J Therm Anal Calorim (2013) 112:345–351 DOI 10.1007/s10973-012-2736-1
Kinetic study of the non-isothermal crystallization process of hematite in ceramic glazes obtained from CRT wastes I. Laza˘u • S. Borca˘nescu • C. Pa˘curariu C. Vancea
•
Received: 27 June 2012 / Accepted: 24 September 2012 / Published online: 20 October 2012 Akade´miai Kiado´, Budapest, Hungary 2012
Abstract Aventurine type crystallization glazes containing hematite (a-Fe2O3) as reddish-brown crystals with golden reflexes are well known for their highly decorative effect. This effect is conditioned by the vitreous matrix composition selected to have a positive influence over the hematite crystallization process producing crystals with suitable shape and size. This paper promote the use of a glass waste—cathode ray tube (CRT) the funnel glass section (15760 wt%)—along with the traditional raw materials (borax, boric acid, quartz, and iron oxide) to obtain the aventurine frits. The mixture of CRT waste and raw materials containing 15720 % Fe2O3 was melted at a temperature of 1,250 C with 30 min soaking time. Black granular frits were obtained after pouring the melts in cold water. The glaze slurry was prepared using the obtained frits (95 %) and kaolin (5 %) as suspension material. The crystallization kinetics of the aventurine type glaze has been investigated using the linear integral isoconversional methods described by Kissinger–Akahira–Sunose, Ozawa– Flynn–Wall, Starink and Tang, and also the non-linear integral isoconversional method described by Vyazovkin. The apparent activation energy of the hematite crystallization in the studied aventurine glazes ranges with the crystallized fraction between 190 and 262 kJ mol-1 for the frit with 3.82 % LiF and between 256 and 281 kJ mol-1 for the frit 3.82 % CaF2, respectively.
The paper was presented at the AICAT 2012 conference. I. Laza˘u S. Borca˘nescu (&) C. Pa˘curariu C. Vancea Faculty of Industrial Chemistry and Environmental Engineering, ‘‘Politehnica’’ University of Timis¸ oara, Piat¸ a Victoriei no. 2, 300006 Timis¸ oara, Romania e-mail:
[email protected]
Keywords Aventurine glazes Hematite crystallization kinetics CRT waste
Introduction The term aventurine glaze is used to define the glazes containing macroscopic crystals which produce a specific decorative effect of scintillation due to the difference in the light-reflection coefficients of the crystalline inclusions and the glass itself and depends on the size, shape, quantity, and arrangement of these crystals and on the observation angle [1, 2]. The elements able to crystallize as oxides and give rise to the aventurine effect are iron, chromium, manganese, uranium, etc. The optimum concentration varies for each type of composition. If low, it remains into the glass matrix and does not produce the effect, while if too high can result in large crystals on the surface and give a metallic appearance, instead of the aventurine effect [3–5]. Fe2O3 is the most reported oxide in the literature as effective generator of aventurine glazes [6, 7]. The mechanism responsible for the aventurine effect in these glazes involves two steps: dissolution in the melt and subsequent hematite (–Fe2O3) crystallization at lower temperature. This mechanism depends on the melt oxide composition and viscosity which affects the hematite proportion and crystal size [8]. In the present paper, CRT waste was used together with the usual raw materials to obtain the glass matrix for the aventurine glaze. A CRT glass unit is composed of three different parts: the viewing section (known as screen, panel, or face plate), made of barium oxide glass, the funnel glass and the neck glass, both containing lead oxide to shield against X-ray radiation released by the high anode
123
I. Laza˘u et al.
346
voltage. The CRT funnel glass composition make it valuable raw material and could represent a way to recycle this type of E waste and to reduce the energy consumption and to shorten production times. This idea derives from the generally accepted concept for which the use of recycled materials is environmentally preferable to that of virgin raw materials [9–11]. The aim of this paper was to establish the optimal composition for the frits containing CRT glass waste to obtain ceramic glazes for terracotta products, having a firing temperature interval between 95071,000 C. The decorative effect was improved using nucleation agents that have a favorable influence upon the crystallization processes. The kinetic studies for the hematite crystallization follow the correlation between the activation energies of the crystallization processes with the decorative effect for two frits, containing different nucleation agents, LiF and CaF2, respectively.
