You are on page 1of 8

JOURNAL Of NUCLEAR And Related TECHNOLOGIES, Volume 6, No.

1, Special Edition, 2009

12

RIETVELD QUANTITATIVE PHASE ANALYSIS OF NONSTOICHIOMETRIC CORDIERITE SYNTHESISED FROM MAINLY


TALC AND KAOLIN: EFFECT OF SINTERING TEMPERATURE
Banjuraizah Johar1,2, Hasmaliza Mohamad2, Zainal Arifin Ahmad2.
1

School of Materials Engineering Universiti Malaysia Perlis


School of Materials and Mineral Resources Engineering Universiti Sains Malaysia
banjuraizah@unimap.edu.my

ABSTRACT
Indialite or -cordierite was synthesized by glass crystallization method using mainly talc and
kaolin and with small amount of MgO, Al2O3, SiO2 to compensate the chemical formulation of
non-stoihiometric compositions of cordierite. (3MgO.1.5Al2O3.5SiO2). B2O3, P2O5 and CaO
was also added to decrease the melting and sintering temperature of cordierite. The glasses
were pelletized and sintered from 850oC up to 1050oC. Phase compositions of both heat treated
glass was quantified by X-ray powder diffraction data by the Rietveld method using TOPAS Ver
3 software. Result shows that about 60wt% of cordierite has successfully crystallized at
850oC. Beside secondary phases (forsterite) which come from initial raw materials, phases from
grinding media were also presents in the sample. The contamination was considered high since
it has reacted with existence phases to form a new phase at higher temperature. Without any
contamination from grinding it was expected to obtain more than 90wt% cordierite using the
same composition.

ABSTRAK
Indialite atau -cordierite telah disintesiskan dengan kaedah penghabluran kaca menggunakan
terutamanya talkum dan kaolin dan dengan jumlah kecil MgO, Al2O3, SiO2 untuk memampasi
perumusan kimia tidak stoihiometric mengarang kordierit. (3MgO.1.5Al2O3.5SiO2). B2O3,
P2O5 dan CaO turut menambah untuk susut meleburkan dan pensinteran suhu kordierit.
Gelas-gelas itu adalah pelletized dan tersinter daripada 850oC sehingga 1050oC. Mengarang
fasa kedua-dua memanaskan mengambil kaca dinyatakan kuantiti oleh pembelauan serbuk
sinar X data oleh kaedah Rietveld menggunakan perisian TOPAS VER 3. Keputusan
menunjukkan yang kira-kira 60wt% cordierite telah dengan jayanya mengkristalkan di
850oC. Di sebelah fasa-fasa sekunder (forsterit) yang mana didatangkan bahan-bahan mentah
awal, fasa-fasa daripada media pengisaran adalah serta dikemukakan dalam sampel.
Pencemaran itu telah dipertimbangkan tinggi sejak ia telah bertindak balas dengan fasa-fasa
kewujudan untuk membentuk satu tahap baru pada suhu lebih tinggi. Tanpa sebarang
pencemaran daripada pengisaran ia dijangka memperoleh lebih daripada 90wt% cordierite
menggunakan komposisi sama.
Keywords: Rietveld, Quantitative analysis, cordierite glass

INTRODUCTION
Indialite or cordierite which have low coefficient of thermal expansion, dielectric constant
and dielectric loss has gained a significant attention during recent as a substrate in electronic
103

JOURNAL Of NUCLEAR And Related TECHNOLOGIES, Volume 6, No. 1, Special Edition, 2009

