Mullitization of aluminosilicate diphasic gel in the presence of nickel oxide additive

Abstract

Aluminosilicate diphasic gel was synthesized by using inorganic salts and characterized by chemical analysis, Fourier transform infrared spectroscopy spectroscopic studies and surface area and bulk density measurements. The gel powder was compacted with nickel oxide additive in three different ratios (wt/wt). Differential thermal analysis of gel samples with additive and without additive was carried out at four different heating rates and for each sample, the activation energy of mullitization was calculated using Kissinger equation. The microstructure and phase analysis were carried out by scanning electron microscopic and X-ray diffraction studies. Observation showed that nickel oxide favored the mullitization process.

Keywords

Ceramics sol–gel mechanical properties scanning electron microscopy

Introduction

As an engineering material, mullite is known as “material par excellence” for several remarkable physicochemical properties, which include good thermal and chemical stability, high melting point, low creep rate,^1–4 and so forth. Based on these qualities, mullite is widely used in the production of heat-resistant material,⁵ electronic packaging,⁶ thin film,⁷ window material for the mid-infrared wavelength range from 3 to 5 µm,⁸ and so forth. Mullite is a non-stoichiometric compound and its molecular formula can be represented as Al₂[Al₂₊₂ _x Si₂₋₂ _x ]O₁₀₋ _x , where x denotes the number of missing oxygen atoms per unit cell, varying between 0.17 and 0.59.⁹ Mullite can be synthesized by various techniques, and when it is prepared by processes such as sol–gel, precipitation, and hydrolysis, it is termed as chemical mullite. Depending on atomic arrangement, mullite gel can be classified as monophasic and diphasic.¹⁰ The advantage of diphasic gel system is that the heat of reaction of the different phases provides extra energy for the densification process.¹¹ Again, in diphasic gel system, homogeneity can be maintained in the nanometer range.¹²

By controlling the synthesis temperature and atmosphere, transition metal cations can be incorporated in mullite structure in different proportions.¹³ Different experiments have been carried out to study the effect of nickel addition on the mullitization process. Roy et al.^14,15 synthesized nickel-doped mullite by using the sol–gel technique. During phase evaluation and microstructural studies, they observed that nickel addition helps the mullitization process. Bagchi et al.¹⁶ observed that mullite can be synthesized in the presence of cobalt and nickel ions even at a low temperature of 600℃ by sol–gel technique. In the gel matrix, increased metal–ligand interaction was found to be responsible for such low temperature mullite phase formation. Tkalcec et al.¹⁷ reported that mullitization of nickel-doped diphasic Al₂O₃–SiO₂ gel proceeded through crystallization of Ni-containing γ-Al₂O₃ into nickel aluminate spinel (NiAl₂O₄) and reaction of Al–Si spinel with amorphous silica forming mullite. Sales et al.^18,19 synthesized NiAl₂O₄ spinel-mullite through sol–gel and observed that Al–Si spinel formed below 1000℃ was converted to mullite between 1100℃ and 1200℃.

Roy et al. investigated the role of different transition metal oxides such as vanadium pentaoxide, titanium oxide, chromium oxide, cobalt oxide, iron oxide, copper oxide on the mullitization of diphasic aluminosilicate precursor.^20–25 So in this present study, the effect of black nickel oxide, Ni₂O₃, on the crystallization of mullite from Al₂O₃–SiO₂ diphasic gel precursor powder derived from inorganic salts was studied by determining the activation energy of mullitization and analyzing the microstructure and mechanical properties of sintered products.

Experimental

Synthesis of nickel-doped mullite

To synthesize mullite precursor gel, the starting materials used were 5% (wt/vol) aluminum nitrate nonahydrate (Al(NO₃)₃.9H₂O, Alumina (Al₂O₃) content 12.98% wt/wt) extra pure (MERCK, India) and 5% (wt/vol) liquid sodium silicate (specific gravity 1.6 and molar ratio of Na₂O: SiO₂ = 1 : 3, silica (SiO₂) content 29.75% wt/wt) (LOBA CHEMIE, India). Liquid sodium silicate (7% wt/vol) was passed through an ion exchanger column packed with Dowex-50 cation exchanging resin with a flow rate of 200 ml/min to prepare silicic acid. Silica sol was obtained by ultrasonic dispersion of the silicic acid in aqueous phase. Al(NO₃)₃.9H₂O solution and silica sol was mixed together in such a way that weight ratio of alumina and silica in the mixture remain 73 : 27, which is slightly higher in alumina content than that in mullite (72 : 28) in order to avoid formation of glassy phase during mullitization. To this mixture, ammonia solution (1 : 1, vol/vol) was added slowly with constant stirring until pH of the solution became ∼9. Mixed sol was aged to get the gel. The gel was then filtered and washed thoroughly followed by drying at 80℃ for overnight.

