- Email: [email protected]
Effect of the precursor nature and preparation mode on the coarsening of La2Zr2O7 compounds
Available online at www.sciencedirect.com
CERAMICS INTERNATIONAL
Ceramics International 42 (2016) 1197–1209 www.elsevier.com/locate/ceramint
Effect of the precursor nature and preparation mode on the coarsening of La2Zr2O7 compounds William Duartea,b, Michel Vardellea, Sylvie Rossignola,n a
Science des Procédés Céramiques et de Traitements de Surface (SPCTS), Limoges, France b Centre National d’Etudes Spatiales (CNES), Paris, France
Received 8 June 2015; received in revised form 1 September 2015; accepted 9 September 2015 Available online 16 September 2015
Abstract This study addresses the thermal behavior of La2Zr2O7 powders prepared via two different synthetic routes and the effect of the zirconium counter-cation on their properties, as carried out by dilatometric experiments. Citrate-based compounds exhibit shrinkage, regardless of the precursors used. Preliminary thermal treatment permits the reduction in shrinkage. Nitrate based compounds calcinated at 600 1C show a higher shrinkage ( 9.4%) than at 1200 1C ( 5.2%) because of the decomposition of the trapped precursors, as revealed by DTA–TGA–MS. Coprecipitated samples present different behaviors as a function of the counter-cation nature. Chloride based compounds leads to shrinkage and phase mixtures, whereas expansion is observed for pure pyrochlore compounds that are synthesized with nitrate. The thermal expansion coefficients of the different compounds are similar to the literature values (9.5–10.10 6 K 1). This study allowed for a better understanding of the densification of the La2Zr2O7 material and its adjusted properties, as a function of the precursor nature and the preparation mode, for application in thermal barrier coatings. & 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Keywords: La2Zr2O7; Coarsening; TBC; Preparation mode; Precursor
1. Introduction Improving the properties of thermal barriers, which are essential elements in the thermal protection of aircraft and rockets engine parts, is the subject of intensive research. This is intended to improve the efficiency of the engine cycle, which can result in reduction fuel consumption and greenhouse gas emissions, by increasing the operating temperature. The development of new alloys, particularly monocrystalline nickel-based superalloys [1,2], initially resulted in an increase in the operating temperature. The use of a ceramic layer (“topcoat”), which was elaborated by plasma spraying or electron-beam physical vapor deposition (EB-PVD), allowed
n Correspondence to: Science des Procédés Céramiques et de Traitements de Surface (SPCTS), 12, rue Atlantis, 87068 Limoges, France. Tel.: þ 33 5 87 50 25 64. E-mail address: [email protected] (S. Rossignol).
http://dx.doi.org/10.1016/j.ceramint.2015.09.051 0272-8842/& 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
the working temperatures to be above the melting point of the superalloys because of the low thermal conductivity [3,4]. Currently, partially stabilized zirconia (Y2O3–ZrO2; YSZ) is widely used because of its low thermal conductivity (2.1 W m 1 K 1) and high thermal expansion coefficient (10.7.10 6 K 1) to accommodate any stresses in the metal substrate (16.10 6 K 1) [4] during thermal cycling. However, its structural instability above 1200 1C (transformation from tetragonal phase to the monoclinic and tetragonal phases) induces a volume expansion and generates compressive stresses in the deposit. Compressive stresses cause cracks to appear, which results in severe damage to the thermal barrier coating [5–7]. Moreover, the sintering resistance of the material is relatively low, which causes a significant densification that is responsible for the increase in thermomechanical stresses, which in turn are detrimental to the coating properties. Research was conducted on the development of new materials to replace YSZ beyond 1200 1C. Doping zirconia
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
The resulting powders were ground in a mortar and calcinated at different temperatures (600, 800, 1000 or 1200 1C) in air for 2 h. For simplification, the following nomenclature was adopted for the calcinated compounds: Zr precursor_Synthesis method_Calcination temperature, and for dilatometric samples: Zr precursor_Synthesis method_TMA Calcination temperature . For example, a powder synthesized using the citrate method (VC) from zirconyl nitrate (ZrN) and lanthanum nitrate (LaN), and calcinated at 600 1C was denoted ZrN_VC_600, and a compound prepared by coprecipitation (CP) from zirconium oxychloride (ZrCl), calcinated at 600 °C, and analyzed by dilatometry was denoted ZrCl_CP_TMA 600 (Table 1). 2.2. Characterization techniques Dilatometric analysis was performed on a vertical Setsys Evolution TMA dilatometer from Setaram. The samples were analyzed with a thermal cycle characterized by a maximum temperature of 1400 1C with heating and cooling rates of 10 1C/min under air. The dimension variation of the samples was measured using a mechanical displacement transducer in alumina, and the applied force was 50 N. Calcinated powders were prepared under cylindrical shapes (Ø ¼ 6 mm et LE 5 mm) by powder pressing. Platinum sheets were placed at each side of the samples to avoid any chemical reaction at high temperature between the lanthanum zirconate sample and the alumina components of the apparatus. Experiments were carried out on compounds that had been synthesized using the citrate route, ZrN_VC, calcinated at 600, 800, 1000 and Table 1 Precursors and solvent nomenclature. Name
Formula
Code
Zirconium oxynitrate Zirconium oxychloride Zirconium acetate hydroxide Lanthanum nitrate Water
ZrO(NO3)2, 6H2O ZrOCl2, 8H2O Zr(OH)3CH3CO2 La(NO3)3, 6H2O H2O
ZrN ZrCl ZrAc LaN H
1200
ο
by rare earth metals, such as lanthanum, was the subject of significant research because it enables the synthesis of crystallizing compounds in the pyrochlore structure (La2Zr2O7). This structure is characterized by a structural stability up to its melting point (42200 1C), an improved sintering resistance and a lower thermal conductivity than YSZ [8–11]. The degradation of the thermal barriers coatings (TBCs) during thermal cycling is complex because of different constituent materials are used (substrate/bond coat/topcoat). The first mode of degradation that is widely studied is the delamination of the topcoat by the growth of an oxide layer of aluminum, also called the thermally grown oxide (TGO), at the interface between the metallic bond coat (generally MCr AlY, M¼ Ni, Co) and the ceramic topcoat. The native Al2O3 layer protects the substrate against corrosion at high temperatures from oxygen diffusion through the ceramic layer. However, its growth induces compressive stresses in the coating that are responsible for the formation and propagation of cracks [12–14]. The second mode of degradation is due to the high temperature densification of the ceramic coating. This results in volume shrinkage and an increase in Young modulus of the material. The rise in these phenomena increase the stresses that occur in the topcoat, where the formation and propagation of the cracks responsible for the component damage occur [15–18]. A previous study [19] was carried out on the synthesis of La2Zr2O7 compounds using different preparation techniques and zirconium precursors. The results showed that the nature of the zirconium counter-cation impacts whether a single phase structure or a mixture of phases of pyrochlore is formed. Larger particles were observed for the phase mixtures compared with the single phase compounds at the same calcination temperature. As such, it would be interesting to study the thermal behavior of the different compounds by dilatometry because few studies have been conducted on the densification process. Indeed, only the final volume shrinkage and relative densities have been determined by the Archimedes method for La2Zr2O7 synthesized by hydrothermal [20,21], and coprecipitation methods [9,22] have been reported. This study focuses on the influence of the nature of the zirconium counter-cation on the La2Zr2O7 densification synthesized via the citrate route and by coprecipitation. The results of dilatometry experiments are correlated with the structural and microstructural analysis of the elaborated compounds.