Experimental The sample preparation The CRT waste was pre-heated at 450 C and then poured in cold water to assure an easy milling. The grinding process was conducted so that the frit powder particles are less than 100 lm on diameter. The CRT composition, determined by X-ray fluorescence using a Niton XL 3 analyzer was (wt%): 58.00 SiO2, 7.03 Na2O, 8.57 K2O, 3.64 CaO, 2.18 MgO, 3.47 BaO, 12.99 PbO, 4.12 Al2O3. The CRT glass powder together with the common raw materials was used to prepare the 12 frits having the composition in the following range (wt%): 30–48 SiO2, 5–11 Na2O, 1–5 K2O, 0.5–2 CaO, 0.4–1.3 MgO, 0.5–2 BaO, 2–8 PbO, 0.6–2.5 Al2O3, 15–40 B2O3, 16–20 Fe2O3. In some of the samples, NaNO3 was used as oxidant and LiF and CaF2 as nucleation agents. The raw materials mixture have been homogenized and melted in an electric furnace with SiC heating elements at 1,280 C for 30 min. The frits obtained after pouring the melts into cold water were used to prepare ceramic glazes containing 95 % frit and 5 % kaolin. The terracotta plates glazed by immersion were dried and then fired in a Nabertherm electric furnace at 980 C for 30 min. For the kinetic study, two frits C8 and C9 were chosen. The composition of the frits (wt%) is: 47.77 SiO2, 5.81 Na2O, 1.33 K2O, 0.56 CaO, 0.34 MgO, 0.54 BaO, 2.01 PbO, 0.64 Al2O3, 21.60 B2O3, 19.39 Fe2O3, and 3.82 LiF in C8 and 3.82 % CaF2 in C9 sample, respectively, (over 100 %) as nucleation agents. The frits were grinded so that the powder has particles less than 100 lm on diameter.
123
Characterization methods The shape, size, and distribution in the glass matrix for the hematite crystals were analyzed by optical microscopy in reflected light, using a Guangzhou L 2020A microscope with digital camera. The kinetic study of the a–Fe2O3 was studied using a Netzsch-STA 449C TG/DTA instrument in dry nitrogen atmosphere at a low rate of 20 mL min-1 at heating rates of 4, 6, 8, 10, 12, and 14 C min-1 using platinum crucibles. The hematite presence in the investigated samples was studied by XRD analysis using a DRON 3 diffractometer with CuKa radiation. Theoretical For non-isothermal conditions, considering the Arrhenius low for the temperature dependence of rate constant and the linear change of temperature with time, the solid-state reaction rate is described by Eq. (1) [12–24]: da A E ¼ e RT f ðaÞ; dT b
ð1Þ
where a is the extent of conversion, T is the temperature, A is the pre-exponential factor, b is the heating rate, E is the activation energy, R is the universal gas constant and f ðaÞ is the reaction model. Integration of Eq. (1) leads to: gðaÞ ¼
Za 0
da A ¼ f ðaÞ b
ZT
E
eRT dT:
ð2Þ
0
The integral in Eq. (2) has no analytical solution and therefore, a number of approximations were made, resulting different methods for the activation energy calculus. Using advanced isoconversional methods, the activation energy can be calculated at any particular value of a, which allows to draw conclusions on the mechanism governing the transformation [4, 14, 15]. The integral isoconversional methods are based on the integration of Eq. (2) and can be classified as linear and non-linear integral methods [14, 21]. The linear integral methods involve the plot of a logarithmic function (which depends on the approximation for the temperature integral used) versus 1=Ta . The general form of these linear equations is [13]: ln
b Ea ¼ A þ C; j Ta RTa
ð3Þ
where j and A are parameters depending on the approximation of the temperature integral employed, C is a constant and the subscript a designates values related to a given extent of conversion.
Kinetic study of the non-isothermal crystallization process
347
Two of the linear integral isoconversional methods, considered in the literature the most accurate, were used in this paper. Using the approximation made by Murray and White [16], the two parameters j and A become: j = 2 and A = 1, and the general Eq. (3) leads to the Kissinger– Akahira–Sunose (KAS) Eq. (4) [17, 18]: ln
b Ea ¼ þ C: Ta2 RTa
b Ta1:92
¼ 1:0008
aðTÞ ¼
Ea þ C: RTa
j6¼i
where: IðEa ; Tai Þ ¼
exp
Ea dT RT
ð7Þ
0
The minimization procedure is repeated for each value of a, to find the dependence of the activation energy on the extent of conversion. In our paper, the integral (Eq. 6) was evaluated using three different approximations (Eqs. 8–10): Gorbachev [29]: Ea IðEa ; Tai Þ ¼ exp dT RT 0 RTai2 Ea exp : ¼ Ea þ 2RTai RTai ZTai
ð8Þ
Cai and Liu [22, 23]: Ea IðEa ; Tai Þ ¼ exp dT RT 0 2 3 2RTai 2 RTai 6 1 Ea 7 Ea ¼ : 4 2 5 exp RT Ea RTai 1 5 Eaai ZTai
AðTÞ AðTotalÞ
ð11Þ
where aðTÞ is the crystallized fraction at the temperature T, AðTÞ is the area at the temperature interval DT and AðTotalÞ is the total area of the crystallization peak.