One part of the dried gel was then calcined at 800℃ for 2 h. Then in a pot mill, one part of the calcined gel was mixed with reagent grade Ni₂O₃ (MERCK, India) additive in three different ratios (1%, 2%, and 3% wt/wt) and thoroughly ground to ensure proper homogeneity in the mixture.

Characterization

The results obtained during chemical analysis (measured following specification laid in ISO 21587-2:2007), surface area (measured by Twin surface area analyzer, Quantachrome), and bulk density (measured by Ultrapyc 1200e Helium Pyconnometer) measurement of the dried gel are given in Table 1. The dried gel was heat treated at eight different temperatures from 200℃ to 1600℃ with an interval of 200℃ each for 1 h and Fourier transform infrared spectroscopy (FTIR) spectra of the dried gel and heat-treated samples were taken using a Perkin-Elmer apparatus (Model: Vertex-70).

Table 1.

Physicochemical properties of the hydrogel.

Composition	Properties
SiO₂ (wt%)	17.86
Al₂O₃ (wt%)	48.23
Ignition loss (%)	33.91
Bulk density (g/cm³)	0.27
Sp. surface area (m²/g)	70

Differential thermal analysis of both the calcined and doped gel were performed at four different heating rates 4, 6, 8 and 10℃/min from room temperature to 1400℃ using differential thermal analyzer (Okay Libratherm instrument, Model: DTA-1500, Bysak, India). The activation energy of mullitization of samples with and without additive was determined from DTA thermogram using Kissinger equation.²⁶

The equation used is

\ln (\frac{T p 2}{υ}) = \ln (\frac{E a}{R}) - \ln ϕ + \frac{E a}{RT p}

(1)

where T_p is the temperature of the exothermic peak, υ is the heating rate, E_a is the activation energy, R is the universal gas constant, and ϕ is the frequency factor.

According to equation (1), the plot of $\ln (T p 2 / υ)$ versus 1/T_p will be a straight line and from the slope of the line, the value of activation energy can be evaluated.

The powder mixes were compacted at 100 MPa, and the compacted masses were fired in an electrically heated muffle furnace at three different final temperatures, 1400℃, 1500℃, and 1600℃ (heating rate was 10℃/min upto 1000℃ and then 2℃/min till final temperature was reached), with 2 h of soaking period in each case. The bulk density and apparent porosity of the sintered masses were measured by following the procedures described in BS 1902, Part 1A, 1966.²⁷ The flexural strength of the sintered samples was determined by the measurement of bending strength with a span of 30 mm and a loading rate of 0.5 mm/min. Fracture toughness was determined with an indentation micro-crack method with a load of 5 kg.²⁸ X-ray diffraction (XRD) pattern of the samples was taken with a Rigaku X-ray diffractometer with Cu target (Miniflex, Japan). The lattice parameters of samples with and without additive were calculated using the following equation

\frac{h 2}{a 2} + \frac{k 2}{b 2} + \frac{l 2}{c 2} = \frac{1}{d 2}

(2)

where h, k, and l are the Miller indices of the peaks in Cartesian coordinates; a, b, and c are the lattice parameters; and d is the spacing of the lattice planes.

The % d-error for the sample and JCPDS (#5-0776) standard d-values for all planes are calculated using the following formula

Relativepercentageerror, i . e . % d error = \frac{| d ex - d |}{d} \times 100

(3)

where d_ex is experimentally obtained d value and d is the standard d-value in JCPDS file.

The average crystallite size was calculated using Debye–Scherrer formula²⁹

D = 0.9 λ / (β \cos θ)

(4)

where D is the diameter of the crystallites forming the film, λ is the wavelength of CuKα line, β is full width at half-maximum (FWHM) in radians, and θ is the Bragg’s angle

β = \sqrt{(β 0) 2} - (b) 2

(5)

where β₀ is the FWHM in radians for the sample, and b is the FWHM in radians for the pure crystal.

