- Email: [email protected]
Synthesis and thermal behavior of La9.33Si2Ge4O26 apatite for SOFCs
Journal of Alloys and Compounds 536S (2012) S480–S484
Contents lists available at SciVerse ScienceDirect
Journal of Alloys and Compounds journal homepage: www.elsevier.com/locate/jallcom
Synthesis and thermal behavior of La9.33 Si2 Ge4 O26 apatite for SOFCs R. Serra a,∗ , C. Alves b , J.C. Oliveira a , F.A.C. Oliveira b , T. Marcelo b , J. Mascarenhas b , B. Trindade a a b
CEMUC, Mechanical Engineering Department, University of Coimbra, Rua Luís Reis Santos, 3030-788 Coimbra, Portugal Laboratório Nacional de Energia e Geologia I.P., Product Engineering Unit, Estrada do Pac¸o do Lumiar, 1649-038 Lisboa, Portugal
a r t i c l e
i n f o
Article history: Received 24 June 2011 Received in revised form 14 November 2011 Accepted 21 November 2011 Available online 28 November 2011 Keywords: SOFC Electrolytes Apatite Mechanical alloying Thermal behavior
a b s t r a c t Powders of La2 O3 , GeO2 and SiO2 were dry milled in a planetary ball mill with different rotation speeds (150–350 rpm) and increasing milling times up to 35 h in order to obtain the La9.33 Si2 Ge4 O26 apatite phase at room temperature. The results showed that the higher the rotation speed the lower the time required for the formation of the apatite phase. No reaction between the starting powders was observed at 150 rpm. Thermal analysis of the unreacted powders milled at 150 rpm showed formation of the apatite phase around 800 ◦ C, with enthalpies ranging from 43.5 and 48.6 kJ mol−1 . An activation energy Ea of 65 kJ mol−1 was obtained applying the Kissinger equation. The mean Avrami exponent n calculated was 1.5, indicating that the apatite phase transformation occurs by a diffusion controlled process. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Solid oxide fuel cells (SOFCs), based on an oxide ion-conducting electrolyte, offer a versatile, efficient, clean and environmental friendly technology to electrochemically generate electricity. Moreover, SOFCs provide some advantages over traditional energy conversion systems, namely fuel adaptability, reliability, modularity and durability [1]. Current commercially available SOFCs operate at temperatures higher than 800 ◦ C, demanding materials and components able to withstand such severe conditions. Accordingly, intensive research work has been carried out in the last years with the aim of decreasing the service temperature and therefore increasing the durability and lowering the cost of materials and components [2]. One way to achieve this goal is to develop new electrolyte materials with higher ionic conductivity at temperatures in the range 500–700 ◦ C, for intermediate temperature solid oxide fuel cells (IT-SOFCs). Amongst the potential materials suitable for such application are apatite-type rare earth based oxides, such as R-doped lanthanum oxides of general formula La9.33 (RO4 )6 O2 with R = Ge, Si, which exhibit high ionic conductivity and low activation energy at moderate temperatures, when compared to the yttria-stabilized zirconia (YSZ) electrolyte commonly used [3,4]. These materials are currently produced by solid-state reaction methods [5]. Several problems in processing such apatite materials include the high temperature required, poor control of
∗ Corresponding author. Tel.: +351 239790745. E-mail address: [email protected] (R. Serra). 0925-8388/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jallcom.2011.11.091
morphology, grain growth [6,7] and germanium volatilization (which is significant above 1250 ◦ C) [8,9]. Mechanical alloying (MA) [10] is a potential alternative processing method to obtain the apatite like lanthanum oxide [3,11]. During the MA process, heavy deformation is introduced into the particles, increasing the defect density, which accelerates the diffusion process by decreasing the diffusion distances. Therefore, sintering is achieved at lower temperatures than the ones required by conventional routes. In previous works, Si- and Ge-doped lanthanum oxide materials with apatite-type phase were produced by MA and subsequent annealing [12,13]. A dependence of apatite-type phase formation on the MA parameters was shown, the higher the MA rotation speed and/or processing time, the lower the annealing temperature required for its formation. In the present work, the influence of the milling atmosphere as well as of the particle size distribution of the raw materials on the structural evolution of powders during MA has been studied. The thermal behavior of the mechanically alloyed mixtures was studied by means of differential scanning calorimetry (DSC). 2. Experimental Powders of crystalline La2 O3 (99.9% purity), SiO2 (quartz with 99.4% purity) and GeO2 (99.9% purity), were used as starting materials. Mixtures with 7La2 O3 :3SiO2 :6GeO2 molar ratios were used in order to obtain the La9.33 Si2 Ge4 O26 phase. The blends were dry milled in a Fritsch planetary ball mill using a 250 ml hardened steel vial and 15 balls with 20 mm diameter of the same material. About 24.5 g of the powder blend was loaded into the vial as to get a ball-to-powder weight ratio of 20:1. Some of the experiments were carried out under protective atmosphere (Ar at 2 bar), the remaining being performed under atmospheric air. The powders
R. Serra et al. / Journal of Alloys and Compounds 536S (2012) S480–S484 Table 1 Characteristic values of the particle size distribution of the raw materials in the as-received condition and after pre-milling. d (m) La2 O3
As-received
Table 2 Lattice parameters of the apatite phase obtained for different milling conditions.