packaging especially for high frequency transportation and high speed applications (Camerucci,
Urretavizcaya, Castro, & Cavalieri, 2001; D. Pal, 1996; Jankovic-Castvan et al., 2007; Mei,
Yang, & Ferreira, 2000; Nandi, 1999). To fulfill these requirement, low sintering temperature
(<1000oC) is required in order to be co-fired with high conductive electrode such as copper,
silver or gold. However, -cordierite has high sintering temperature (>1350oC ) which is near to
the incongruent melting point of cordierite (Jankovic-Castvan et al., 2007). Study by various
researchers (Chen, 2008; Chen & Liu, 2007; Goel, Shaaban, Melo, Ribeiro, & Ferreira, 2007;
Lo & Yang, 1998; Oh, Hing, & Fong, 1999; Villegas & Alarcon, 2002; Wang, Zhou, & Luo,
2003) found that -cordierite can be crystallized at or below 950oC via glass crystallization
method, and most of them used high purity of oxides except two (C.Oprea, C. Stan, E. Rotiu, &
Popescu, 1999; Nandi, 1999) used mineral as raw materials. In addition, none of them melted
below than 1400oC. Low melting temperature (1350oC) was used in the presence study based
on the statement taken from Oprea et. al (C.Oprea et al., 1999) who stated that that the melting
temperature of MgO-Al2O3-SiO2 from minerals talc and kaolin can lower than pure oxides.
In all applications, the properties and performance of heat treated cordierite glass depend on
size, morphology, crystallinity and purity of phase. However purity of phase, crystallinity and
crystallite size will varies with sintering temperature, initial raw material and processing
conditions. Qualitative analysis as well as comparison based on count of intensity couldnt give
strong reason for the decreasing or increasing of certain properties e.g relative density, hardness
etc. Although many research have been done on synthesized cordierite, however not much
research have been carried out to understand and evaluate its quantitative phase transformation
characteristic as a function of sintering temperature.
Rietveld analysis is standardless method which allows the weight fraction of each phase in
unknown mixture to be calculated using the entire diffraction profile. This method requires
knowledge of crystal structure of all the phases present in the sample as the process comprises
the comparison of the measured and calculated patterns. In this refinement or fitting procedure,
the difference between the experimental powder diffraction pattern and the calculated one,
together with the background function, is minimized. This methodology has several advantages
over others because it can minimized overlap peak, preferred orientation and sample broadening
effect. Furthermore, absorption coefficient can be refined. This is because microabsorption
effect is one of the quantitative analysis problems especially when dealing with a binary
mixture of two or more crystalline phases. The intensity beam will decrease due to absorption
and if the mass attenuation coefficient of one phase is much higher than the other, then the
diffraction intensity will be much less than the calculated intensity. As a results, the wt% of one
phase will be overestimated and the other one will be underestimated.
The purpose of this presents study is to quantified secondary phase which was exist from initial
raw materials and grinding media, as a function of sintering temperature. Relative amount of
this secondary phase is difficult to estimate based on the observation of intensities since each
phase has different linear absorption coefficient and different structure factor which will effects
the intensity of the diffraction line.

MATERIALS AND METHODS


Synthesis of Cordierite Glass Powder
Cordierite glass powder was synthesized using kaolin, talc, MgO, Al2O3, SiO2, and CaO. The
composition of oxides was formulated using nonstoichiometric composition of cordierite
(3MgO.1.5Al2O3. 5SiO2). Compositions of the mixed powders were calculated based on
chemical reaction between chemical formulation of kaolin and talc to obtain required metal
oxide compositions. Small amount of MgO, Al2O3 was added to compensate the formulation. In
addition, in this formulation 3 wt% of MgO was replaced with CaO. 2wt% of B2O3 and 3wt%
P2O5 from the total weight percent of the mixture was use as sintering aids in both samples. The
104

JOURNAL Of NUCLEAR And Related TECHNOLOGIES, Volume 6, No. 1, Special Edition, 2009

oxides compositions is similar to Chen et al.(Chen, 2008), who has successfully obtained high
purity of cordierite at low sintering temperature (950oC) with all properties measured meet the
criteria for substrate in high frequency application. Mixed powder was melted in alumina
crucibles at 1350oC for 4 hours then immediately quenched in distilled water to form frits. Frits
was crushed and dried milled for 6 hours using zirconia ball and further heat treated at different
heat treatment temperature (850oC 1050oC for 2 h).
XRD were performed on diffractometer with CuK. The reading was taken on Bragg Brentano
geometry between 10 to 90o. Counting time was 0.2 s for each 0.03 2 steps. The Vantec
diffractometer consisted of primary soller slit 173mm. One 2mm scattered radiation slit was
placed downstream of the sample, followed by a system of secondary Soller slit and 0.2 mm
detector slit. The X-ray tube operated at 40 kV and 30mA. Sample was mounted on a flat Alholder by side loading in order to minimize preferred orientation of the crystal. TOPAS (ver. 3)
software was used to analyze X-ray diffraction spectra for quantitative analysis in this research.
Phase analysis by Rietveld method was carried out using TOPAS ver 3. Instrument parameters
and crystal structure data for each phase present in the samples were fed into the TOPAS
programme. Global optimized parameters were background coefficient (Chebychev polynomial
functions with five orders of the series), lattice parameter, zero-shift error, scale factor, peak
shape parameters, and absorption coefficient.
The crystal structure data which was used to interpret the powder pattern was taken from the
Inorganic Crystal Structure Database (ICSD). A split-pseudo-Voigt function was selected as a
peak profile function for all phase including the asymmetry correction of Finger et al. and the
surface roughness microabsorption effect was corrected by the Pitchke function. Preferred
orientation for some crystalline phase was taken into account using the March-Dollase
algorithm. Other parameters that were optimized are background polynomial coefficient, unit
cell parameter, scale factor, and zero-shift error. Atomic position, occupancies and temperature
factor was fixed during phase analysis. The refinement was done in stages and reflection profile
can be calculated by convolution of mathematical functions of these parameters.