Scanning electron microscopic investigation of the samples was carried with FEI Quanta microscope (US).

Results and discussion

The precursor gel had very low bulk density (0.27 g/cm³) and high surface area (70 m²/g) indicating high surface activity of the gel. The water content of the gel was almost 34%. To prevent excessive shrinkage during sintering, the gel was calcined at 800℃ for 2 h.

FTIR spectral analysis of the undoped sample

Figure 1(a) and (b) represent the FTIR spectra of the dried and heat-treated samples at different temperatures. Bands that appeared at 3151 and 1104 cm⁻¹ in the spectra of the dried gel were assigned to the Al–OH stretching and bending vibrations, respectively. Shoulder appeared at 1104 cm⁻¹ and was related to the overlapping of Si–O–Si stretching vibration of SiO₄⁴⁻ tetrahedra with Al–OH bending vibration.³⁰ Si–O–Si bending vibration appeared at 477 cm⁻¹. Again stretching of Al–O–Al linkage was observed at 747 cm⁻¹. But no characteristic peak corresponding to Al–O–Si linkage was observed in the FTIR spectra of the gel sample. It confirmed that both the gel, aluminum hydroxide, Al(OH)_3, and silicon hydroxide, Si(OH)₄, formed simultaneously. So precursor gel was diphasic in nature.

Figure 1.

(a) FTIR spectra of the aluminosilicate gel after heat treatment at different temperatures within 2000–4000 cm⁻¹ and (b) FTIR spectra of the aluminosilicate gel after heat treatment at different temperatures within 400–2000 cm⁻¹.

Si–O–Al linkage first appeared in the spectra of the sample heated at 600℃. The peak at 803 cm⁻¹ was assigned for symmetric stretching vibration of Si–O–Si linkages. This peak was slightly shifted to 806 cm⁻¹ in the spectra of the sample heated at 800℃. This result confirmed the formation of Al–Si spinel in the gel structure. After heating at 1000–1200℃, the major change in the spectra was observed in the shifting of peaks at 832–833 cm⁻¹ from 806 cm⁻¹. This indicated the formation of more Si–O–Al linkages resulting in the crystallization of mullite phases. The peak observed at 843 cm⁻¹ in the spectra of the sample heated at 1400℃ became wider, indicating the formation of more Si–O–Al linkages in the mullite crystals. Again peaks at 586 cm⁻¹ due to the stretching vibration of octahedrally coordinated Al became sharper, indicating the formation of Al₂O₃ in the structure. In the spectra of the sample heated at 1600℃, the peak corresponding to Si–O–Si linkage became broader, indicating complete crystallization of mullite.³¹

Roll of nickel ion as additive

Black nickel oxide used in the present experiment contains some non-stoichiometric NiO.³² Therefore, the following defect reactions are proposed for both NiO and Ni₂O₃ with Al₂O₃. Melting point of NiO is about 600℃.³³ So during heat treatment, mobile nickel ions could substitute Al³⁺ in AlO₆ octahedra generated during crystallization of pseudo-boehmite gel, isovalently. The 3d⁷ electrons of Ni³⁺ in an octahedral crystal field are split from the energetic ground state to the high spin state [(t_2g)⁶(e_g)¹]. The cationic size of six-coordinated Ni³⁺ under high spin state is 72 pm, which is larger than that of Al³⁺.³⁴ The e_g electron occupies the d_z² orbital, and this electron can repel the electrons of the respective oxygen ligands. As a result, the z-axis of the octahedron is lengthened. This distortion is known as Jahn-Teller distortion³⁵ due to which deformation in oxygen octahedral can take place. This deformation is responsible for expansion of lattice along b-axis and shortening along a-axis.

The incorporation of Ni (II) ions into the aluminosilicate samples does not produce strain because this cation has a stable d⁸ electronic configuration with a spherically symmetric charge distribution. But cation can induce defect in the structure in the following way

NiO + Al 2 O 3 \to Ni Al / + Al Al xx + O O xx + 0.5 \overset{··}{V} O . .