Temperature ( C)
1198
2. Experimental procedure 2.1. Compounds synthesis The lanthanum zirconate compounds were synthesized using lanthanum nitrate (La(NO3)3, 6H2O) and three zirconium precursors characterized by different counter-cations: nitrate (ZrO(NO3)2, 6H2O), chloride (ZrOCl2, 8H2O) and acetate (Zr(OH)3CO2CH3). Lanthanum and zirconium precursors were dissolved in stoichiometric amounts in water. The two synthetic modes used were the citrate route and the coprecipitation, as described previously [19].
1000
800
600 0
20
40
60
80
100
120
140
160
180
200
Thermal treatment time (min)
Fig. 1. Thermal treatment time (□: calcination dwell time and ■: TMA time from calcination temperature to 1400 1C) for ZrN_VC calcinated at different temperatures.
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
(a)
CO (/50)
2,0
98
Weight (%)
600°C
1,0
94 92
Heat flow (W/g)
1,5 96
0,5
Relative intensity (a.u.)
100
88 200
400
600
800
1000
0,0 1200
25
200
Temperature (°C) (a)
800
2,0
1000
1200
CO (/50)
1,0
94 92
Heat flow (W/g)
1,5 96 Weight (%)
600
Temperature (°C)
98
800°C
400
0,5
R e la t iv e in t e n s it y ( a .u .)
100
NO
COOH
90 (b)
25
1199
NO COOH
90 (b)
88 25
200
400
600
800
1000
0,0 1200
25
200
Temperature (°C) (a)
1,0
94 92
Heat flow (W/g)
1,5 96 Weight (%)
800
1000
0,5
NO CO (/50)
90
COOH
(b)
88 25
200
400
600
800
1000
0,0 1200
25
200
(a)
800
92 0,5
R elative in ten sity (a.u .)
Weight (%)
1,0
94
Heat flow (W/g)
1,5 96
90
600
1000
NO CO (/50)
(b)
COOH
88 25
1200
2,0
98
1200°C
400
Temperature (°C)
Temperature (°C)
100
1200
2,0
98
1000°C
600
Temperature (°C)
R elative intensity (a.u .)
100
400
200
400
600
800
Temperature (°C)
1000
0,0 1200
25
200
400
600
800
1000
1200
Temperature (°C)
Fig. 2. (A) Thermal analysis curves ((a) weight loss and (b) heat flow) as a function of temperature and (B) relative intensity m/z¼ 18(H2O), 30(NO), 44(CO2), 45 (COOH) and 46(NO2) for ZrN_VC compounds calcinated at (1, 1′) 600 (ZrN_VC_600), (2, 2′) 800 (ZrN_VC_800), (3, 3′) 1000 (ZrN_VC_1000) and (4, 4′) 1200 1C (ZrN_VC_1200) (some intensities were divided/50).
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
1200 1C, ZrCl_VC, ZrAc_VC, and two compounds elaborated by coprecipitation, ZrN_CP and ZrCl_CP, all calcinated at 600 and 1000 1C. After the dilatometric experiments, the compounds that were calcinated at different temperatures were used in the experiments at different heat treatment times, as shown in Fig. 1. The heat treatment times included 2 h of calcination at the desired temperature (600, 800, 1000 or 1200 1C) and the time required for the temperature to increase to 1400 1C during the dilatometric experiments. With a heating rate of 10 1C/min for dilatometric analysis, this time varied from 20 to 80 min. Differential thermal analysis (DTA) and thermogravimetric analysis (TGA) coupled with mass spectrometry (MS) were performed on an SDT Q600 apparatus from TA Instruments coupled with a heated capillary column and a mass spectrometer from Hiden. Thermal analysis of the ZrN_VC powders was performed up to 1200 1C at a heating rate of 20 1C/min to 550 1C and then at a rate of 5 1C/min under air in platinum crucibles. Signal masses, m/z, of 18 (H2O), 44 (CO2), 45 (COOH) and 46 (NO2) were recorded. Thermal analyzes were carried out on the ZrN_CP and ZrCl_CP samples up to 1400 1C at a heating rate of 10 1C/min under air in platinum crucibles. Powder X-ray diffraction analysis was performed on a D8 Advance instrument from Bruker with a Bragg–Brentano geometry using copper Kα radiation (λ¼ 1,5406 Å).
∑ ni d 3i
(b)
0.0
25
200
400
600
800
1000
1200
(1’)
(a)
1.5
98
0.5
90
(2 )
1.0
96 0.5 94
(b)
0.0
92
90
-0.5 1400
25
200
400
Temperature (°C)
(2)
(a)
1,5
100
96 0,5 94 (b)
0,0
92 90 25
200
400
600
800
1000
Temperature (°C)
800
1000
1200
-0.5 1400
1200
-0,5 1400
(2’)
(a)
1,5
98
1,0
Weight (%)
Weight (%)
98
1000°C
600
Temperature (°C)
Heat flow (W/g)
100
Heat flow (W/g)
100
1.0
92
(1 )
∑ ni d 4i
1.5
96
94
∑ ni d2i
Weight (%)
Weight (%)
Γv=
∑ ni d 3i
(1)
(a)
98
600°C
Γs=
Heat flow (W/g)
100
Diffraction patterns were collected from 101 to 601, with a step size of 0.0121 and a dwell time of 0.86 s. Diffraction peaks were indexed using JCPDS files and the DiffracPlusEVA software (Bruker). Raman spectra were collected with a T64000 spectrometer from Horiba-Jobin Yvan using a Ar þ /Kr þ laser with an excitation wavelength of 514 nm and operating at a power of 30 mW. Scattered light was collected in backscattering mode using a long working distance objective ( 50) with a triple diffraction grating (1800 lines/mm). The spectral range was 50–900 cm 1. The acquisition time was 10 s. The microstructure of the samples was observed using a JSM 7400F field emission scanning electron microscope from Jeol. The particle size was measured using the ImageJ software. From these measurements, the values of surface size Γs (Eq. (1)) and volume size Γv (Eq. (2)) [23,24] were determined using the n values of the diameter d.
1,0
96 0,5 94 (b)
0,0
92 90 25
200
400
600
800
1000
1200
Heat flow (W/g)
1200
-0,5 1400
Temperature (°C)
Fig. 3. Thermal analysis curves ((a) weight loss and (b) heat flow) as a function of temperature for compounds (A) ZrN_CP and (B) ZrCl_CP calcinated at (1) 600 1C (ZrN_CP_600, ZrCl_CP_600) and (2) 1000 1C (ZrN_CP_1000, ZrCl_CP_1000).