Pre-milled
d10 d50 d90
2.0 9.4 49.7
0.7 2.4 11.6
SiO2
d10 d50 d90
2.8 26.4 118.3
1.0 4.8 15.1
GeO2
d10 d50 d90
2.5 37.6 72.8
0.6 2.0 4.2
were mechanically alloyed at rotational speeds of vial of 150, 250 and 350 rpm for 15–35 h either in the as-received condition or after milling the starting materials separately at 350 rpm for 10 min (hereafter referred to as “pre-milled condition”). With this procedure, it was expected to start from more homogenous powder blends to promote the apatite single phase formation (by eliminating large particles and agglomerates) as well as to achieve MA mixtures with smaller particle size population that in principle can be easily sintered. Both raw materials and as-milled mixtures were characterized by means of laser scattering (particle size distribution), X-ray diffraction (XRD) with Co K␣ radiation and scanning electron microscopy (SEM). DSC was performed up to 1200 ◦ C under static air. Different heating rates, in the range 10–40 ◦ C min−1 , were used in order to determine the thermal parameters corresponding to the formation of apatite phase using the Kissinger method [14].
3. Results and discussion Fig. 1 shows scanning electron micrographs of the raw materials, as-received. It can be seen that they have different particle size distributions and morphologies. The GeO2 particles are round whereas the ones of La2 O3 and SiO2 show some flat surfaces (cleavage) typical of brittle materials. Table 1 shows the characteristic values of the particle size distribution of the oxide particles as-received and after pre-milling; dx (x = 10, 50, 90) means that the x volume fraction of the measured particles is at or below the diameter d. A significant decrease of d10 , d50 and d90 values of the starting materials was observed after premilling, as a result of the powders brittleness and deagglomeration, which may enhance sintering behavior. Fig. 2(a) and (b) show the particle size distributions (differential and cumulative curves) of the MA mixtures synthesized at 350 rpm in argon with and without pre-milling. Both MA mixtures are characterized by multimodal distributions with d50 values around 6 and 7 m for mixtures with and without pre-milling, respectively. The mixtures processed at 250 and 350 rpm were milled using the minimum time for lanthanum germanosilicate formation, as previously reported [12,13]. Fig. 3(a) shows the XRD patterns of the mechanically alloyed powder mixtures obtained in this work. As anticipated, formation of the apatite phase was obtained after 15 and 25 h at 350 and 250 rpm, respectively, while the mixture milled at 150 rpm for 35 h only shows peaks indexed to the starting materials.
S481
a (Å) c (Å) c/a
350 rpm (as-received) Milled in air
350 rpm (pre-milled) Milled in air
350 rpm (pre-milled) Milled in argon
9.92 7.25 0.731
9.96 7.25 0.728
9.90 7.27 0.734
The effect of pre-milling of the starting powders and of milling atmosphere was studied on samples prepared at 350 rpm for 15 h. No significant differences on apatite-type structure were observed by X-ray diffraction after MA (Fig. 3(b)). The apatite phase was formed after 15 h of milling independently of the milling atmosphere (Ar or air) and the initial state of the raw materials (asreceived or pre-milled). The crystallite size of the apatite phase was determined by the Scherrer equation. All the samples are nanocrystalline with very similar crystallite sizes (in the range of 12–18 nm). The hexagonal lattice parameters of the La9.33 Si2 Ge4 O26 apatite phase were determined after MA from the XRD patterns shown in Fig. 3. The values obtained for a and c are summarized in Table 2. The a parameter in all samples is higher than the value reported in the ˚ As the a value of ICCD card of the La9.33 Si3 Ge3 O26 phase (a = 9.86 A). apatite phase increases with Ge content, this result confirms that more than half of the positions in the Si/Ge–O tetrahedral structures of the unit cell are occupied by Ge atoms, in agreement with the SiO2 /GeO2 molar ratio of 1/2 used in the starting mixtures. Within the experimental error, the c parameter values of the mixtures after MA are similar to the one reported for the La9.33 Si3 Ge3 O26 phase ˚ This result may be due to either a weak dependence of (c = 7.27 A). the c parameter on the Ge/Si atomic ratio in the apatite phase, as compared to the a parameter or to a distortion of the unit cell with Ge content increase. Thermal behavior of the MA mixtures prepared at 150 rpm under different conditions was evaluated by DSC analysis. Fig. 4(a) shows the curve obtained at 30 ◦ C min−1 as a typical example. Fig. 4(b) shows the influence of the heating rate on the temperature of the apatite formation peak. No peaks were observed during cooling and therefore this part of the curves is not shown. The DSC curves show two endothermic peaks in the ranges 360–400 and 520–560 ◦ C (Fig. 4(a)), depending on the heating rate, related to the decomposition process of the La(OH)3 phase into LaOOH and subsequent dehydration with formation of La2 O3 . Typically, decomposition of the hydroxide and hydroxycarbonate occurs at 280 ◦ C and dehydration of LaOOH occurs at 420 ◦ C whereas decomposition of La2 O2 CO3 occurs at 600 ◦ C [15]. It is likely that the difference in recorded temperatures can be attributed to differences in moisture and CO2 contents. The presence of La(OH)3 was not confirmed by XRD analysis of the 150 rpm MA mixtures (probably its content is below the equipment detection limit or amorphous nature). However, its presence in La2 O3 powder is almost unavoidable taking into account the high reactivity of La2 O3 in the presence of moisture and carbon
Fig. 1. Powders used in this work (as-received): (a) La2 O3 ; (b) SiO2 ; (c) GeO2 .