(6)

In the same was Ni₂O₃ can also create defect in the Al₂O₃ sub lattice. The defect generated in the Al₂O₃ sub lattice assisted the densification of the aluminosilicate system. Again, expansion along a crystallographic direction is also possible since the ionic radii of six-coordinated Ni²⁺ [83 pm] is much larger than that of Al³⁺.³⁶

Moreover, NiO can produce a spinel phase by reaction with Al₂O₃ gel in the following way

NiO + Al 2 O 3 \to NiAl 2 O 4

(7)

Ni₂O₃ can react with silica gel to form nickel silicate

Ni 2 O 3 + SiO 2 \to Ni 2 SiO 5

(8)

Mullitization kinetics

DTA curves of samples with and without additive (Figure 2(a)) predicted that mullitization process was exothermic in nature. Again with increase heating rate, peaks were shifted toward higher temperature region. For gel without additive, primary mullitization was noticed in the temperature range of 975–990℃, whereas in case of sample with additive, mullitization was started at comparatively lower temperature (965–985℃). The exothermic peaks were shifted to lower temperature with increasing additive content, but the change was very little (∼1–2℃). Now by applying Kissinger equation (equation (1)), calculations were performed in each case. The plots of $\ln (ϕ / T m 2)$ values against (1/T_m) ×10³ are shown in Figure 2(b), and the values of activation energy are given in Table 2. The value of activation energy of mullitization of sample without additive was equal to ∼772 KJ/mole, which was much less, compared with that available in the literature.^26,37,38 With increasing additive content, the activation energy of mullitization was reduced. In the case of samples with 1% and 2% nickel oxide additive, the percentage of reduction of activation energy values were very little (<1%), but significant percentage of reduction of the same was observed in sample with 3% Ni₂O₃ additive (∼7.4%). So by increasing nickel concentration, mullitization process can be faster, possibly due to improved densification.

Figure 2.

(a) DTA curves of: (i) undoped gel and (ii) 1%, (iii) 2%, and (iv) 3% Ni₂O₃-doped samples and (b) plot of ln(ϕ/T_m²) versus 10³× T⁻¹ of: (i) undoped gel and (ii) 1%, (iii) 2%, and (iv) 3% Ni₂O₃-doped samples.

Table 2.

Values of activation energy.

Sample	E (kJ/mole)	Reduction of activation energy with regard to undoped sample (%)
Calcined gel	772
1% Ni₂O₃	764	0.98
2% Ni₂O₃	764	0.98
3% Ni₂O₃	714	7.5

Crystallographic information

Nickel aluminate phases were detected in the XRD diagram of the samples with additive (Figure 3(b) to (d)). These phases promoted the formation of mullite by reducing the energy barrier for the reaction of alumina and silica. At higher sintering temperature, more peaks corresponding to mullite phases were detected in the XRD diagram of the sample with additive compared with the sample without additive.

Figure 3.

XRD diagrams of: (a) undoped gel sintered at 1600℃, (b) 1% Ni₂O₃-doped samples sintered at 1600℃, (c) 2% Ni₂O₃-doped samples sintered at 1600℃, and (d) 3% Ni₂O₃-doped samples sintered at 1600℃. XRD: X-ray diffraction.

The calculated values of lattice parameters of samples with and without additive are given in Table 3. These values were compared with the unit cell parameters of orthorhombic mullite given in the literature.^39–42 In all cases, percentage of change in cell volume decreased with the increase in sintering temperature. Again, in case of sample with nickel oxide additive, at lower sintering temperature, expansion mostly took place along b- and c-axes, but with the increase in the temperature, the expansion took place along a-axis. It indicated the development of some asymmetric nature in the morphology of the mullite crystal as a result of nickel oxide additive.

Table 3.

Values of lattice parameters of 3%-doped sample at different sintering temperature.

Sample	Firing temperature (℃)	Theoretical cell parameters				Experimental cell parameters				Change in cell volume (%)
Sample	Firing temperature (℃)	a (A^o)	b (A^o)	c (A^o)	abc (A^o)³	a (A^o)	b (A^o)	c (A^o)	abc (A^o)³	Change in cell volume (%)
Undoped Gel	1400	7.60	7.70	2.90	169.71	7.63	7.74	2.92	172.44	1.61
	1500					7.61	7.72	2.91	170.96	0.74
	1600					7.605	7.7	2.91	170.41	0.41
3% Ni₂O₃	1400					7.61	7.75	2.94	173.39	2.17
	1500					7.61	7.74	2.91	171.40	0.99
	1600					7.63	7.71	2.91	171.19	0.87
2% Ni₂O₃	1400					7.61	7.76	2.94	173.62	2.30
	1500					7.615	7.72	2.92	171.66	1.15
	1600					7.64	7.71	2.92	172.0	1.35
1% Ni₂O₃	1400					7.62	7.79	2.93	173.92	2.48
	1500					7.62	7.75	2.91	171.85	1.26
	1600					7.64	7.71	2.91	171.41	1.0

