. Tkalcec, S. Kurajica “The influence of La2O3 on sol-gel derived mullite densification behaviour”
2
placed in Pt crucibles and heated at a rate of 10 °Cmin-1 to 1000 °C in a synthetic air flow of
30 cm3min-1, -alumina was used as a reference.
The crystal phases were identified by powder X-ray diffraction (XRD) on diffractometer
Philips 1830 with CuK radiation. Data were collected between 10 and 120 o2θ in a step scan
mode with steps of 0.02 o and counting time of 2 s.
Different pellets diameters are the result of different sintering temperatures and La2O3
contents. The shrinkage was determined by multiple measuring of the pellets diameters. The
average value was calculated. The density of the sintered ceramics bodies was measured by
means of Archimedes method.
3. Results and discussion
Thermal behavior
0
200
400
600
800
1000 1200 1400 1600
LM0
LM1
100
90
-0.5
412°C
316°C
242°C
70
-1.0
60
TGA derivative
LM3
80
TGA /%
0.0
DTGA
-1.5
50
TGA
183°C
40
0
200
400
600
800
-2.0
1000 1200 1400 1600
T/ C
Figure 1: TGA and derivative TGA (DTGA) curves of
the samples doped with various amount of La2O3
The thermal behavior of the samples is characterized by multistep weight loss in the
temperature range of 100-500 °C (Fig. 1). In the lower temperature region the DTA curves
consist of series endothermic and exothermic effects with various intensity and mutual
superposition.
LM0
Intensity /a.u.
LM1
LM2
LM3
988.5°C
Vmg-1
983°C
984°C
960
987.5 C
980
1000
Temperature /°C
Figure 2: DTA curves of the samples doped with various amount
of La2O3 in the temperature range between 950 and 1010 ºC
V. Mandic, E. Tkalcec, S. Kurajica “The influence of La2O3 on sol-gel derived mullite densification behaviour”
3
Literature indicates that sol-gel gained TEOS and Al-nitrate-nonahydrate premullite
system releases adsorbed water at 100 ºC, chemisorbed water from 190-200 ºC, respectively.
At 160-170 ºC ethanol degrades, at ~230 °C Si-OC2H5 bonds decompose and C2H5 groups
degrade and at 160, 230, 290 and 330 ºC nitrates decompose. (Nass 1995)
In high temperature area premullite powders exhibit only exothermic peak at ~980 ºC
resembling Schneider’s type I precursor (Schneider 1993, Schneider 1994) (Fig. 2). After
calcinations (Fig. 3) the effects in lower temperature region disappear and the exothermic
peak at ~980 ºC is more intense and its position depends on doping level. With the increase of
doping level the temperature of exothermic peak is shifted to higher temperature, in
accordance with literature (Regiani 2002). On the other hand its intensity decreases with
La2O3 content in the sample.
LM0
Intensity / a.u.
LM1
LM2
LM3
Vmg-1
970
975
980
985
990
995
1000
O
T/ C
Figure 3: DTA curves of calcined samples doped
with various amount of La2O3
XRD patterns
XRD analysis of all samples sintered at various temperatures (Figure 4) was performed in
order to determinate phases present in the studied system. Mullite diffraction lines appear in
XRD patterns of all samples as a dominate phase, while alumina appears only in those with
higher La2O3 content or sintered at higher temperature.
The only phases evident from XRD patterns are mullite and α-alumina. Neither La2O3 nor
La related phases were found in the samples. Furthermore, the content of crystalline mullite
decreases and the background increases with the increase of La2O3 content, suggesting the
increase of amorphous phase. -alumina peaks also increase with the increase of La2O3
content. Regarding the La2O3 two possibilities exist; (1) complete incorporation of lanthanum
in mullite structure or (2) dissolution of lanthanum in the Si-rich glassy phase. Whereas the
amorphous phase increases with the increase of La2O3 on the account of the decrease of the
crystalline phases, the dissolution of La2O3 in the glassy phase should be primarily supposed,
rather than the incorporation of La2O3 into mullite.
XRD patterns also reveal increased mullitization with increase of thermal treatment
temperature.
V. Mandic, E. Tkalcec, S. Kurajica “The influence of La2O3 on sol-gel derived mullite densification behaviour”
4
Corrundum
Mullite
LM0
1600C
1500C
1400C
1300C
Intensity /a.u.
Intensity /a.u.
LM1
1600C
1500C
1400C
1300C
1200C
1200C
1100C
10
20
30
40
50
60
1100C
70
10
20
30
2/
40
1400C
1300C
40
50
60
Intensity /a.u.
Intensity /a.u.