RESULTS
Figure 1 shows phase evolution of sample after heat treated at 850oC up to 1050oC. The
diffraction line doesn't change up to 900oC. At and above 950oC, peaks belong to zirconium
silicate phase was observed and the intensity of this peak is increasing as a function of
temperature while other peaks belong to zirconia polymorph continuously decreased. It was
difficult to predict the whole reaction process without quantify all phases presence in the
sample. Therefore rietveld method was carried out using the same diffraction data.
Figure 2 presents refinement plots for all samples. Table 1 summarized phase content in all
sample at different sintering temperature, while table 2 demonstrate the refinement results.
About 30wt% of zircornia worn out during grinding and contaminate the sample. This can be
explained by the fact that the hardness of the quenched sample is near to the hardness of
grinding media. This is because it was melted at low temperature.
As shown in Table 1, the amount of cordierite phase is increasing when sintered from 850oC
to 900oC. Amorphous phase or disordered atom which was remain at 850oC crystallized at
900oC causing the amount of cordierite phase increased. However, above 900oC, the amount
of cordierite decreased as a function of sintering temperature.

105

JOURNAL Of NUCLEAR And Related TECHNOLOGIES, Volume 6, No. 1, Special Edition, 2009

Figure 1 Phase evolution of glass sample heat treated at 850oC up to


1050oC. a-alpha cordierite, S-spinel, F-forsterite, ZM-zirconia
monoclinic, ZT-zirconia tetragonal, ZC-zirconia cubic.
Table 1: Phase content in heat treated glass powder at various sintering temperature
T(oC)

cordierite

Spinel

Forsterite

Zircon

850
900
950
1000
1050

59.683
57.408
54.846
48.853
38.853

0
0
0.726
5.28
10.33

10.071
10.466
9.440
10.418
11.037

0
0
4.483
16.079
30.518

1,800
1,600

Zircorni
Cubic
0.473
0.235
0
0
0

Zirconia
tetragonal
13.266
9.643
3.870
2.052
0.662

Zircornia
monoclinic
16.507
22.248
26.634
17.319
8.597

Rwp
8.44
8.31
7.39
8.9
9.14

Alpha cordierite
ZrO2 tetragonal
ZrO2 monoclinic
ZrO2 cubic
Forsterite

850oC

1,400
1,200

59.68 %
13.27 %
16.51 %
0.47 %
10.07 %

1,000
800
600
400
200
0
-200
-400
-600
10

15

20

25

30

35

40

45

50

106

55

60

65

70

75

80

85

90

JOURNAL Of NUCLEAR And Related TECHNOLOGIES, Volume 6, No. 1, Special Edition, 2009
1,800