The change in cell volume was caused by the substitution of Al³⁺ by Ni³⁺ in the mullite structure. With the increase in sintering temperature and additive content, the cell volume of mullite decreased due to the expulsion of Ni³⁺ from the mullite lattice, and its accumulation at the grain boundary of mullite crystals resulted in the reduction of crystal size.

From the XRD data, the % d-error_, which is a measure of the shift in mullite crystal planes due to structural deformation and/or incomplete formation, has also been calculated in the presence of nickel oxide at different temperatures using equation (3), and the results are given in Table 4. From the results, it can be predicted that with the increase of additive content and the sintering temperature, % d-error decreased, implying that the crystal structure of mullite became more perfect.

Table 4.

% d-error for mullite phase in presence of nickel oxide additive at different sintering temperature.

Additive content (%) →	1	2	3
Sintering temperature (℃) ↓	1	2	3
1400	1.93	1.82	1.74
1500	1.41	1.3	1.05
1600	1.29	1.17	0.94

The average crystallite size after calculation using equation (5) for mullite sample with nickel oxide additive has been given in Table 5. With the increase in temperature, the size of the mullite crystallites became shorter, suggesting the accumulation of significant amount of nickel oxide-bearing phases at the grain boundary of mullite samples.

Table 5.

Average crystallite size (µm) of nickel-doped sample at different sintering temperature.

Additive content (%) →	1	2	3
Sintering temperature (℃) ↓	1	2	3
1400	3.01	2.82	2.26
1500	2.79	2.57	2.14
1600	2.43	2.28	1.92

Microstructural analysis

From the SEM micrographs of the sintered samples (Figures 4 and 5), it is apparent that sol–gel-derived mullite without additive formed small crystallites. The incorporation of nickel ions in the sol–gel mullite induced tabular crystal growth parallel to the crystallographic c-axis. With the increase in the nickel ion content in mullite, formation of more equiaxed but smaller sized crystallites was observed. From the EDX analysis of the sample with 3% Ni₂O₃ additive sintered at 1600℃ (Figure 6), the composition was found to be like this: Al₂O₃ = 71.88%, SiO₂ = 26.17, and Ni₂O₃ = 1.95%. It can be concluded that the residual nickel oxide was present at grain boundary of mullite crystal as nickel aluminate spinel and nickel silicate.

Figure 4.

Scanning electron micrograph of the undoped gel samples sintered at: (a) 1400℃, (b) 1500℃, and (c) 1600℃.

Figure 5.

Scanning electron micrograph of the sintered gel samples with 3% Ni₂O₃ additive sintered at: (a) 1400℃, (b) 1500℃, and (c) 1600℃.

Figure 6.

EDX spectrum of 3% Ni₂O₃-doped gel sintered at 1600℃.

Physico-mechanical properties

The variation in bulk density and apparent porosity of the samples with sintering temperature is shown in Figures 7 and 8. From the figures, it is clear that nickel oxide exhibited a positive effect on the densification of the mullite ceramics. During re-crystallization of mullite, nickel ions that came out from the crystal structure produced a liquid phase. For this reason, the densification of the sample was enhanced at elevated temperatures. The flexural strength and fracture toughness (Figures 9 and 10) of the samples also increased in the presence of nickel oxide additive. It can also be related to the improved densification of the samples in the presence of Ni₂O₃ additive. The effect was more pronounced for the batches sintered at 1500℃ compared with batches sintered at other temperatures. The additive in the liquid form reduced the stress surrounding the pores in the samples by displacing the gases, resulting in more compactness of the microstructure, which can be related toward the improvement of the mechanical strength. It was the reason for improvement of mechanical strength.⁴³ The formation of more interlocked crystalline phases in the presence of Ni₂O₃ additives can be another reason for the improvement of the mechanical strength. A small amount of highly viscous silica or aluminosilicate glass can exist at the grain boundaries, which would minimize the contribution of grain boundary sliding during the flexural stress. At higher sintering temperatures, both the flexural strength and fracture toughness did not improve much because of the development of more amorphous character in the sintered masses.