1500C
30
70
LM3
1600C
20
60
2/
LM2
10
50
1600C
1500C
1400C
1300C
1200C
1200C
1100C
1100C
70
10
20
30
2/
40
50
60
70
2/
Figure 4: XRD patterns of the doped samples sintered at temperatures from 1100-1600ºC
Lattice parameters
Mullite crystallizes in orthorhombic system and its unit cell is characterized with three lattice
parameters. (Figure 5)
Unit cell parameters /
7,70
3,0
7,69
b-axis
c-axis 2,9
7,68
7,56
a-axis
2,8
7,54
0
4
8
12
16
La2O3 / mol%
Figure 5: Unit cell parameters of mullite in samples
fired at 1600ºC for 4 h
V. Mandic, E. Tkalcec, S. Kurajica “The influence of La2O3 on sol-gel derived mullite densification behaviour”
5
Unit cell parameters of mullite are given in Fig. 5. While a-axis of mullite shows
elongation (Δa/a=0.165%), the b-axis is characterized in contraction (Δb/b= –0.04%) by
addition of 4.76 mol% of La2O3; on the other hand the elongation of c-axis is very small (Δc/c
= 0.038%) No further change of axes is evident by increased addition of La2O3. The observed
change of lattice parameters are not necessarily the consequence of lanthanum entrance in
mullite lattice and could also be explained with the change of Al2O3 / SiO2 ratio in mullite
lattice. It has to be noticed that by the increase of doping level the amount of α-alumina
increases, with simultaneous decrease of mullite amount in the samples. No change of unit
cell parameters (or very small change) at higher La2O3 amounts can be explained by
lanthanum tendency to form glassy phase with SiO2 and Al2O3 (Sadiki 2006). Excess
lanthanum does not form any of stabile crystal phase under these conditions, therefore we can
assume formation of lanthanum glassy amorphous phase that can not be detected and
confirmed by XRD analysis. (Figure 4)
To confirm existence of lanthanum rich SiO2-glassy phase and to explain its formation
ability, an additional investigation of this system is necessary. Namely, micrographic
investigation of sample surface, especially grain boundaries, as well as exact determination of
chemical composition in crystal grains versus glass would offer a valuable information.
Density
The sintering progress can be monitored through shrinkage and volume changes as a function
of sintering temperature and lanthanum content, respectively. (Figure 6)
3000
0,30
LM0
LM1
LM2
LM3
2000
1500
LM0
LM1
LM2
LM3
0,25
0,20
l/%
V/mm
3
2500
0,15
0,10
0,05
1000
1100
1200
1300
1400
T/°C
1500
1600
1100
1200
1300
1400
1500
1600
T/°C
Figure 6: Pellet’s volume and linear shrinkage as a functions of sintering temperature;
the lines are added as a guidelines for the eye
Figure 6 displays a volume decrease with increasing sintering temperature and La2O3
amount, respectively. Further, the figure displays maximum shrinkage at highest sintering
temperature (1600 °C). As sintering process accelerates with the temperature increase,
particles get closer, leading to lower porosity, namely higher pellets density. The biggest
shrinkage difference at the same temperature can be observed between samples with low
La2O3 content, while for those with higher La2O3 concentration the difference is much
smaller. The sintering process accelerates at higher temperatures. The lanthanum addition has
a positive densification effect.
V. Mandic, E. Tkalcec, S. Kurajica “The influence of La2O3 on sol-gel derived mullite densification behaviour”
6
3200
LM0
LM1
LM2
LM3
Density /kgm
-3
2800
2400
2000
1600
1100
1200
1300
1400
1500
1600
T/ C
Figure 7: Pellet’s density as a function of sintering temperature;
the lines are added as a guidelines for the eye
Experimental data gained trough Archimedes method reveal the increase of samples
density with temperature increase. They also reveal that the sample density is almost reaching
mullite theoretical density (3.16 gcm-3) (Figure 7). Beside the sample density increase this is
also the consequence of high La2O3 density (6.51 gcm-3). Samples LM0, LM1, LM2 and LM3
at 1600 ºC exhibit maximum density; 1.70, 2.63, 3.03 and 3.08 gcm-3.
4. Conclusion
The influence of La2O3 on sol-gel gained mullite density and crystallization path was
studied.
The gels exhibit crystallization path resembling Schneider’s type I precursors.
Two crystal phases present in the samples are mullite and -alumina. Neither La-silicate
nor La-aluminate phases crystallize in the system.
Mullite unit cell parameters reveal slight changes with the amount of La2O3 added. It
remains inconclusive whether incorporation of any lanthanum in mullite occurs, or the whole
of lanthanum is contained in glassy phase.
The addition of La2O3 was advantageous to densification in single phase aluminosilicate
gels. The sample density increases with sintering temperature and lanthanum amount. For
sample LM3 sintered at 1600 ºC almost theoretical mullite density was reached.
Additional investigation of system tendency to form lanthanum rich glassy amorphous
phase, as well as the phase itself is necessary.