S482
R. Serra et al. / Journal of Alloys and Compounds 536S (2012) S480–S484
Fig. 2. Particle size distributions of the MA mixtures after milling at 350 rpm for 15 h; comparison between as-received and pre-milled starting powders.
Fig. 3. XRD patterns of the mechanically alloyed mixtures: (a) influence of milling rotation speed and time; (b) influence of pre-milling and protective atmosphere (350 rpm).
Fig. 4. DSC curves of the 150 rpm milled sample: (a) 30 ◦ C min−1 ; (b) apatite formation peaks at different heating rates.
R. Serra et al. / Journal of Alloys and Compounds 536S (2012) S480–S484
S483
Fig. 6. Kissinger plot of the apatite peaks obtained from a set of heating rates.
Fig. 5. XRD patterns of the mechanically alloyed mixtures at 150 rpm for 35 h after DSC runs up to 900 and 1200 ◦ C.
dioxide. Some authors [15,16] claim that an amorphous phase of formula La2 (OH)4 CO3 ·nH2 O can be formed during the hydration of the La2 O3 that would explain La(OH)3 absence in the La2 O3 starting material. Tao and Irvine [17] used differential thermal analysis (DTA) to study the thermal behavior of the amorphous La10 Si6 O27 and La9.33 Si6 O26 phases obtained by sol–gel process. They observed, for each composition, just one exothermic low intensity peak with Tp peak temperature close to 959 ◦ C (in the case of La10 (SiO4 )6 O3 ) and 874 ◦ C (for La9.33 Si6 O26 phase) corresponding to the crystallization or formation peak of the lanthanum silicate phase. Tzvetkov and Minkova [18] have reported DTA and TG results obtained from powder mixtures of lanthanum oxide and amorphous silica synthesized by mechanical alloying. They observed that an amorphous phase was formed after MA as well as the existence of two exothermic peaks in the DTA traces, one in the range 855–870 ◦ C and another in the range 935–940 ◦ C. Both peaks were ascribed to apatite-type lanthanum silicate formation, the first from the amorphous precursor phase and the second from unreacted La2 O3 and SiO2 milled powders. In the present work the first exothermic DSC peak occurs in the temperature range 810–825 ◦ C, depending on the heating rate, and is ascribed to the formation of apatite. In a previous work on the annealing of MA’ed powder mixtures with and without germanium [12] it was shown that Ge is responsible for lowering the temperature of formation of the apatite phase. XRD analysis was performed after DSC runs in order to ascribe the exothermic peak to the formation of apatite (Fig. 5). The XRD pattern obtained for the DSC run up to 900 ◦ C showed the coexistence of the GeO2 , La(OH)3 and the apatite phase (Fig. 5). The presence of La(OH)3 rather than the initial La2 O3 phase is the result of its hydration, which occurred after DSC. Accordingly in addition to apatite, other phases with different chemical compositions are formed upon heating after MA. At higher temperatures, a second exothermic peak at 1110 ◦ C is observed (Fig. 4(a)), which is associated to the formation of La2 GeO5 (ICDD card 073-9203), as can be seen in Fig. 5. Solid state reactions occurring by nucleation and growth can be described by the Johnson–Mehl–Avrami (JMA) model [19]. According to this model, the fraction of transformed material after a hold time at a given temperature is expressed by the equation: x(t) = 1 − exp(−k · t n )
(1)
where x(t) is the volume fraction of the initial material transformed at time t, n is the Avrami exponent which is related to the nucleation rate and the growth morphology, and k is the temperaturedependent reaction rate constant. The Kissinger method is based on the JMA model and uses the highest rate of a transformation at maximum peak (Tp ) for interpreting the DSC results. It requires data recorded at different heating rates so that different values for Tp are obtained and can be interpreted by the following equation:
ln
˛ Tp2
=−
Ea R · Tp
(2)
where Tp is the temperature at maximum peak, ˛ = ∂T/∂t is the heating rate and Ea the activation energy necessary for the reaction. Ea is the slope of the straight line obtained by plotting ln(˛/Tp2 ) vs 1/Tp . The Tp and the enthalpy values of the apatite phase formation are shown in Table 3. As expected for a diffusion controlled reaction, the higher the heating rate the higher the temperature at maximum peak at which the apatite phase is formed. Inversely, the enthalpy decreased with increasing of heating rate (less time to complete the reactions and therefore lower energy release). Based on the Tp values, the Kissinger plot was drawn (Fig. 6) and the Ea was calculated from the slope of the straight line. A value of 65 kJ mol−1 was obtained. The Avrami exponent (n) is usually determined by isothermal DSC/DTA methods. However, Zhaosheng et al. [20] developed a non-isothermal method of multi-scanning heating rate for the determination of the Avrami exponent for an amorphous alloy. In this method, applicable to any phase transformation occurring by a diffusion process, n is given by: n=−
ln[− ln(1 − x1)] − ln[− ln(1 − x2)] ln ˛1 − ln ˛2
(3)
where ˛1 and ˛2 are the heating rates and x1 and x2 are the phase transformation fractions at the same temperature. In the present study, the temperatures of 812, 816 and 822 ◦ C (all in the temperature range of the apatite formation) were considered for 10–20, 20–30 and 30–40 ◦ C min−1 pairs of heating rates, respectively. The values of the Avrami exponent obtained using Eq. (3) are summarized in Table 4. Depending on the heating rates used, n values in the range 1.3–1.6 were obtained. This suggests that phase transformation occurs through diffusion under conditions were negligible nucleation takes place at the initial stages of transformation (i.e. site saturation) [21].