2
2
0
0
-2
-2
-4
800°C
∇
∇
-6
L/L (%)
600°C
L/L (%)
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
1201
-4 -6
-8
-8
-10
-10 -12
-12 0
200
400
600
800
1000
1200
1400
0
200
400
2
2
0
0
-2
-2
-4
1200°C
∇
∇
-6
800
1000
1200
1400
1200
1400
-4 -6
-8
-8
-10
-10 -12
-12 0
200
400
600
800
1000
1200
0
1400
200
Temperature (°C)
2
2
0
0
-2
-2
-4
1000°C
600
800
1000
-4 -6
∇
∇
-6
400
Temperature (°C)
L/L (%)
L/L (%)
600
Temperature (°C)
L/L (%)
1000°C
L/L (%)
Temperature (°C)
-8
-8
-10
-10
600°C
-12 0
200
400
600
800
1000
1200
1000°C 600°C
-12 1400
Temperature (°C)
0
200
400
600
800
1000
1200
1400
Temperature (°C)
Fig. 4. Dilatometric curves of the compounds at various temperatures (A) ZrN_VC_TMA (a) 600, (b) 800, (c) 1000 and (d) 12001C and (B) (a) ZrCl_VC_TMA 600 (─) and 10001C (- -) and (b) ZrAc_VC_TMA 600 (─) and 10001C (- -).
3. Results and discussion 3.1. Effect of thermal treatment of synthesized compounds studied by thermal analyzes 3.1.1. Detailed study of a compound prepared by citrate route (ZrN_VC) Compounds elaborated by citrate route with zirconium nitrate precursor (ZrN_VC) and calcinated at various temperatures (600, 800, 1000 and 1200 1C) were analyzed by TG– DTA–MS to identify the chemical reaction occurring during thermal treatment of the powders and the effect of the calcination temperature. Fig. 2 presents the thermal analysis curves (Fig. 2(A)) and the relative intensity curves (Fig. 2(B)) of the collected gas from the decomposition of ZrN_VC
compounds that were used in the dilatometric study as a function of temperature for a m/z¼ 44 (CO2), 45 (COOH) and 46 (NO2) under air. Below 1000 1C, intense peaks appear within the temperature range. No chemical species are detected for compounds calcinated at temperatures higher than 1000 1C. The thermal analysis curves of the compounds calcinated at 600 1C (ZrN_VC_600) (Fig. 2 (A.1)) show two small weight losses at 66 and 300 1C that are specific to the desorption of water and surface rearrangement, which were not observed for the other compounds. Two important weight losses at 604 and 836 1C, which are associated with two endothermic peaks, are observed. The peak at 604 1C is attributed to the thermal decomposition of the residual precursors due to the presence of intense peaks, with a m/z¼ 44, 45 and 46, which are associated with carbon dioxide (CO2), a carboxyl group (COOH) and
1202
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
nitrogen dioxide (NO2), respectively (Fig. 2(B.1′)). The peak at 836 1C may be due to the thermal decomposition of the residual precursors because intense peaks with a m/z¼ 44, 45 and 46 are observed, but they may also be due to the initiation of the compounds crystallization into the fluorite structure [25,26]. The weight loss at 1051 1C is associated with a weak peak that has a m/z¼ 44 and is involved in the decomposition of carbon dioxide due to the formation of lanthanum carbonate [27]. The ZrN_VC_800 compound presents a weight loss at 840 1C that is associated with an exothermic peak (Fig. 2 (A.2)). Again, the collected gases have peaks at m/z¼ 44, 45 and 46 (Fig. 2(B.2′)) and are intense, which reveals that the residual precursors trapped within the compound are decomposed. The compounds calcinated at higher temperatures, ZrN_VC_1000 and ZrN_VC_1200, as shown in Fig. 2(A.3, A.4) and (B.3′, B.4′), do not show a weight loss, and the intensities of the collected gases are negligible. These results highlight that a small part of the precursors are trapped inside the compound and decompose below 1000 1C. Thanks to these data two calcination temperature were selected (600 and 1000 1C) to the study of compounds elaborated by coprecipitation method with two different zirconium precursors.
4
4
2
2
0
0
L/L (%)
600°C
L/L (%)
3.1.2. Compounds synthesized by coprecipitation Coprecipitated compounds were also investigated by TG– DTA to confirm the thermal decomposition of the trapped
precursors in the 600–900 1C range as previously observed in the citrate based powders. Fig. 3 presents the thermal analysis curves as a function of temperature for ZrN_CP and ZrCl_CP calcinated at 600 and 1000 1C. The heat flow curve of ZrN_CP_600 (Fig. 3(A.1.b)) shows five exothermic peaks at 310, 650, 890, 1095 and 1210 1C. The first three peaks are associated with important weight losses (Fig. 3(A.1.a)) and can be attributed to the decomposition of the precursors and the beginning of the fluorite crystallization of the compound. The peak at 1095 1C could be associated to the degradation of lanthanum carbonate [27]. The last peak at 1210 1C is likely associated with the complete crystallization into the pyrochlore structure. For the ZrN_CP_1000 compound, two weak exothermic peaks are present at 949 and 1124 1C (Fig. 3 (A.2.b)). The peak at 949 1C is not associated with a weight loss, so it is less likely due to the complete fluorite crystallization of the sample. In contrast, the peak at 1124 1C is associated with a slight weight loss. Its origin may be the decomposition of lanthanum carbonate [27]. Similar behavior is observed for the chloride-based compound ZrCl_CP. Five exothermic peaks at 352, 698, 870, 1020 and 1240 1C can also be observed for the ZrCl_CP_600 sample (Fig. 3(B.1′.b)). Peaks at a temperature below 900 1C are characteristic of the decomposition reactions of the precursors. Conversely to the previous compound, the peak at 1020 1C also correlates to the weight loss due to the
∇
∇
-2
-4
-2
-4
-6
-6
0
200
400
600
800
1000
1200
1400
0
200
400
600
800
4
4
2
2
0
0
-2
∇
∇
1000
1200
1400
1000
1200
1400
Temperature (°C)
L/L (%)
1000°C
L/L (%)
Temperature (°C)
-4
-2
-4
-6
-6
0
200
400
600
800
1000
Temperature (°C)
1200
1400
0
200
400
600
800
Temperature (°C)
Fig. 5. Dilatometric curves of the compounds (A) ZrN_CP and (B) ZrCl_CP calcinated at (a) 6001C (ZrN_CP_TMA 600, ZrCl_CP_TMA 600) and (b) 1000 1C (ZrN_CP_TMA 1000, ZrCl_CP_TMA 1000).