Alpha cordierite
ZrO2 tetragonal
ZrO2 monoclinic
ZrO2 cubic
Forsterite

900oC

1,600
1,400
1,200

57.41 %
9.64 %
22.25 %
0.23 %
10.47 %

1,000
800
600
400
200
0
-200
-400
-600
10

15

20

25

30

35

40

45

50

55

60

65

70

75

1,800

1,400

85

Spinel
Alpha cordierite
ZrO2 tetragonal
ZrO2 monoclinic
Forsterite
Zircon

950oC

1,600

80

1,200

90

0.73 %
54.85 %
3.87 %
26.63 %
9.44 %
4.48 %

1,000
800
600
400
200
0
-200
-400
-600
10

15

20

25

30

35

40

45

50

55

60

65

70

75

2,200
2,000

85

Spinel
Alpha cordierite
Zircon
ZrO2 tetragonal
ZrO2 monoclinic
Forsterite

1000oC

1,800

80

1,600
1,400

90

5.28 %
48.85 %
16.08 %
2.05 %
17.32 %
10.42 %

1,200
1,000
800
600
400
200
0
-200
-400
-600
-800
10

15

20

25

30

35

40

45

50

55

60

65

70

75

4,500
4,000
3,500

80

85

Spinel
Alpha cordierite
Zircon
ZrO2 tetragonal
ZrO2 monoclinic
Forsterite

1050oC

3,000

90

10.33 %
38.85 %
30.52 %
0.66 %
8.60 %
11.04 %

2,500
2,000
1,500
1,000
500
0
-500
-1,000
-1,500
10

15

20

25

30

35

40

45

50

107

55

60

65

70

75

80

85

90

JOURNAL Of NUCLEAR And Related TECHNOLOGIES, Volume 6, No. 1, Special Edition, 2009

Figure 2 Rietveld refinement results of glass ceramic heat treated at 850oC up to


1050oC. The observed diffraction lines are displayed in blue, with the
calculated diffraction line drawn in red and the difference curve (Io-Ic) in
grey is displayed near the bottom of the graph. Rietveld plot shows the good
quality of refinement.
The decreasing of cordierite phase at and above 950oC is probably due to increment in
zirconium silicate phase, where at and above this temperature some of zirconia has reacted with
cordierite to form zirconium silicate and spinel, and as a results, the amount of spinel also
increase with temperature.
5ZrO2 + 2MgO.2Al2O3.5SiO2 5ZrSiO4 + 2Mg2Al2O4 (spinel)
(1)
Some of zirconia tetragonal and cubic was also dissolved in the silicate liquid to gradually
allow crystallization of zirconium silicate by below reaction:
SiO2 + ZrO2 ZrSiO4
(2)
This reaction has caused the weight percent of cordierite and both type of zirconia polymorph
drops. At the same time, tetragonal zirconia has change to monoclinic zirconia. This can be seen
from refinement result at temperature 900oC and 950oC respectively, where the amount of
tetragonal zirconia has decreased and the amount of zirconia monoclinic is increasing.
However, above 950oC some of zirconia monoclinic also reacted with excess silica and cordierite to form zirconium silicate and spinel. The amount of forsterite is in the range of 9.8
to 11.6 % and no correlation is observed with temperature. The fluctuate results may come from
the probability of excess Mg2+ ion in non-stoichiometric composition to compete with Zr2+ ion
to form a bonding with SiO4 network.
As indicated in Table 2, refinement was converged for all samples with satisfied goodness of
fit. This means that the crystal structure data that was used as initial model together with other
refinement parameter to simulate all diffraction peaks achieve perfect agreements between
measured and calculated pattern.
Table 2: Results of quality of refinement
Sample

Rwp

Rp

Re

Crystal density (g/cm3)

Refinement result
S-850
S-900
S-950
S-1000
S-1050

8.44
8.31
7.39
8.90
9.14

6.52
6.3
5.64
6.76
6.81

5.65
5.69
5.6
5.73
5.44

Crystallite size

cordierite
1.48
1.77
1.34
1.55
1.68

2.509
2.508
2.503
2.498
2.495

87.7
99.8
85.9
89.3
92.4

Figure 3 plots the intensity and weight percent of cordierite phase as a function of
temperature. The trend on count of intensity of cordierite is the same as its concentration
measured by Rietveld method. Traditional method which used strongest peak to account the
concentration of phase is not accurate due to many factors that contribute to an error in count of
intensity such as probability of peak broadening come from instrument parameter and lower
particle size, lower intensity due to preferred orientation and different in mass absorption factor
and etc. This is because count of intensity was selected at certain plane. Although zirconium
silicate peak have higher intensity at 1050oC, but it doesnt mean that zirconium silicate is a
major phase in this sample. This is due to structure factors of zirconium silicate which caused it
has higher degree of crystallinity at certain plane.
108

JOURNAL Of NUCLEAR And Related TECHNOLOGIES, Volume 6, No. 1, Special Edition, 2009

Figure

3: Count of intensity (measured at 29.7o 2thetha) and weight


percent of cordierite phase as a function of sintering
temperature.