Figure 7.

Variation in bulk density of Ni₂O₃-doped sintered aluminosilicate gel with firing temperature (℃).

Figure 8.

Variation in apparent porosity of Ni₂O₃-doped sintered aluminosilicate gel with firing temperature (℃).

Figure 9.

Variation in flexural strength of Ni₂O₃-doped sintered aluminosilicate gel with firing temperature (℃).

Figure 10.

Variation in fracture toughness of Ni₂O₃-doped sintered aluminosilicate gel with firing temperature (℃).

Conclusions

Chemical mullite was synthesized by the colloidal interaction of silicic acid and aluminum nitrate nonahydrate solution. The gel powder possessed very low density and high surface area and consisted of separate non-linked units of alumina and silica gel. Ni₂O₃ was used as sintering additive for the processing of mullite ceramics in different proportions. Nickel oxide favored the mullitization process by reducing its activation energy. The sintering behavior of mullite was promoted by Ni₂O₃, and different mechanisms were put forward to explain the favorable effect of Ni₂O₃. The lattice expansion mostly occurred along b-axis. The crystallite size of mullite was also modified by Ni₂O₃. The mechanical properties of the sintered masses were also improved significantly due to enhanced microstructure and favorable phase compositions.

References

Schneider

Eberhard

. Thermal expansion of mullite. J Am Ceram Soc 1990; 73: 2073–2076.

Hynes

Doremus

. High-temperature compressive creep of polycrystalline mullite. J Am Ceram Soc 1991; 74: 2469–2475.

Kollenberg

Schneider

. Microhardness of mullite at temperartures to 1000℃. J Am Ceram Soc 1989; 72: 1739–1740.

Aksay

Dabbs

Sarikaya

. Mullite for structural electronic and optical applications. J Am Ceram Soc 1991; 74: 2343–2358.

Belogurova

Grishin

. Highly heat-resistant mullite-silicon carbide materials. Refract Ind Ceram 2009; 49: 466–468.

Ramakrishnan

Goo

Roldan

. Microstructure of mullite ceramics used for substrate and packaging applications. J Mater Sci 1992; 27: 6127–6130.

Kurihara T, Horiuchi M, Takeuchi Y, et al. Mullite ceramic substrate for thin-film application. In: Proceedings of the 40th components and technology conference, IEEE, New York, 3--6 April 1990, pp. 68–75.

Shinohara

Dabs

Aksay

. Infrared transparent mullite through densification of monolithic gels at 1250℃. Proc SPIE 1986; 683: 19–24.

Cameron

. Mullite: a substituted alumina. Am Miner 1977; 62: 747–755.

10.

Schneider

Okada

Pask

. Mullite and mullite ceramics, England, UK: John Wiley and Sons Ltd, 1994, pp. 123–123.

11.

Roy R, Komarneni S and Roy DM. Better ceramics through chemistry. Brinker CJ, Clark DE and Ulrich DR (eds). North-Holland, New York: Elsevier, 1984; pp. 347–360.

12.

Wei

Halloran

. Transformation kinetics of diphasic aluminosilicate gels. J Am Ceram Soc 1998; 71: 581–587.

13.

Schneider

. Transition metal distribution in mullite. Ceram Trans 1990; 6: 135–138.

14.

Roy

Bagchi

Bhattacharya

. A comparative study of densification of sol-gel-derived nano-mullite due to the influence of iron, nickel and copper ions. Int J Appl Ceram Technol 2013; 1.

15.

Roy

Bagchi

Das

. Electrical and dielectric properties of sol-gel derived mullite doped with transition metals. Mater Chem Phys 2013; 138: 375–383.

16.

Bagchi

Das

Bhattacharya

. Effect of nickel and cobalt ions on low temperature synthesis of mullite by sol–gel technique. J Sol-Gel Sci Technol 2010; 55: 135–141.

17.

Tkalcec

Kurajica

Schmauch

. Crystallization of amorphous Al₂O₃–SiO₂ precursors doped with nickel. J Non-Cryst Solids 2007; 353: 2837–2844.

18.

Sales

Valentin

Alarcon

. Reaction sequence in the preparation of NiAl₂O₄ spinel—mullite composites by sol—gel. J Am Ceram Soc 1997; 80: 1798–1804.