S484
R. Serra et al. / Journal of Alloys and Compounds 536S (2012) S480–S484
Table 3 Tp and enthalpy values of the DSC peak of the apatite phase formed under different heating rates.
◦
Tp ( C) Enthalpy
V = 10 ◦ C min−1
V = 20 ◦ C min−1
V = 30 ◦ C min−1
V = 40 ◦ C min−1
808.9 −21.13 J g−1 (−43.48 kJ mol−1 )
815.3 −22.20 J g−1 (−45.68 kJ mol−1 )
819.3 −23.61 J g−1 (−48.58 kJ mol−1 )
822.9 −23.41 J g−1 (−48.17 kJ mol−1 )
Table 4 Avrami exponent n obtained for three pairs of heating rates used.
References
Heating rates (◦ C min−1 )
ln ˛
x
ln x
10 20
2.302585 2.995732
0.8073 0.4262
−0.21406 −0.85285
1.6
20 30
2.995732 3.401197
0.6208 0.4334
−0.47675 −0.83609
1.3
30 40
3.401197 3.688879
0.6803 0.5178
−0.38522 −0.65817
1.6
n
4. Conclusions Mechanical alloying of La2 O3 , GeO2 and SiO2 powders induced the formation of the La9.33 Si2 Ge4 O26 apatite phase at room temperature at rotation speeds higher than 150 rpm. Neither pre-milling of the starting powders nor the composition of the milling atmosphere had any influence on the structure of the as-milled powders at 350 rpm for 15 h. No reaction between the La2 O3 , GeO2 and SiO2 powders has occurred at 150 rpm for up to 35 h. In this case, the apatite-type lanthanum germanosilicate is formed upon heating in the range 809–823 ◦ C with enthalpies between −43.5 and −48.6 kJ mol−1 . Applying the Kissinger equation to the reaction kinetics, an activation energy Ea of 65 kJ mol−1 was obtained. The mean Avrami exponent n calculated was 1.5 which suggests a diffusion controlled phase transformation with site saturation. Acknowledgment Financial support provided by FCT through project PTDC/EMEPME/102837/2008 is gratefully acknowledged as well as the research fellowship granted to C. Alves.
[1] E.J. Abram, D.C. Sinclair, A.R. West, J. Mater. Chem. 11 (2001) 1978–1979. [2] L. Leon-Reina, E.R. Losilla, M. Martínez-Lara, S. Bruque, M.A.G. Aranda, J. Mater. Chem. 14 (2004) 1142–1149. [3] Y. Higuchi, M. Sugawara, K. Onishi, M. Sakamoto, S. Nakayama, Ceram. Int. 36 (2010) 955–959. [4] S. Nakayama, M. Sakamoto, J. Mater. Sci. Lett. 20 (2001) 1627–1629. [5] H. Yao, J. Wang, D. Hu, J. Li, X. Lu, Z. Li, Solid State Ionics 181 (2010) 41–47. [6] G. Tzvetkov, N. Minkova, Mater. Lett. 39 (1999) 354–358. [7] E. Rodrıíguez-Reyna, A.F. Fuentes, M. Maczka, J. Hanuza, K. Boulahya, U. Amador, J. Solid State Chem. 179 (2006) 522–531. [8] E. Rodríguez-Reyna, A.F. Fuentes, M. Maczka, J. Hanuza, K. Boulahya, U. Amador, Solid State Sci. 8 (2006) 168–177. [9] S.P. Jiang, L. Zhang, H.Q. He, R.K. Yap, Y. Xiang, J. Power Sources 189 (2009) 972–981. [10] C. Suryanarayana, Prog. Mater. Sci. 46 (2001) 1–184. [11] E.J. Abram, C.A. Kirk, D.C. Sinclair, A.R. West, Solid State Ionics 176 (2005) 1941–1947. [12] M.M. Vieira, J.C. Oliveira, A. Cavaleiro, B. Trindade, Mater. Sci. Forum 587–588 (2008) 128–132. [13] M.M. Vieira, J.C. Oliveira, A. Cavaleiro, B. Trindade, Rev. Adv. Mater. Sci. 18 (2008) 344–348. [14] H.E. Kissinger, Anal. Chem. 29 (1957) 1702–1706. [15] P.J. Panteix, I. Julien, D. Bernache-Assollant, P. Abélard, Mater. Chem. Phys. 95 (2006) 313–320. [16] S. Bernal, F.J. Botana, R. García, J.M. Rodríguez-Izquierdo, Thermochim. Acta 66 (1993) 139–145. [17] S. Tao, J.T.S. Irvine, Mater. Res. Bull. 36 (2001) 1245–1258. [18] G. Tzvetkov, N. Minkova, J. Mater. Sci. 35 (2000) 2435–2441. [19] M. Avrami, J. Chem. Phys. 8 (1940) 212–224. [20] T. Zhaosheng, Z. Bangwei, Z. Hen Tan, J. Mater. Sci. 27 (1992) 4021–4023. [21] R. Cahn, P. Haasen, Physical Metallurgy, vol. 2, 4th ed., North-Holland, Amsterdam, 1996, ISBN 0444898751.