CONCLUSIONS
Based on the observation of highly crystalline sample, we can predicted the possibility of
obtaining more than 88wt% of cordierite by using the same stoichimetric composition and
initial raw materials if problem related to pulverized process was solved. Low concentration of
cordierite phase at higher sintering temperature was related to the reaction between
contaminated phases and chemicals in the sample. Consequently only 38.8 wt% of cordierite
phase left after sintered at 1050oC. Results of quantitative analysis is very useful in determining
type of reaction occurred especially when multi-phases are presence. Although quantitative
phase analysis can be determined using a variety of methods, but the rietveld method is the
most appropriate and accurate method in phase analysis since no internal standard need to be
prepared and it can be run in a few minutes.

ACKNOWLEDGEMENTS
The authors gratefully acknowledge the financial supports from Islamic Development Bank
(IDB) and Fundamental Research Grant Scheme (FRGS 9003-00171) Universiti Malaysia
Perlis and technical assistants from Universiti Sains Malaysia (USM).

REFERENCE
C.Oprea, C. Stan, E. Rotiu, & Popescu, C. (1999). Non-Isothermal Crystallization of cordierite
glasses. J. Therm. Anal. Calorim., 56, 611-615.

109

JOURNAL Of NUCLEAR And Related TECHNOLOGIES, Volume 6, No. 1, Special Edition, 2009

Camerucci, M. A., Urretavizcaya, G., Castro, M. S., & Cavalieri, A. L. (2001). Electrical
properties and thermal expansion of cordierite and cordierite-mullite materials. J. Eur.
Ceram. Soc., 21(16), 2917-2923.
Chen, G.-h. (2008). Sintering, crystallization, and properties of CaO doped cordierite-based
glass-ceramics. J. Alloys Compd., 455(1-2), 298-302.
Chen, G.-H., & Liu, X.-Y. (2007). Sintering, crystallization and properties of MgO-Al2O3SiO2 system glass-ceramics containing ZnO. J. Alloys Compd., 431(1-2), 282-286.
D. Pal, A. K. C., Suchitra Sen. (1996). The Synthesis, Characterization, and Sintering of Sol-gel
derived Cordierite Ceramics for electronic applications. J. Matter. Sci., 31, 3995-4005.
Goel, A., Shaaban, E. R., Melo, F. C. L., Ribeiro, M. J., & Ferreira, J. M. F. (2007). Nonisothermal crystallization kinetic studies on MgO-Al2O3-SiO2-TiO2 glass. J. NonCryst. Solids., 353(24-25), 2383-2391.
Jankovic-Castvan, I., Lazarevic, S., Jordovic, B., Petrovic, R., Tanaskovic, D., & Janackovic,
D. (2007). Electrical properties of cordierite obtained by non-hydrolytic sol-gel
method. J. Eur. Ceram. Soc., 27(13), 3659-3661.
Lo, S.-H., & Yang, C.-F. (1998). The Sintering Characteristics of Bi2O3 added MgO-CaOAl2O3-SiO2 Glass Powder. Ceram. Inter., 24, 139-144.
Mei, S., Yang, J., & Ferreira, J. M. F. (2000). Microstructural evolution in sol-gel derived
P2O5-doped cordierite powders. J. Eur. Ceram. Soc., 20(12), 2191-2197.
Nandi, A. K. (1999). Thermal expansion Behavior of Boron-doped glass ceramic. J. Am.
Ceram. Soc., 82(3), 789-790.
Oh, J. T., Hing, P., & Fong, H. S. (1999). Effect of sintering temperature mismatch on the
thermal diffusivity of cordierite-AlN composite. J. Mater. Process Tech., 89-90, 497500.
Villegas, M. A., & Alarcon, J. (2002). Mechanism of crystallization of Co-cordierites from
stoichiometric powdered glasses. J. Eur. Ceram. Soc., 22(4), 487-494.
Wang, S., Zhou, H., & Luo, L. (2003). Sintering and crystallization of cordierite glass ceramics
for high frequency multilayer chip inductors. Mater. Res. Bull., 38(8), 1367-1374.

110