ð5Þ
To increase the accuracy of activation energy assessment, Vyazovkin [25–28] developed a non-linear isoconversional method. For a set of n experiments carried out at different heating rates, the activation energy can be determined at any value of a by finding the value of Ea , which minimizes the function: n X n X IðEa ; Tai Þbj ð6Þ X¼ IðEa ; Taj Þbi i¼1 j¼1
ZTai
ð10Þ
The crystallized fraction aðTÞ was determined from the DTA curves using Eq. (11):
ð4Þ
One of the most accurate equations was proposed by Starink [13, 15]. In this case, j = 1.92 and A = 1.0008, and the general linear Eq. (3) become: ln
Ea dT IðEa ;Tai Þ ¼ exp RT 0 RTai2 Ea =RTai þ 0:66691 Ea : exp ¼ Ea Ea =RTai þ 2:64943 RTai ZTai
Results and discussion All the investigated compositions were elaborated at 1,280 C, their viscosity depending on the SiO2 amount. The presence of LiF and CaF2 used as nucleation agents determined a decrease of the melt viscosity. The crystallization degree and aspect of the glazes made using the frits depends on their composition. The best results were obtained for the C8 and C9 frits, as shown in the photo images presented in Fig. 1. The decorative effect of these glazes, caused by the light reflection on the hematite crystals depends on their size and distribution and also on the thickness of the glaze and the texture of the ceramic support. Figure 2 presents the optical microscopy images of the glaze obtained with C8 frit, for three different areas (a, b, and c), shown in Fig. 1. The presence of the hematite (a-Fe2O3) as unique crystalline phase in the studied glazes was confirmed by X-ray diffraction. Hematite crystals with rhombohedral symmetry, having about 10–120 lm can be observed. The kinetic study for the crystallization process was conducted for the two chosen frits (C8 and C9). Figure 3 presents the DTA crystallization curves of frit C8, recorded at six different heating temperatures. The crystallized fraction a, evaluated from the DTA curves using Eq. (11) is presented in Fig. 4 as a function of temperature, for all the heating rates used.
Chen and Liu [24]:
Kissinger–Akahira–Sunose method
ð9Þ
The activation energies of the crystallization process Ea were calculated from the slope of the linear fitted function of lnðb=Ta2 Þ versus Ta1 using Eq. (4). The KAS plots are shown in Figs. 5 and 6 for different crystallized fraction for the frit C8 and C9.
123
348
I. Laza˘u et al.
Fig. 1 Visual aspect of the glazes made with a C8 and b C9 frits
Fig. 2 Optical microscopy images for different areas of the glaze obtained with C8 frit
Starink method Using the Starink method, the values of Ea was calculated from the slope of the linear fitted function of lnðb=Ta1:92 Þ versus Ta1 [Eq. (5)] for constant crystallized fraction. The Starink plots for different crystallized fractions are shown in Figs. 7 and 8 for the frit C8 and C9. Figure 9 shows the variation of the activation energy as a function of crystallized fraction, evaluated by the methods of KAS, Starink and Vyazovkin [using Gorbachev (Eq. 8), Agrawal and Sivasubramanian (AS) (Eq. 9), respectively Cai (Eq. 10) approximations] for both C8 and C9 frits.