Contents lists available at SciVerse ScienceDirect
Journal of Alloys and Compounds journal homepage: www.elsevier.com/locate/jallcom
Synthesis and thermal behavior of La9.33 Si2 Ge4 O26 apatite for SOFCs R. Serra a,∗ , C. Alves b , J.C. Oliveira a , F.A.C. Oliveira b , T. Marcelo b , J. Mascarenhas b , B. Trindade a a b
CEMUC, Mechanical Engineering Department, University of Coimbra, Rua Luís Reis Santos, 3030-788 Coimbra, Portugal Laboratório Nacional de Energia e Geologia I.P., Product Engineering Unit, Estrada do Pac¸o do Lumiar, 1649-038 Lisboa, Portugal
a r t i c l e
i n f o
Article history: Received 24 June 2011 Received in revised form 14 November 2011 Accepted 21 November 2011 Available online 28 November 2011 Keywords: SOFC Electrolytes Apatite Mechanical alloying Thermal behavior
a b s t r a c t Powders of La2 O3 , GeO2 and SiO2 were dry milled in a planetary ball mill with different rotation speeds (150–350 rpm) and increasing milling times up to 35 h in order to obtain the La9.33 Si2 Ge4 O26 apatite phase at room temperature. The results showed that the higher the rotation speed the lower the time required for the formation of the apatite phase. No reaction between the starting powders was observed at 150 rpm. Thermal analysis of the unreacted powders milled at 150 rpm showed formation of the apatite phase around 800 ◦ C, with enthalpies ranging from 43.5 and 48.6 kJ mol−1 . An activation energy Ea of 65 kJ mol−1 was obtained applying the Kissinger equation. The mean Avrami exponent n calculated was 1.5, indicating that the apatite phase transformation occurs by a diffusion controlled process. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Solid oxide fuel cells (SOFCs), based on an oxide ion-conducting electrolyte, offer a versatile, efficient, clean and environmental friendly technology to electrochemically generate electricity. Moreover, SOFCs provide some advantages over traditional energy conversion systems, namely fuel adaptability, reliability, modularity and durability [1]. Current commercially available SOFCs operate at temperatures higher than 800 ◦ C, demanding materials and components able to withstand such severe conditions. Accordingly, intensive research work has been carried out in the last years with the aim of decreasing the service temperature and therefore increasing the durability and lowering the cost of materials and components [2]. One way to achieve this goal is to develop new electrolyte materials with higher ionic conductivity at temperatures in the range 500–700 ◦ C, for intermediate temperature solid oxide fuel cells (IT-SOFCs). Amongst the potential materials suitable for such application are apatite-type rare earth based oxides, such as R-doped lanthanum oxides of general formula La9.33 (RO4 )6 O2 with R = Ge, Si, which exhibit high ionic conductivity and low activation energy at moderate temperatures, when compared to the yttria-stabilized zirconia (YSZ) electrolyte commonly used [3,4]. These materials are currently produced by solid-state reaction methods [5]. Several problems in processing such apatite materials include the high temperature required, poor control of
∗ Corresponding author. Tel.: +351 239790745. E-mail address: [email protected] (R. Serra). 0925-8388/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.jallcom.2011.11.091
morphology, grain growth [6,7] and germanium volatilization (which is significant above 1250 ◦ C) [8,9]. Mechanical alloying (MA) [10] is a potential alternative processing method to obtain the apatite like lanthanum oxide [3,11]. During the MA process, heavy deformation is introduced into the particles, increasing the defect density, which accelerates the diffusion process by decreasing the diffusion distances. Therefore, sintering is achieved at lower temperatures than the ones required by conventional routes. In previous works, Si- and Ge-doped lanthanum oxide materials with apatite-type phase were produced by MA and subsequent annealing [12,13]. A dependence of apatite-type phase formation on the MA parameters was shown, the higher the MA rotation speed and/or processing time, the lower the annealing temperature required for its formation. In the present work, the influence of the milling atmosphere as well as of the particle size distribution of the raw materials on the structural evolution of powders during MA has been studied. The thermal behavior of the mechanically alloyed mixtures was studied by means of differential scanning calorimetry (DSC). 2. Experimental Powders of crystalline La2 O3 (99.9% purity), SiO2 (quartz with 99.4% purity) and GeO2 (99.9% purity), were used as starting materials. Mixtures with 7La2 O3 :3SiO2 :6GeO2 molar ratios were used in order to obtain the La9.33 Si2 Ge4 O26 phase. The blends were dry milled in a Fritsch planetary ball mill using a 250 ml hardened steel vial and 15 balls with 20 mm diameter of the same material. About 24.5 g of the powder blend was loaded into the vial as to get a ball-to-powder weight ratio of 20:1. Some of the experiments were carried out under protective atmosphere (Ar at 2 bar), the remaining being performed under atmospheric air. The powders
R. Serra et al. / Journal of Alloys and Compounds 536S (2012) S480–S484 Table 1 Characteristic values of the particle size distribution of the raw materials in the as-received condition and after pre-milling. d (m) La2 O3
As-received
Table 2 Lattice parameters of the apatite phase obtained for different milling conditions.