19.

Sales

Valentin

Alarcon

. Effect of NiO and/or TiO₂ mullite formation and microstructure from gels. J Mater Sci 1998; 33: 4435–4446.

20.

Roy

Bandyopadhyay

Das

. Effect of CoO on the formation of mullite ceramics from diphasic Al₂O₃-SiO₂ gel. J Eng Sci Technol Rev 2010; 3: 136–141.

21.

Roy

Bandyopadhyay

Das

. Effect of TiO₂ on the formation of mullite ceramics from diphasic Al₂O₃-SiO₂ gel. Interceram 2010; 59: 213–217.

22.

Roy

Bandyopadhyay

Das

. Role of Cr₂O₃ on the formation of mullite ceramics from diphasic Al₂O₃-SiO₂ gel. Ceramica (Braz.) 2010; 56: 273–278.

23.

Roy

Bandyopadhyay

Das

. Role of V₂O₅ on the formation of chemical mullite from aluminosilicate precursor. Ceram Int 2010; 36: 1603–1608.

24.

Roy

Bandyopadhyay

Das

. Effect of copper ions on mullite formation from aluminosilicate precursor. Ceram Silik 2010; 54: 128–132.

25.

Roy

Bandyopadhyay

Das

. Effect of synthetic Fe₂O₃ on the properties of mullite ceramics from diphasic Al₂O₃-SiO₂ gel. J Aus Ceram Soc 2010; 46: 15–22.

26.

Okada

. Activation energy of mullitization from various starting materials. J Eur Ceram Soc 2008; 28: 377–382.

27.

ISBN 0-580-17809-9. Methods of testing refractory materials sampling and physical tests, British Standard Institution.

28.

de la Lastra

Leblud

Leriche

. K_IC calculations for some mullite-zirconia composites prepared by reaction sintering. J Mater Sci Lett 1985; 4: 1099–1101.

29.

Skinner

Cook

Potter

. Effect of TiO₂, Fe₂O₃ and alkali on mineralogical and physical properties of mullite type and mullite forming Al₂O₃-SiO₂ mixtures. J Am Ceram Soc 1953; 36: 349–356.

30.

Okada

Otsuka

. Characterization of the spinel phase from SiO₂-Al₂O₃ xerogels and the formation process of mullite. J Am Ceram Soc 1986; 69: 652–656.

31.

Roy

Bandyopadhyay

Das

. Studies on the formation of mullite from diphasic Al₂O₃-SiO₂ gel by Fourier transform infrared spectroscopy. Iran J Chem Chem Eng 2011; 30: 65–71.

32.

Greenwood

Earnshaw

. Chemistry of the elements, 2nd ed. Oxford, UK: Butterworth Heinemann, 1997.

33.

Aggarwal

Goswami

. An oxide of tervalent nickel. J Phys Chem 1961; 65: 2105–2105.

34.

www.webelements.com (accessed 5 April 2010).

35.

Jahn

Teller

. Stability of polyatomic molecules in degenerate electronic states I orbital degeneracy. Proc R Soc Lond Ser A 1937; 161: 220–235.

36.

Shanon

. Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides. Acta Crystallogr Sec A 1976; A32: 751–767.

37.

Okadaa

Kanedaa

Kameshimaa

. Crystallization kinetics of mullite from polymeric Al₂O₃–SiO₂ xerogels. Mater Lett 2003; 57: 3155–3159.

38.

Tan

Ding

Zhang

. Activation energy for mullitization of gel fibres obtained from aluminium isopropoxide. Bull Mater Sci 2012; 35: 833–837.

39.

Pask

. Importance of starting materials on reactions and phase equilibria in the Al₂O₃-SiO₂ system. J Eur Ceram Soc 1996; 16: 101–108.

40.

Thomson

. Mullite formation from nonstoichiometric diphasic precursors. J Am Ceram Soc 1991; 74: 2382–2387.

41.

Ossaka

. Tetragonal mullite-like phase from co-precipitated gels. Nature 1961; 191: 1000–1001.

42.

Thomson

. Tetragonal to orthorhombic transformation during mullite formation. J Mater Res 1991; 6: 819–824.

43.

Gupta

Dubikova

French

. Effect of CO₂ gasification on the transformation of coke minerals at high temperatures. Energy Fuels 2007; 21: 1052–1061.