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209 4 2
L/L (%)
0
∇
decomposition of the carbonate lanthanum. Finally, the peak at 1240 1C, which is not associated with a weight loss, is due to the completion of the pyrochlore crystallization of the compound. Thermal analysis curves of the compound calcinated at 1000 1C (ZrCl_CP_1000) (Fig. 3(B.2′)) shows a very weak weight loss above 1100 1C that is associated with an exothermic peak at 1135 1C, which could be due to the decomposition of the carbonate lanthanum species. As previously observed with samples that were prepared using the citrate route, precursors are trapped inside the coprecipitated compounds and are decomposed at a temperature below 1000 1C.
1203
R 0.160
0.165
0.170
0.175
0.180 (Å)
-2 1200
-4
1000
-6
800
-8
600
-10 -12
3.2. Dilatometric study
Fig. 6. Final shrinkage values as a function of the counter-cation radius for different compounds (■: citrate route, ▲: coprecipitation) calcinated at various temperature.
3.2.1. Effect of the precursor nature on citrate route compounds Dilatometric experiments were carried out on the ZrN_VC compounds to study the effect that the calcination temperature had on the La2Zr2O7 densification. Fig. 4(A) presents the dilatometric curves as a function of temperature for the compounds that were synthesized using the citrate route (ZrN_VC) from an aqueous solution and calcinated at different temperatures. Regardless of the calcination temperature, the dilatometric curves presented a slight expansion at lower temperatures, then significant shrinkage occurred until the cooling was complete. For a calcination temperature of 600 1C (ZrN_VC_TMA 600) (Fig. 4(A.a)), the curve shows a slight expansion between 30 and 540 1C due to the desorption of adsorbed water and surface rearrangements [28]. Then, different shrinkage events are observed. The first shrinkage (540–820 1C) is attributed to the thermal decomposition of residual precursors that were trapped within the material after calcination but were not eliminated as previously shown by thermal analysis coupled with mass spectrometry of the powder. Between 820 and 870 1C, a second shrinkage appears and is corroborated by the transformation of weakly crystallized structure into fluorite. The appearance of the fluorite structure in the La2Zr2O7 compound occurs at a temperature lower than 900 1C [29,30]. The fluorite to pyrochlore transformation would occur at a temperature lower than 1000 1C. Various studies [20,21,26] have indicated that it is difficult to determine the polymorphic transformation temperature for powders calcinated between 900 and 1100 1C by XRD because of the very weak intensity of the (331) and (511) characteristic diffraction peaks of the pyrochlore structure. Beyond 1080 1C, shrinkages are specific to the material densification. During the cooling, the shrinkage is linear and reaches a final value of -9.4%. The coefficient of thermal expansion (CTE) of ZrN_VC_TMA 600 between 1300 and 500 1C is 9.8.10 6 K 1. The ZrN_VC_TMA 800 compound calcinated at 800 1C (Fig. 4(A.b)) presents an expansion followed by a first shrinkage at a temperature of 870 1C (þ 0.3%) due to surface rearrangements [28]. A shrinkage characteristic of the densification is observed above 1000 1C. The final shrinkage value is 6.8%, and the CTE between 1300 and 500 1C is
9.5.10 6 K 1. For a calcination temperature of 1000 1C (ZrN_VC_TMA 1000) (Fig. 4(A.c)), an expansion (þ 0.8%) is observed until 1020 1C, and a significant shrinkage then occurs due to the densification phenomena. The final shrinkage is 6.7%, and the CTE is constant at 9.5.10 6 K 1. The ZrN_VC_TMA 1200 compound calcinated at the maximum temperature of 1200 1C (Fig. 4(A.d)) shows an expansion (þ 1.2%) at 1130 1C, followed by the final shrinkage. The shrinkage value is the lowest ( 5.2%), but the CTE remains constant at 9.5.10 6 K 1. The increase in the calcination temperature leads to a decreased of the final shrinkage but also removes the residual precursors trapped within the compound that are responsible for the significant shrinkage. Fig. 4(B) shows the dilatometric curves as a function of temperature for compounds that were synthesized using the citrate route in an aqueous medium with different zirconium counter-cations of chloride (ZrOCl2, 8H2O) or acetate (Zr(OH)3CH3CO2). The powders are then calcinated at two different temperatures, 600 and 1000 1C. The compounds that were calcinated at 600 1C, ZrCl_VC_TMA 600 (Fig. 4(B.a)) and ZrAc_VC_TMA 600 (Fig. 4(B.b)), behave similarly to the previously studied compound (ZrN_VC_TMA 600) with significant shrinkages at temperatures below 870 1C due to the thermal decomposition of the residual precursors that are still trapped within the compounds after calcination at a low temperature. However, ZrCl_VC_TMA 600 shows a higher shrinkage at 870 1C ( 6.9%) than ZrAc_VC_TMA 600 ( 3.1%). This difference in the shrinkage could be explained by the different thermal decompositions of the residual precursors. The thermal decomposition of the acetate- and nitrate-based precursors can occur through the formation of hydrocarbons, nitrogen carbonates and oxides [31–33]. However, the chloride-based precursor can only decompose into the chloride ion or dichloride [34]. Nevertheless, both compounds have a similar final shrinkage and CTE, which values of of 10.4% and 9.8.10 6 K 1 for ZrCl_VC_TMA 600, and 10.4% and 9.8.10 6 K 1 for ZrAc_VC_TMA 600. The importance of calcination at high temperature to avoid shrinkage due to the decomposition of residual precursors is confirmed by ZrCl_VC_TMA 1000 (Fig. 4(B.a)) and ZrAc_VC_TMA 1000 (Fig. 4(B.b)) with an expansion of þ 0.80% at 1000 1C and þ 0.7% at 870 1C, followed by a
1204
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
Fig. 7. (A) XRD patterns and (B) Raman spectra of different compounds synthesized by citrate route from ZrN-LaN (a) as calcinated and (b) after dilatometric test (JCPDS file| P (La2Zr2O7 pyrochlore): 00-017-0450).