Pre-milled
d10 d50 d90
2.0 9.4 49.7
0.7 2.4 11.6
SiO2
d10 d50 d90
2.8 26.4 118.3
1.0 4.8 15.1
GeO2
d10 d50 d90
2.5 37.6 72.8
0.6 2.0 4.2
were mechanically alloyed at rotational speeds of vial of 150, 250 and 350 rpm for 15–35 h either in the as-received condition or after milling the starting materials separately at 350 rpm for 10 min (hereafter referred to as “pre-milled condition”). With this procedure, it was expected to start from more homogenous powder blends to promote the apatite single phase formation (by eliminating large particles and agglomerates) as well as to achieve MA mixtures with smaller particle size population that in principle can be easily sintered. Both raw materials and as-milled mixtures were characterized by means of laser scattering (particle size distribution), X-ray diffraction (XRD) with Co K␣ radiation and scanning electron microscopy (SEM). DSC was performed up to 1200 ◦ C under static air. Different heating rates, in the range 10–40 ◦ C min−1 , were used in order to determine the thermal parameters corresponding to the formation of apatite phase using the Kissinger method [14].
3. Results and discussion Fig. 1 shows scanning electron micrographs of the raw materials, as-received. It can be seen that they have different particle size distributions and morphologies. The GeO2 particles are round whereas the ones of La2 O3 and SiO2 show some flat surfaces (cleavage) typical of brittle materials. Table 1 shows the characteristic values of the particle size distribution of the oxide particles as-received and after pre-milling; dx (x = 10, 50, 90) means that the x volume fraction of the measured particles is at or below the diameter d. A significant decrease of d10 , d50 and d90 values of the starting materials was observed after premilling, as a result of the powders brittleness and deagglomeration, which may enhance sintering behavior. Fig. 2(a) and (b) show the particle size distributions (differential and cumulative curves) of the MA mixtures synthesized at 350 rpm in argon with and without pre-milling. Both MA mixtures are characterized by multimodal distributions with d50 values around 6 and 7 m for mixtures with and without pre-milling, respectively. The mixtures processed at 250 and 350 rpm were milled using the minimum time for lanthanum germanosilicate formation, as previously reported [12,13]. Fig. 3(a) shows the XRD patterns of the mechanically alloyed powder mixtures obtained in this work. As anticipated, formation of the apatite phase was obtained after 15 and 25 h at 350 and 250 rpm, respectively, while the mixture milled at 150 rpm for 35 h only shows peaks indexed to the starting materials.
S481
a (Å) c (Å) c/a
350 rpm (as-received) Milled in air
350 rpm (pre-milled) Milled in air
350 rpm (pre-milled) Milled in argon
9.92 7.25 0.731
9.96 7.25 0.728
9.90 7.27 0.734
The effect of pre-milling of the starting powders and of milling atmosphere was studied on samples prepared at 350 rpm for 15 h. No significant differences on apatite-type structure were observed by X-ray diffraction after MA (Fig. 3(b)). The apatite phase was formed after 15 h of milling independently of the milling atmosphere (Ar or air) and the initial state of the raw materials (asreceived or pre-milled). The crystallite size of the apatite phase was determined by the Scherrer equation. All the samples are nanocrystalline with very similar crystallite sizes (in the range of 12–18 nm). The hexagonal lattice parameters of the La9.33 Si2 Ge4 O26 apatite phase were determined after MA from the XRD patterns shown in Fig. 3. The values obtained for a and c are summarized in Table 2. The a parameter in all samples is higher than the value reported in the ˚ As the a value of ICCD card of the La9.33 Si3 Ge3 O26 phase (a = 9.86 A). apatite phase increases with Ge content, this result confirms that more than half of the positions in the Si/Ge–O tetrahedral structures of the unit cell are occupied by Ge atoms, in agreement with the SiO2 /GeO2 molar ratio of 1/2 used in the starting mixtures. Within the experimental error, the c parameter values of the mixtures after MA are similar to the one reported for the La9.33 Si3 Ge3 O26 phase ˚ This result may be due to either a weak dependence of (c = 7.27 A). the c parameter on the Ge/Si atomic ratio in the apatite phase, as compared to the a parameter or to a distortion of the unit cell with Ge content increase. Thermal behavior of the MA mixtures prepared at 150 rpm under different conditions was evaluated by DSC analysis. Fig. 4(a) shows the curve obtained at 30 ◦ C min−1 as a typical example. Fig. 4(b) shows the influence of the heating rate on the temperature of the apatite formation peak. No peaks were observed during cooling and therefore this part of the curves is not shown. The DSC curves show two endothermic peaks in the ranges 360–400 and 520–560 ◦ C (Fig. 4(a)), depending on the heating rate, related to the decomposition process of the La(OH)3 phase into LaOOH and subsequent dehydration with formation of La2 O3 . Typically, decomposition of the hydroxide and hydroxycarbonate occurs at 280 ◦ C and dehydration of LaOOH occurs at 420 ◦ C whereas decomposition of La2 O2 CO3 occurs at 600 ◦ C [15]. It is likely that the difference in recorded temperatures can be attributed to differences in moisture and CO2 contents. The presence of La(OH)3 was not confirmed by XRD analysis of the 150 rpm MA mixtures (probably its content is below the equipment detection limit or amorphous nature). However, its presence in La2 O3 powder is almost unavoidable taking into account the high reactivity of La2 O3 in the presence of moisture and carbon
Fig. 1. Powders used in this work (as-received): (a) La2 O3 ; (b) SiO2 ; (c) GeO2 .