Acknowledgment
The financial support of the Ministry of Science, Education and Sports of Republic of Croatia
within the framework of the project Noº 125-1252970-2981 “Ceramic nanocomposites
obtained by sol-gel process” is greatly acknowledged.
V. Mandic, E. Tkalcec, S. Kurajica “The influence of La2O3 on sol-gel derived mullite densification behaviour”
7
References
























Colombo P. and Hellmann J.R., ‘‘Ceramics Foams from Preceramic Polymers,’’
Mater. Res. Innovat., 6, 260–72 (2002).
Felsche J., Naturwissenschaften 56, 325 (1969).
Hong S.H. and Messing G.L., ‘‘Development of Textured Mullite by Templated Grain
Growth,’’ J. Am. Ceram. Soc., 82 [4] 867–72 (1999).
ICDD 73-1253 Durovic S., Chem Zvesti, 23, 113 (1969).
International Tables for X-ray Crystallography, Vol. 1, Kynoch Press, England (1952).
Ivankovic H., Tkalcec E., Nass R., and Schmidt H., ‘‘Correlation of the Precursor
Type with Densification Behavior and Microstructure of Sintered Mullite Ceramics,’’
J. Eur. Ceram. Soc., 23, 283–92 (2003).
Kim H.S. and Nicholson P.S., ‘‘Use of Mixed-Rare-Earth Oxide in the Preparation of
Reaction-Bonded Mullite at ≤1300 °C,’’ J. Am. Ceram. Soc., 85 [7] 1730–4 (2002).
Kim Y.-W., Kim S.H., Kim H.D., and Park C.B., ‘‘Processing of Closed-Cell Silicon
Oxycarbide Foams from a Preceramic Polymer,’’ J. Mater. Sci., 39, 5647–52 (2004).
Kong L.B., Zhang T.S., Ma J., Boey F., Zhang R.F., Journal of the European
Ceramic Society 23 (2003) 2247-2256
Ma T.I., Mazdiyasni K.S., Mechanical properties of mullite, J. Am. Ceram. Soc. 66
1983) 699-703
Mazdiyasni K.S. and Brown L.M., ‘‘Synthesis and Mechanical Properties of
Stoichiometric Aluminum Silicate (Mullite),’’ J. Am. Ceram. Soc., 55, 548–52 (1972).
Mazza D., Ronchetti S., Matter. Res. Bull. 34 (9) (1999) 1375-1382
McCarthy P., J. Nucl. Inorg. Chem. 29, 253 (1967).
Metcalfe B.L., Sant J.H., Trans. J. Br. Ceram. Soc. 74 (7) (1975) 193–201
Nass R., Tkalcec E., “Single phase mullite gels doped with chromium”, J. Am. Ceram.
Soc., 78 [11] 3097-1995.
Osendi M.I. and Baudin C., ‘‘Mechanical Properties of Mullite Materials,’’ J. Eur.
Ceram. Soc., 16, 217–24 (1996).
Peng H.X., Fan Z., Evans J.R.G., and Busfield J.J.C., ‘‘Microstructure of Ceramic
Foam,’’ J. Eur. Ceram. Soc., 20, 807–13 (2000).
Regiani W., Magalhães L.E., De Souza D.P.F., Paiva-Santos C.O., De Souza M.F., J.
Am. Ceram. Soc. 85 (1) (2002) 232-238
Sadiki N., Coutures J.P., Fillet C., Dussossoy J.L., “Crystallization of Lanthanum and
Yttrium aluminosilicate glasses,” Journal of Nuclear Materials (2006) 70-78
Schneider H., Saruhan B., Voll D., Merwin L., Sebald A., “Mullite precursor phases,”
Journal of European Ceramic Society 11 (1993) 87.
Schneider, H., Voll, D., Saruhan, B., Sanz, J., Schraeder, G., Ruescher C., Mosset, A.,
“Synthesis and structural characterization of non-crystalline mullite precursors,” J.
Non-Cryst. Sol., 1994, 78, 262-271.
She J., Mechnich P., Schmucker M., and Schneider H., ‘‘Reaction-Bonding Behavior
of Mullite Ceramics with Y2O3 Addition,’’ J. Eur. Ceram. Soc., 22, 323–8 (2002).
Suttor D., Kleebe H.-J., and Ziegler G., ‘‘Formation of Mullite from Filled
Siloxanes,’’ J. Am. Ceram. Soc., 80 [10] 2541–8 (1997).
Yamuna A., Devanarayanan S., and Lalithambika M., ‘‘Phase-Pure Mullite from
Kaolinite,’’ J. Am. Ceram. Soc., 85 [6] 1409–13 (2002).
V. Mandic, E. Tkalcec, S. Kurajica “The influence of La2O3 on sol-gel derived mullite densification behaviour”
8