shrinkage due to densification. However, the final shrinkage and CTE are different. ZrCl_VC_TMA 1000 shows the lowest final shrinkage ( 4.2%) compared with ZrAc_VC_TMA 1000 ( 8.2%) and ZrN_VC_TMA 1000 ( 6.7%). In contrast, ZrAc_VC_TMA 1000 presents the highest CTE of 11.3.10 6 K 1 against 10.3.10 6 K 1 for ZrCl_VC_TMA 1000 and 9.5.10 6 K 1 for ZrN_VC_TMA 1000. For the citrate route, the nature of the zirconium countercation has a minimal effect on the general behavior observed in the dilatometric analysis, but the mineral counter-cations ( NO−3 , Cl ) seem to promote a lower final shrinkage than does the organic counter-cation (CH3COO ). 3.2.2. Effect of the synthesized method: coprecipitated compounds Fig. 5 presents the dilatometric curves as a function of temperature for compounds that were elaborated by coprecipitation in an aqueous medium with nitrate- and chloride-based precursors and were then calcinated at 600 and 1000 1C. The compound ZrN_CP_TMA 600 (Fig. 5(A.a)) shows a slight expansion (þ 0.3%) between 30 and 408 1C due to the
desorption of adsorbed water and surface rearrangements [28]. Two phenomena are observed: a shrinkage (408– 890 1C), followed by an expansion that intensifies from 1150 1C. The first phenomenon can be attributed to the thermal decomposition of the precursors and the fluorite transformation, as previously described. The second phenomenon may be due to the appearance of a mixture of crystalline phases. During the cooling step, the shrinkage is linear ( 0.36%), and the CTE is close to that of the citrate-based compounds (9.5.10 6 K 1). The ZrN_CP_TMA 1000 (Fig. 5(A.b)) presents an expansion during the temperature rise that sharply increases from 1150 1C and attains a maximum value of þ 3.2% at 1400 1C. The expansion linearly decreases during the cooling (þ 1.9%), and the CTE is similar (9.9.10 6 K 1). The expansion of the ZrN_CP_TMA 1000 compound was confirmed by experiments carried out on other samples. Conversely, the ZrCl_CP shows similar behavior to those of the compounds elaborated from citrate route. The ZrCl_CP_TMA 600 samples (Fig. 5(B.a)) present a first shrinkage ( 2.2%) at 880 1C, with a final value of 5.0%. The CTE remains close to 9.5.10 6 K 1. When the compound
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
nsP
1000°C (1400°C-24h) TMA 1000°C
1000°C (1400°C-24h)
nsP
P
TMA 1000°C
P
20
30
40
50
60
10
20
nsP
P
P
P
200
30
40
50
60
P
P
50
nsP
1000°C
1000°C 10
1205
400
P
P
P
600
800
900
-1
Raman shift (cm )
50
200
400
PP
600
800
900
-1
Raman shift (cm )
Fig. 8. (A) XRD patterns as calcinated and after dilatometric test and (B) Raman spectra of calcinated compounds synthesized by coprecipitation (a) ZrN_CP_1000 and (b) ZrCl_CP_1000 (JCPDS files | P (La2Zr2O7 pyrochlore): 00-017-0450; nsP (La1.88Zr2.09O7): 04-018-5483, ● (ZrO2 tetragonal) :00-050-1089).
is calcinated at a temperature of 1000 1C (ZrCl_CP_TMA 1000) (Fig. 5(B.b)), an expansion is observed ( þ 0.9%) until 1060 1C, followed by shrinkage due to densification. The final shrinkage value is 2.9%, and the CTE is slightly higher (10.2.10 6 K 1). Unlike the citrate route, the effect of the zirconium counter-cation is important in coprecipitation. Indeed, the use of nitrate leads to completely different thermal behavior with a definitive material expansion that is not observed for the coprecipitated chloride compound or citrate based powders.
3.2.3. Comparison of different preparations Fig. 6 shows the final shrinkage for the different compounds that were prepared using the citrate route and by coprecipitation at various temperatures as a function of the atomic radius of the zirconium counter-cation (acetate, chloride and nitrate at, respectively, 0.162, 0.172 and 0.178 Å). These representations, considering all of the shrinkages, exacerbate the effect of the
preparation mode relative to the counter-cation effect. Citratebased compounds exhibit similar behavior for each calcination temperature. Low-temperature compounds present shrinkage values that are approximately 9% higher, regardless of the precursor used. This may occur because of the higher porosity that is caused by the combustion synthesis and the residual precursors revealed by mass spectrometry. Increasing the calcination temperature results in a reduced shrinkage due to a higher densification rate and avoiding residual precursors. Coprecipitated compounds exhibit low shrinkage due to their higher density and lower porosity. Expansion is even observed for the ZrN_CP_TMA 1000 sample. The countercation appears to impact the shrinkage for coprecipitated compounds. Indeed, for 600 and 1000 1C, the ZrCl_CP sample presents a higher final shrinkage than the ZrN_CP_TMA 1000; thus, the shrinkage decreases with the increase in the countercation radius. The difference in the thermal behavior of the two coprecipitated compounds is not well understood. However, the presence of different structures as a function of the
1206
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
precursor nature may affect the expansion or shrinkage of the compound. More experiments should be performed to understand this phenomenon more completely. 400
Particles size (nm)
300
200
100
0
140 min
160 min
ZrN_VC_1200 ZrN_VC_1000
180 min ZrN_VC_800
200 min ZrN_VC_600
Fig. 9. Particles size values (Γs: □ and Γv: ■) determined by SEM as a function of the heat treatment time (calcinationsþ TMA until 1400 1C) of ZrN_VC compounds.
3.3. Effect of various thermal treatments on structural and microstructural data 3.3.1. Crystalline phases investigations The XRD patterns and Raman spectra of the different ZrN_VC compounds as calcinated and after dilatometric analysis are shown in Fig. 7. The ZrN_VC_600 and ZrN_VC_800 (Fig. 7(A.a)) samples show characteristic XRD patterns of a weakly crystallized structure. At 1000 1C (ZrN_VC_1000), the fluorite structure is identified, but at 1200 1C (ZrN_VC_1200), (331) and (511) diffraction peaks appear that are characteristic of the pyrochlore structure. These results are confirmed by the Raman spectra (Fig. 7(B.a)), which show that at a local level, the pyrochlore structure is the unique phase in the ZrN_VC_1200 compounds due to the presence of characteristic Raman bands (299 (F2g), 395 (Eg) and 520 cm 1 (F2g)). After dilatometric tests at 1400 1C, both ZrN_VC_TMA 600 and ZrN_VC_TMA 1000 samples present a single pyrochlore structure by XRD and Raman spectroscopy (Fig. 7(A.b and B.b)). Thus, the presence of unreacted precursors trapped within the compound calcinated at 600 1C
100 nm
100 nm
100 nm
100 nm
Fig. 10. SEM photographs after dilatometric experiments of compounds synthesized by (A) citrate route (a) ZrN_VC_TMA 1200 and (b) ZrN_VC_TMA 600, and (B) coprecipitation (a) ZrN_CP_TMA 1000 and (b) ZrCl_CP_TMA 1000.