S482
R. Serra et al. / Journal of Alloys and Compounds 536S (2012) S480–S484
Fig. 2. Particle size distributions of the MA mixtures after milling at 350 rpm for 15 h; comparison between as-received and pre-milled starting powders.
Fig. 3. XRD patterns of the mechanically alloyed mixtures: (a) influence of milling rotation speed and time; (b) influence of pre-milling and protective atmosphere (350 rpm).
Fig. 4. DSC curves of the 150 rpm milled sample: (a) 30 ◦ C min−1 ; (b) apatite formation peaks at different heating rates.
R. Serra et al. / Journal of Alloys and Compounds 536S (2012) S480–S484
S483
Fig. 6. Kissinger plot of the apatite peaks obtained from a set of heating rates.
Fig. 5. XRD patterns of the mechanically alloyed mixtures at 150 rpm for 35 h after DSC runs up to 900 and 1200 ◦ C.
dioxide. Some authors [15,16] claim that an amorphous phase of formula La2 (OH)4 CO3 ·nH2 O can be formed during the hydration of the La2 O3 that would explain La(OH)3 absence in the La2 O3 starting material. Tao and Irvine [17] used differential thermal analysis (DTA) to study the thermal behavior of the amorphous La10 Si6 O27 and La9.33 Si6 O26 phases obtained by sol–gel process. They observed, for each composition, just one exothermic low intensity peak with Tp peak temperature close to 959 ◦ C (in the case of La10 (SiO4 )6 O3 ) and 874 ◦ C (for La9.33 Si6 O26 phase) corresponding to the crystallization or formation peak of the lanthanum silicate phase. Tzvetkov and Minkova [18] have reported DTA and TG results obtained from powder mixtures of lanthanum oxide and amorphous silica synthesized by mechanical alloying. They observed that an amorphous phase was formed after MA as well as the existence of two exothermic peaks in the DTA traces, one in the range 855–870 ◦ C and another in the range 935–940 ◦ C. Both peaks were ascribed to apatite-type lanthanum silicate formation, the first from the amorphous precursor phase and the second from unreacted La2 O3 and SiO2 milled powders. In the present work the first exothermic DSC peak occurs in the temperature range 810–825 ◦ C, depending on the heating rate, and is ascribed to the formation of apatite. In a previous work on the annealing of MA’ed powder mixtures with and without germanium [12] it was shown that Ge is responsible for lowering the temperature of formation of the apatite phase. XRD analysis was performed after DSC runs in order to ascribe the exothermic peak to the formation of apatite (Fig. 5). The XRD pattern obtained for the DSC run up to 900 ◦ C showed the coexistence of the GeO2 , La(OH)3 and the apatite phase (Fig. 5). The presence of La(OH)3 rather than the initial La2 O3 phase is the result of its hydration, which occurred after DSC. Accordingly in addition to apatite, other phases with different chemical compositions are formed upon heating after MA. At higher temperatures, a second exothermic peak at 1110 ◦ C is observed (Fig. 4(a)), which is associated to the formation of La2 GeO5 (ICDD card 073-9203), as can be seen in Fig. 5. Solid state reactions occurring by nucleation and growth can be described by the Johnson–Mehl–Avrami (JMA) model [19]. According to this model, the fraction of transformed material after a hold time at a given temperature is expressed by the equation: x(t) = 1 − exp(−k · t n )
(1)
where x(t) is the volume fraction of the initial material transformed at time t, n is the Avrami exponent which is related to the nucleation rate and the growth morphology, and k is the temperaturedependent reaction rate constant. The Kissinger method is based on the JMA model and uses the highest rate of a transformation at maximum peak (Tp ) for interpreting the DSC results. It requires data recorded at different heating rates so that different values for Tp are obtained and can be interpreted by the following equation:
ln
˛ Tp2
=−
Ea R · Tp
(2)
where Tp is the temperature at maximum peak, ˛ = ∂T/∂t is the heating rate and Ea the activation energy necessary for the reaction. Ea is the slope of the straight line obtained by plotting ln(˛/Tp2 ) vs 1/Tp . The Tp and the enthalpy values of the apatite phase formation are shown in Table 3. As expected for a diffusion controlled reaction, the higher the heating rate the higher the temperature at maximum peak at which the apatite phase is formed. Inversely, the enthalpy decreased with increasing of heating rate (less time to complete the reactions and therefore lower energy release). Based on the Tp values, the Kissinger plot was drawn (Fig. 6) and the Ea was calculated from the slope of the straight line. A value of 65 kJ mol−1 was obtained. The Avrami exponent (n) is usually determined by isothermal DSC/DTA methods. However, Zhaosheng et al. [20] developed a non-isothermal method of multi-scanning heating rate for the determination of the Avrami exponent for an amorphous alloy. In this method, applicable to any phase transformation occurring by a diffusion process, n is given by: n=−
ln[− ln(1 − x1)] − ln[− ln(1 − x2)] ln ˛1 − ln ˛2
(3)
where ˛1 and ˛2 are the heating rates and x1 and x2 are the phase transformation fractions at the same temperature. In the present study, the temperatures of 812, 816 and 822 ◦ C (all in the temperature range of the apatite formation) were considered for 10–20, 20–30 and 30–40 ◦ C min−1 pairs of heating rates, respectively. The values of the Avrami exponent obtained using Eq. (3) are summarized in Table 4. Depending on the heating rates used, n values in the range 1.3–1.6 were obtained. This suggests that phase transformation occurs through diffusion under conditions were negligible nucleation takes place at the initial stages of transformation (i.e. site saturation) [21].