123
The activation energies evaluated by KAS, Starink, and Vyazovkin methods show very close values in the entire range of the crystallized fraction for both C8 and C9 frits. These results are in agreement with the literature data [26, 27] and recommends the methods used in this paper as the most precise methods for activation energies calculation. The activation energy varies with the crystallized fraction regardless of the method used. This behavior confirms the complex mechanism of the crystallization process. The values of the apparent activation energy for the hematite crystallization in the investigated aventurine glazes ranges between 190 and 262 kJ mol-1 for the frit
Kinetic study of the non-isothermal crystallization process
349
819.1 –10.8
14 K min–1
10K min–1
800.2
2 α
In( β /T
12 K min–1
812.7
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
–11.2
–11.4
8K min–1
793.1
α
–11.0
(
DTA signal/a.u. Exo
816.2
–11.6
6K min–1
–11.8 1.14
740
760
780
800
820
840
860
880
1.16
1.20
1.18
1.22
1.24
1/Tα× 103/K–1
900
T/K
Fig. 3 DTA crystallization curves of frit C8 at different heating rates
Fig. 6 Plots of lnðb=Ta2 Þ versus Ta1 at different crystallized fractions for frit C9
–10.2
1.0
α
–10.4
Crystallized fraction α
0.8 6 K min–1 8 K min–1
0.6
10 K min–1
In( β /T
1.92 α
(
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
–10.6
–10.8
12 K min–1 0.4
–11.0
14 K min–1
0.2
–11.2 1.16
1.18
1.20
1.22
1.26
1.24
1.28
1/Tα× 103/K–1
0.0 740
760
780
800
820
840
860
880
T/K
Fig. 7 Plots of lnðb=Ta1:92 Þ versus Ta1 at different crystallized fractions for frit C8
Fig. 4 The variation of the crystallized fraction a versus temperature for the frit C8 at different heating rates
–10.2
α –10.4
–10.6
α
–10.8
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
–11.0
In( β /T
2 α
( –11.2 –11.4 –11.6 –11.8 1.18
1.20
1.22
1.24
1.26
1.28
1/Tα× 103/K–1
Fig. 5 Plots of lnðb=Ta2 Þ versus Ta1 at different crystallized fractions for frit C8
In( β /T
1.92 α
(
0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
–10.6 –10.8 –11.0 –11.2 1.14
1.16
1.18
1.20
1.22
1/Tα× 103/K–1
Fig. 8 Plots of lnðb=Ta1:92 Þ versus Ta1 at different crystallized fractions for frit C9
C8 and between 256 and 281 kJ mol-1 for the frit C9. The slightly lower values for the activation energy for the C8 frit compared to those for C9 frit is due to the different
123
I. Laza˘u et al.
Eα /kj mol–1
350 300 290 280 270 260 250 240 230 220 210 200 190 180 170 160 150 140
C8–KAS C8–Starink C8–V(Gorbachev) C8–V(Cai) C8–V(AS) C9–KAS C9–Starink C9–V(Gorbachev) C9–V(Cai) C9–V(AS) 0.0
0.2
0.4
0.6
0.8
1.0
α
Fig. 9 Variation of Ea as a function of a evaluated by KAS, Starink, and Vyazovkin methods for the frits C8 and C9
behavior of the two nucleation agents used: LiF and CaF2, respectively. The Li? cation from the C8 frit has a lower ionic radius and therefore, a higher mobility than the Ca2? cation present in C9 frit. The obtained results from the kinetic studies for the hematite crystallization prove the usefulness of the nucleation agents that reduce the activation energy, to enhance the decorative aspect of those aventurine glazes.
Conclusions The CRT waste can be used as valuable raw material to obtain aventurine glazes and could represent a way to recycle this type of E waste, to reduce the energy consumption and to shorten production times. The decorative effect on these glazes is caused by the uneven distribution of hematite crystals (Fe2O3), depending on the glass composition. The kinetic study for the crystallization process was conducted for the two chosen frits (sample C8 and sample C9). The apparent activation energy of the hematite crystallization in the studied aventurine glazes ranges with the crystallized fraction between 190 and 262 kJ•mol-1 for the frit C8 and between 256 and 281 kJ•mol-1 for the frit C9, respectively. The activation energies evaluated by KAS, Starink, and Vyazovkin methods show very close values in the entire range of the crystallized fraction for both samples C8 and C9. The activation energy for the crystallization process of the hematite in the C8 frit was slightly lower than that for the C9 frit. This behavior can be explained based on the Li? cation present in C8 frit, having a lower ionic radius and a higher mobility than the Ca2? ion present in C9 frit. The activation energy varies with the crystallized fraction regardless of the method used confirming the complex mechanism of the hematite crystallization process in the aventurine glaze.
123
Acknowledgements This work was partially supported by the strategic Grant POSDRU/89/1.5/S/57649, Project ID 57649 (PERFORM-ERA), co-financed by the European Social Found—Investing in People, within the Sectoral Operational Programme Human Resources Development 2007–2013.