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
do not affect the structure of the sample after a hightemperature treatment. Fig. 8 presents the XRD patterns and Raman spectra of the compounds that were prepared using coprecipitation, ZrN_CP_1000 and ZrCl_CP_1000, which were calcinated and heat treated for an additional 24 h at 1400 1C after the dilatometric analysis. The ZrN_CP_1000 and ZrCl_CP_1000 (Fig. 8(A.a and A.b)) samples show XRD patterns that are characteristic of the fluorite crystallized structure. After dilatometric tests at 1400 1C, the ZrN_CP_TMA 1000 compounds showed a single pyrochlore structure that was due to the appearance of the characteristic (331) and (511) diffraction peaks. These results are corroborated by the Raman spectrum (Fig. 8(B.a)), where the characteristic Raman bands associated with the pyrochlore structure (299 (F2g), 395 (Eg) and 520 cm 1 (F2g)) are present. It must be noted that the unpressed powder, after being heat treated at 1400 1C for 24 h (ZrN_CP_1000(1400 1C-24 h)), is also a single structure that is a non-stoichiometric pyrochlore (La1.88Zr2.09O7) (Fig. 8(A.a)). From previous work [19], unpressed ZrN_CP_1200 compounds without dilatometric tests show a phase mixture that is composed of pyrochlore and fluorite zirconia. Thus, it can be stated that the shaping and/or heat treatment at 1400 1C influences the crystallization of the compound. However, the ZrCl_CP_TMA 1000 compound presents a mixture of crystalline phases: pyrochlore and tetragonal zirconia. Indeed, the XRD pattern (Fig. 8(A.b)) is characteristic of the pyrochlore structure, but a peak at approximately 301 indicates that the tetragonal zirconia phase is also present. The Raman spectrum of the compound (Fig. 8(B.b)) confirmed the phase mixture at the local order because the characteristic Raman bands of the pyrochlore structure (299 (F2g), 395 (E2) and 520 cm 1 (F2g)) are observed, as are the tetragonal zirconia bands (146, 456 and 646 cm 1). A phase mixture is also observed by XRD for the compounds that were heat treated at 1400 1C for 24 h ((ZrCl_CP_1000(1400 1C-24 h)), with the presence of non-stoichiometric pyrochlore (La1.88Zr2.09O7) and tetragonal zirconia (Fig. 8 (A.b)). The brief thermal treatment of the powders at 1400 1C that occurred in the dilatometric study led to the full crystallization of the compounds from fluorite to pyrochlore. However, pressing the powder appears to favor a phase mixture with the formation of tetragonal zirconia in ZrCl_CP_TMA 1000, whereas the single pyrochlore structure is only observed for the ZrN_CP_TMA 1000 sample. Moreover, the heat treatment time also affects unpressed powders. Indeed, after 24 h at 1400 1C, the fluorite phase observed in calcinated powders (ZrN_CP_1000 and ZrCl_CP_1000) was crystallized into non-stoichiometric pyrochlore (La1.88Zr2.09O7), which is in accordance with the ZrO2–La2O3 binary diagram [35]. As observed for the previous compound (ZrN_CP), the unpressed ZrCl_CP_1200 compound exhibited a different crystallization because a single pyrochlore structure was found [19]. This result confirms that the pressure shaping and/or heat treatment at 1400 1C modified the crystallization of the coprecipitated compound. 3.3.2. Microstructure by SEM investigations Fig. 9 presents the particle size values (surface size (Γs) and volume sizes (Γs)) of the different ZrN_VC calcinated compounds as a function of the heat treatment time. For each
1207
sample, the volume size (Γs) is slightly higher than the surface size (Γs), but they remain relatively close. Comparing all of the samples, the particle size increases with the heat treatment time. High-temperature calcinated compounds (ZrN_VC_1200) are characterized by the smallest particle size. This can be explained by the fact that the particle growth grain size occurs when the temperature during the dilatometric test exceeds the calcination temperature, in this case 1200 1C. Thus, the ZrN_VC_1200 particles only have a very short time to grow, 20 min (Fig. 1). For the ZrN_VC_600 compound, the particles have more time to grow (80 min) because a lower calcination temperature was used (600 1C) (Fig. 1). The difference in particle size is highlighted clearly in the SEM images of the two citrate based compounds presented in Fig. 10(A). The ZrN_VC_TMA 600 compounds have larger particles (Γs¼ 284 nm and Γv¼ 323 nm) than do the ZrN_VC_TMA 1200 compounds (Γs ¼ 170 nm and Γv ¼ 187 nm). Calcination at 1200 1C favors the presence of small particles during the dilatometric process, which are of interest for decreasing the thermal conductivity due to the presence of more grain boundaries. However, high-temperature calcination also leads to material densification and has the opposite effect to the small particle size, which increases the thermal conductivity. Thus, the choice of calcination temperature must be a compromise between thermal conductivity and densification. The SEM images of the samples after the dilatometric tests at 1400 1C (ZrN_CP_TMA 1000 and ZrCl_CP_TMA 1000) show similar microstructures that consist of a mixture of large and small particles (Fig. 10(B)), though they are different crystalline structures. Moreover, the particle size measurements do not show any difference between the compounds. Indeed, the surface size (Γs) and volume size (Γv) of the ZrN_CP_TMA 1000 sample of 263 and 286 nm, respectively, are close to those of the ZrCl_CP_TMA 1000 sample of 241 and 265 nm, respectively. 4. Conclusions Ceramic materials, which are used as topcoats in thermal barrier coatings, are subjected to densification during the operating step of an engine. This phenomenon is partly responsible for the severe degradation of TBC, but little information is available. The dilatometric study of La2Zr2O7 compounds that were synthesized by the citrate route or by coprecipitation revealed that the preparation mode has a significant impact on the shrinkage of the material. The compounds that were elaborated using the citrate route with different zirconium precursors showed significant shrinkage for all of the samples analyzed. A preliminary thermal treatment results in a reduced shrinkage. For instance, from 9.4% (ZrN_VC_TMA 600) to 5.2% (ZrN_VC_TMA 1200), The DTA-TGA-MS results highlighted that the shrinkage during low-temperature calcination (o 1000 1C) is due to the decomposition of residual precursors that are trapped within the compounds. Therefore, preliminary thermal treatments are important for controlling the shrinkage. Moreover,
1208
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
the higher the calcination temperature is, the smaller the particle size of the citrate-based compound is after the dilatometric tests. The formation of small particles is interesting because of their potential thermal barrier applications due to the presence of more grain boundaries, which decreases the thermal conductivity. However, higher densification can also be achieved at a high temperature, which increases the thermal conductivity. Compounds elaborated by coprecipitation exhibited varying results compared with the citrate-based powders. Their thermal behaviors are dependent on the zirconium counter-cation, contrary to the citrate-based compounds. Chloride countercations favored shrinkage, whereas the nitrate sample with a calcination at 1000 1C resulted in an expansion of the dilatometric sample. Thus, the counter-cation radius affects the thermal behavior of the coprecipitated compounds. XRD and Raman spectroscopy revealed that the powder shaping by pressing combined with the high temperature (1400 1C) that was experienced during the dilatometric experiments modified the structure compared with the raw powder that was calcinated at 1200 1C. The counter-cation also influences the coprecipitated compounds, with a phase mixture of pyrochlore and tetragonal zirconia observed for the chloride counter-cation and a single pyrochlore structure observed for nitrate, as for the citrate-based compounds. The coefficient of thermal expansion for the various compounds prepared by either the citrate or coprecipitation method was approximately 10.10 6 K 1, which is consistent with the literature values. These results show that the thermal behavior of La2Zr2O7 powders can be modified by the preparation mode, the nature of the precursor, and the thermal treatment. To limit the shrinkage responsible for the degradation of the thermal barrier coating, coprecipitated compounds appear to be quite promising. Moreover it would be interesting to study the thermal behavior of synthesized powders after their treatment in a plasma jet. Acknowledgments The authors gratefully acknowledge the DGA (French Ministry of Defense) and CNES National R&T program for their financial support. References [1] T.M. Pollock, S. Tin, Nickel-based superalloys for advanced turbine engines: chemistry, microstructure, and properties, J. Propuls. Power 22 (2006) 361–374. [2] G.E. Fuchs, Solution heat treatment response of a third generation single crystal Ni-base superalloy, Mater. Sci. Eng. A 300 (2001) 52–60. [3] D.R. Clarke, M. Oechsner, N.P. Padture, Thermal barrier coatings for more efficient gas turbine engines, MRS Bull. 37 (2012) 891–898. [4] X.Q. Cao, R. Vassen, D. Stover, Ceramic materials for thermal barrier coatings, J. Eur. Ceram. Soc. 24 (2004) 1–10. [5] V. Lughi, D.R. Clarke, High temperature aging of YSZ coatings and subsequent transformation at low temperature, Surf. Coat. Technol. 200 (2005) 1287–1291.