S484
R. Serra et al. / Journal of Alloys and Compounds 536S (2012) S480–S484
Table 3 Tp and enthalpy values of the DSC peak of the apatite phase formed under different heating rates.
◦
Tp ( C) Enthalpy
V = 10 ◦ C min−1
V = 20 ◦ C min−1
V = 30 ◦ C min−1
V = 40 ◦ C min−1
808.9 −21.13 J g−1 (−43.48 kJ mol−1 )
815.3 −22.20 J g−1 (−45.68 kJ mol−1 )
819.3 −23.61 J g−1 (−48.58 kJ mol−1 )
822.9 −23.41 J g−1 (−48.17 kJ mol−1 )
Table 4 Avrami exponent n obtained for three pairs of heating rates used.
References
Heating rates (◦ C min−1 )
ln ˛
x
ln x
10 20
2.302585 2.995732
0.8073 0.4262
−0.21406 −0.85285
1.6
20 30
2.995732 3.401197
0.6208 0.4334
−0.47675 −0.83609
1.3
30 40
3.401197 3.688879
0.6803 0.5178
−0.38522 −0.65817
1.6
n
4. Conclusions Mechanical alloying of La2 O3 , GeO2 and SiO2 powders induced the formation of the La9.33 Si2 Ge4 O26 apatite phase at room temperature at rotation speeds higher than 150 rpm. Neither pre-milling of the starting powders nor the composition of the milling atmosphere had any influence on the structure of the as-milled powders at 350 rpm for 15 h. No reaction between the La2 O3 , GeO2 and SiO2 powders has occurred at 150 rpm for up to 35 h. In this case, the apatite-type lanthanum germanosilicate is formed upon heating in the range 809–823 ◦ C with enthalpies between −43.5 and −48.6 kJ mol−1 . Applying the Kissinger equation to the reaction kinetics, an activation energy Ea of 65 kJ mol−1 was obtained. The mean Avrami exponent n calculated was 1.5 which suggests a diffusion controlled phase transformation with site saturation. Acknowledgment Financial support provided by FCT through project PTDC/EMEPME/102837/2008 is gratefully acknowledged as well as the research fellowship granted to C. Alves.
[1] E.J. Abram, D.C. Sinclair, A.R. West, J. Mater. Chem. 11 (2001) 1978–1979. [2] L. Leon-Reina, E.R. Losilla, M. Martínez-Lara, S. Bruque, M.A.G. Aranda, J. Mater. Chem. 14 (2004) 1142–1149. [3] Y. Higuchi, M. Sugawara, K. Onishi, M. Sakamoto, S. Nakayama, Ceram. Int. 36 (2010) 955–959. [4] S. Nakayama, M. Sakamoto, J. Mater. Sci. Lett. 20 (2001) 1627–1629. [5] H. Yao, J. Wang, D. Hu, J. Li, X. Lu, Z. Li, Solid State Ionics 181 (2010) 41–47. [6] G. Tzvetkov, N. Minkova, Mater. Lett. 39 (1999) 354–358. [7] E. Rodrıíguez-Reyna, A.F. Fuentes, M. Maczka, J. Hanuza, K. Boulahya, U. Amador, J. Solid State Chem. 179 (2006) 522–531. [8] E. Rodríguez-Reyna, A.F. Fuentes, M. Maczka, J. Hanuza, K. Boulahya, U. Amador, Solid State Sci. 8 (2006) 168–177. [9] S.P. Jiang, L. Zhang, H.Q. He, R.K. Yap, Y. Xiang, J. Power Sources 189 (2009) 972–981. [10] C. Suryanarayana, Prog. Mater. Sci. 46 (2001) 1–184. [11] E.J. Abram, C.A. Kirk, D.C. Sinclair, A.R. West, Solid State Ionics 176 (2005) 1941–1947. [12] M.M. Vieira, J.C. Oliveira, A. Cavaleiro, B. Trindade, Mater. Sci. Forum 587–588 (2008) 128–132. [13] M.M. Vieira, J.C. Oliveira, A. Cavaleiro, B. Trindade, Rev. Adv. Mater. Sci. 18 (2008) 344–348. [14] H.E. Kissinger, Anal. Chem. 29 (1957) 1702–1706. [15] P.J. Panteix, I. Julien, D. Bernache-Assollant, P. Abélard, Mater. Chem. Phys. 95 (2006) 313–320. [16] S. Bernal, F.J. Botana, R. García, J.M. Rodríguez-Izquierdo, Thermochim. Acta 66 (1993) 139–145. [17] S. Tao, J.T.S. Irvine, Mater. Res. Bull. 36 (2001) 1245–1258. [18] G. Tzvetkov, N. Minkova, J. Mater. Sci. 35 (2000) 2435–2441. [19] M. Avrami, J. Chem. Phys. 8 (1940) 212–224. [20] T. Zhaosheng, Z. Bangwei, Z. Hen Tan, J. Mater. Sci. 27 (1992) 4021–4023. [21] R. Cahn, P. Haasen, Physical Metallurgy, vol. 2, 4th ed., North-Holland, Amsterdam, 1996, ISBN 0444898751.