References 1. Wanie W. Kristalline glasuren und Kristallglasuren. Glas-EmailKeramo-Techik. 1965;16(11):413–7. 2. Levitskii IA. Mechanism of phase formation in aventurine glaze. Glass Ceram. 2001;58(5–6):223–6. 3. Dakhai S, Orlova LA, Mikhailenko NYu. Types and compositions of crystalline glazes (a review). Glass Ceram. 1999;56(5–6): 177–80. 4. Dvornichenko IN, Matsenko SV. Production of iron-containing crystalline glazes. Glass Ceram. 2000;57(1-2):67–9. 5. SYu Eliseev, Rodtsevich SP, Dostanko EV. Enamel with the aventurine effect. Glass Ceram. 2000;57(3–4):140–2. 6. Casasola R, Ma Rinco’n J, Romero M. Glass–ceramic glazes for ceramic tiles: a review. J Mater Sci. 2012;47:553–82. 7. Pa˘curariu C, Laza˘u RI, Laza˘u I. Aventurine decorative glazes— kinetic study of a-Fe2O3 crystallization. Rom J Mater. 2011; 41(4):346–51. 8. Pa˘curariu C, Laza˘u RI, Laza˘u I, Tit¸ a D, Dumitrel A. Non-isothermal crystallization kinetics of some aventurine decorative glaze. J Therm Anal Calorim. 2011;105:435–41. 9. Andreola F, Barbieri L, Corradi A, Ferrari AM, Lancellotti I, Neri P. Recycling of EOL CRT glass into ceramic glaze formulations and its environmental impact by LCA approach. Int J LCA. 2007;12(6):448–54. 10. Dima V, Eftimie M, Volceanov A, Voicu G, Dinca˘ R. Possibilities for recovery of some glass waste from cathode ray tubes. Rom J Mater. 2012;42:76–81. 11. Andreola F, Barbieri L, Corradi A, Lancellotti I. CRT glass state of the art—a case study: recycling in ceramic glazes. J Eur Ceram Soc. 2007;27:1623–9. 12. Vyazovkin S, Wight CA. Model-free and model-fitting approaches to kinetic analysis of isothermal and non-isothermal data. Thermochim Acta. 1999;340–341:53–68. 13. Starink MJ. The determination of activation energy from linear heating rate experiments: a comparison of the accuracy of isoconversion methods. Thermochim Acta. 2003;404:163–76. 14. Vyazovkin S, Burnham AK, Criado JM, Pe´rez-Maqueda LA, Popescu C, Sbirrazzuoli N. ICTAC kinetics committee recommendations for performing kinetic computations on thermal analysis data. Thermochim Acta. 2011;520:1–19. 15. Starink MJ. Activation energy determination for linear heating rate experiments: deviations due to neglecting the low temperature end of the temperature integral. J Mater Sci. 2007;42:483–9. 16. Murray P, White J. Kinetics of the thermal dehydration of clays. IV. Interpretation of the differential thermal analysis of the clay minerals. Trans Brit Ceram Soc. 1955;54:204–38. 17. Kissinger HE. Reaction kinetics in differential thermal analysis. Anal Chem. 1957;29:1702–6. 18. Akahira T, Sunose T. Trans joint convention of four electrical institutes, Research report, Chiba Institute of Technology. Sci Technol. 1971;16:22–31. 19. Criado JM, Sanchez-Jimenez PE, Perez-Maqueda LA. Critical study of the isoconversional methods of kinetic analysis. J Therm Anal Calorim. 2008;92:199–203. 20. Cai JM, Liu RH. Precision of integral methods for the determination of the kinetic parameters. Use in the kinetic analysis of solid-state reactions. J Therm Anal Calorim. 2008;91(1):275–8.
Kinetic study of the non-isothermal crystallization process 21. Farjas J, Roura P. Isoconversional analysis of solid state transformations. J Therm Anal Calorim. 2011;105:757–66. 22. Cai JM, Liu RH. Non-isothermal kinetics in solids. The precision of some integral methods for the determination of the activation energy without neglecting the temperature integral at the starting temperature. J Therm Anal Calorim. 2008;94: 313–6. 23. Pratap A, Rao TLS, Lad KN, Dhurandhar HD. Isoconversional vs. model fitting methods. A case study of crystallization kinetics of a Fe-based metallic glass. J Therm Anal Calorim. 2007;89(2): 399–405. 24. Chen HX, Liu NA. Approximations for the temperature integral. Their underlying relationship. J Therm Anal Calorim. 2008; 92(2):573–8.
351 25. Burnham AK, Dinh LN. A comparison of isoconversional and model-fitting approaches to kinetic parameter estimation and application predictions. J Therm Anal Calorim. 2007;89(2):479–90. 26. Vyazovkin S, Dollimore D. Linear and nonlinear procedures in isoconversional computations of the activation-energy of non-isothermal reactions in solids. J Chem Inf Comp Sci. 1996;36:42–5. 27. Vyazovkin S. Advanced isoconversional method. J Therm Anal Calorim. 1997;49:1493–9. 28. Vyazovkin S, Dranca I. Isoconversional analysis of combined melt and glass crystallization data. Macromol Chem Phys. 2006;207:20–5. 29. Gorbachev VM. A solution of the exponential integral in the nonisothermal kinetics for linear heating. J Thermal Anal Calorim. 1975;8(2):349–50.
123