[6] J. Moon, H. Choi, H. Kim, C. Lee, The effects of the heat treatment on the pahse transformation behavior of plasma sprayed stabilized ZrO2 coatings, Surf. Coat. Technol. 155 (2002) 1–10. [7] C.U. Hardwicke, Y.C. Lau, Advances in thermal spray coatings for gas turbines and energy generation : a review, J. Therm. Spray Technol. 22 (2013) 564–576. [8] D. Zhu, R.A. Miller, Development of advanced low conductivity thermal barrier coatings, Int. J. Ceram. Technol. 1 (2004) 86–94. [9] R. Vassen, X. Cao, F. Tietz, D. Basu, D. Stöver, Zirconates as new materials for thermal barrier coatings, J. Am. Ceram. Soc. 83 (2000) 2023–2028. [10] H. Kido, S. Komarneni, R. Roy, Preparation of La2Zr2O7 by sol–gel route, J. Am. Ceram. Soc. 74 (1991) 422–424. [11] J. Nair, P. Nair, E.B.M. Doesburg, J.G. Van Ommen, J.R.H. Ross, A. J. Burggraff, F. Mizukami, Preparation and characterization of lanthanum zirconate, J. Mater. Sci. 33 (1998) 4517–4523. [12] A. Rabiei, A.G. Evans, Failure mechanisms associated with the thermally grown oxide in plasma sprayed thermal barrier coating, Acta Mater. 48 (2000) 3963–3976. [13] K.W. Schlichting, N.P. Padture, E.H. Jordan, M. Gell, Failures modes in plasma sprayed thermal barrier coatings, Mater. Sci. Eng. A 342 (2003) 120–130. [14] A.G. Evans, D.R. Mumm, J.W. Hutchinson, G.H. Meier, F.S. Petit, Mechanisms controlling the durability of thermal barrier coatings, Prog. Mater. Sci. 46 (2001) 505–553. [15] C. Chen, H. Guo, S. Gong, X. Zhao, P. Xiao, Sintering of electron beam physical vapor deposited thermal barrier coatings under flame shock, Ceram. Int. 39 (2013) 5093–5102. [16] D. Zhu, R.A. Miller, Sintering and creep behavior of plasma sprayed zirconia and hafnia based thermal barrier coating, Surf. Coat. Technol. 108–109 (1998) 114–120. [17] A. Cipitria, I.O. Golosnoy, T.W. Clyne, A sintering model for plasma sprayed zirconia thermal barrier coatings. Part II: Coatings bonded to a rigid substrate, Acta Mater. 57 (2009) 993–1003. [18] G. Thurn, G.A. Schneider, F. Aldinger, High temperature deformation of plasma sprayed ZrO2 thermal barrier coatings, Mater. Sci. Eng. A 233 (1997) 176–182. [19] W. Duarte, A. Meguekam, M. Colas, M. Vardelle, S. Rossignol, Effect of the counter-cation nature and preparation method on the structure of La2Zr2O7, J. Mater. Sci. 50 (2015) 463–475. [20] C. Wang, Y. Wang, Y. Cheng, W. Huang, Z.S. Khan, X. Fan, Y. Wang, B. Zou, X. Cao, Preparation and thermophysical properties of nano-sized Ln2Zr2O7 (Ln¼ La, Nd, Sm and Gd) ceramics with pyrochlore structure, J. Mater. Sci. 47 (2012) 4392–4399. [21] Y. Wang, C. Li, C. Wang, A. Zhang, Y. Chen, Synthesis of La2O3–ZrO2–Y2O3 with various Y2O3 contents via hydrothermal method : a phase evolution and sintering-resistance properties study, Int. J. Appl. Ceram. Technol. 12 (2014) 65–70. [22] J. Wang, S. Bai, H. Zhang, C. Zhang, The structure, thermal expansion coefficient and sintering behavior of Nd3 þ doped La2Zr2O7 for thermal barrier coatings, J. Alloy. Compd. 476 (2009) 89–91. [23] R.D. Cadle, Particle Size Theory and Industrial Applications, Reinhold Publishing Corp, New York, 1965. [24] T. Allen, Particle Size Measurement, third ed., Chapman and Hall, London, 1981. [25] A. Zhang, M. Lu, G. Zhou, S. Wang, Y. Zhou, Combustion synthesis and photoluminescence of Eu3 þ , Dy3 þ doped La2Zr2O7 nanocrystals, J. Phys. Chem. Solids 67 (2006) 2430–2434. [26] D. Prusty, A. Pathak, A. Chintha, B. Mukherjee, A. Chowdhury, Structural investigations on the compositional anomalies in lanthanum zirconate system synthesized by coprecipitation method, J. Am. Ceram. Soc. 97 (2014) 718–724. [27] J.C. Grivel, Y. Zhao, X. Tang, P.G.P.A. Pallewatta, A. Watenphul, M. V. Zimmermann, Thermal decomposition of lanthanum(III) butyrate in argon atmosphere, Thermochim. Acta 566 (2013) 112–117. [28] H. Chen, Y. Gao, Y. Liu, H. Luo, Coprecipitation synthesis and thermal conductivity of La2Zr2O7, J. Alloy. Compd. 480 (2009) 843–848. [29] A. Zhang, M. Lu, G. Zhou, S. Wang, Y. Zhou, Combustion synthesis and photoluminescence of Eu3 þ , Dy3 þ doped La2Zr2O7 nanocrystals, J. Phys. Chem. Solids 67 (2006) 2430–2434.
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209 [30] D. Prusty, A. Pathak, A. Chintha, B. Mukherjee, A. Chowdhury, Structural investigations on the compositional anomalies in lanthanum zirconate system synthesized by coprecipitation method, J. Am. Ceram. Soc. 97 (2014) 718–724. [31] K. Tonsuaadu, A. Zalga, A. Beganskiene, A. Kareiva, Thermoanalytical study of the YSZ precurors prepared by aqueous sol-gel synthesis route, J. Therm. Anal. Calorim. 110 (2012) 77–83. [32] J.C. De Jesus, I. Gonzalez, A. Quevedo, T. Puerta, Thermal decomposition of nickel acetate tetrahydrate an integrated study by TGA, QMS and XPS techniques, J. Mol. Catal. A Chem. 228 (2005) 283–291.
1209
[33] G.A.M. Hussein, Rare earth metal oxides: formation, characterization and catalytic activity. Thermoanalytical and applied pyrolysis, J. Anal. Appl. Pyrolysis 37 (1996) 111–149. [34] T. Settu, Characterisation of MgO–ZrO2 precursor powders prepared by in-situ peptisation of coprecipitated oxalate gel, Ceram. Int. 26 (2000) 517–521. [35] A. Rouanet, Contribution à l’étude des systèmes zircone-oxydes des lanthanides au voisinage de la fusion, Rev. Int. Hautes Temp. Réfract. 8 (1971) 161–180.
CERAMICS INTERNATIONAL
Ceramics International 42 (2016) 1197–1209 www.elsevier.com/locate/ceramint
Effect of the precursor nature and preparation mode on the coarsening of La2Zr2O7 compounds William Duartea,b, Michel Vardellea, Sylvie Rossignola,n a
Science des Procédés Céramiques et de Traitements de Surface (SPCTS), Limoges, France b Centre National d’Etudes Spatiales (CNES), Paris, France
Received 8 June 2015; received in revised form 1 September 2015; accepted 9 September 2015 Available online 16 September 2015
Abstract This study addresses the thermal behavior of La2Zr2O7 powders prepared via two different synthetic routes and the effect of the zirconium counter-cation on their properties, as carried out by dilatometric experiments. Citrate-based compounds exhibit shrinkage, regardless of the precursors used. Preliminary thermal treatment permits the reduction in shrinkage. Nitrate based compounds calcinated at 600 1C show a higher shrinkage ( 9.4%) than at 1200 1C ( 5.2%) because of the decomposition of the trapped precursors, as revealed by DTA–TGA–MS. Coprecipitated samples present different behaviors as a function of the counter-cation nature. Chloride based compounds leads to shrinkage and phase mixtures, whereas expansion is observed for pure pyrochlore compounds that are synthesized with nitrate. The thermal expansion coefficients of the different compounds are similar to the literature values (9.5–10.10 6 K 1). This study allowed for a better understanding of the densification of the La2Zr2O7 material and its adjusted properties, as a function of the precursor nature and the preparation mode, for application in thermal barrier coatings. & 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
Keywords: La2Zr2O7; Coarsening; TBC; Preparation mode; Precursor
1. Introduction Improving the properties of thermal barriers, which are essential elements in the thermal protection of aircraft and rockets engine parts, is the subject of intensive research. This is intended to improve the efficiency of the engine cycle, which can result in reduction fuel consumption and greenhouse gas emissions, by increasing the operating temperature. The development of new alloys, particularly monocrystalline nickel-based superalloys [1,2], initially resulted in an increase in the operating temperature. The use of a ceramic layer (“topcoat”), which was elaborated by plasma spraying or electron-beam physical vapor deposition (EB-PVD), allowed
n Correspondence to: Science des Procédés Céramiques et de Traitements de Surface (SPCTS), 12, rue Atlantis, 87068 Limoges, France. Tel.: þ 33 5 87 50 25 64. E-mail address: [email protected] (S. Rossignol).
http://dx.doi.org/10.1016/j.ceramint.2015.09.051 0272-8842/& 2015 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
the working temperatures to be above the melting point of the superalloys because of the low thermal conductivity [3,4]. Currently, partially stabilized zirconia (Y2O3–ZrO2; YSZ) is widely used because of its low thermal conductivity (2.1 W m 1 K 1) and high thermal expansion coefficient (10.7.10 6 K 1) to accommodate any stresses in the metal substrate (16.10 6 K 1) [4] during thermal cycling. However, its structural instability above 1200 1C (transformation from tetragonal phase to the monoclinic and tetragonal phases) induces a volume expansion and generates compressive stresses in the deposit. Compressive stresses cause cracks to appear, which results in severe damage to the thermal barrier coating [5–7]. Moreover, the sintering resistance of the material is relatively low, which causes a significant densification that is responsible for the increase in thermomechanical stresses, which in turn are detrimental to the coating properties. Research was conducted on the development of new materials to replace YSZ beyond 1200 1C. Doping zirconia
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
The resulting powders were ground in a mortar and calcinated at different temperatures (600, 800, 1000 or 1200 1C) in air for 2 h. For simplification, the following nomenclature was adopted for the calcinated compounds: Zr precursor_Synthesis method_Calcination temperature, and for dilatometric samples: Zr precursor_Synthesis method_TMA Calcination temperature . For example, a powder synthesized using the citrate method (VC) from zirconyl nitrate (ZrN) and lanthanum nitrate (LaN), and calcinated at 600 1C was denoted ZrN_VC_600, and a compound prepared by coprecipitation (CP) from zirconium oxychloride (ZrCl), calcinated at 600 °C, and analyzed by dilatometry was denoted ZrCl_CP_TMA 600 (Table 1). 2.2. Characterization techniques Dilatometric analysis was performed on a vertical Setsys Evolution TMA dilatometer from Setaram. The samples were analyzed with a thermal cycle characterized by a maximum temperature of 1400 1C with heating and cooling rates of 10 1C/min under air. The dimension variation of the samples was measured using a mechanical displacement transducer in alumina, and the applied force was 50 N. Calcinated powders were prepared under cylindrical shapes (Ø ¼ 6 mm et LE 5 mm) by powder pressing. Platinum sheets were placed at each side of the samples to avoid any chemical reaction at high temperature between the lanthanum zirconate sample and the alumina components of the apparatus. Experiments were carried out on compounds that had been synthesized using the citrate route, ZrN_VC, calcinated at 600, 800, 1000 and Table 1 Precursors and solvent nomenclature. Name
Formula
Code
Zirconium oxynitrate Zirconium oxychloride Zirconium acetate hydroxide Lanthanum nitrate Water
ZrO(NO3)2, 6H2O ZrOCl2, 8H2O Zr(OH)3CH3CO2 La(NO3)3, 6H2O H2O
ZrN ZrCl ZrAc LaN H
1200
ο
by rare earth metals, such as lanthanum, was the subject of significant research because it enables the synthesis of crystallizing compounds in the pyrochlore structure (La2Zr2O7). This structure is characterized by a structural stability up to its melting point (42200 1C), an improved sintering resistance and a lower thermal conductivity than YSZ [8–11]. The degradation of the thermal barriers coatings (TBCs) during thermal cycling is complex because of different constituent materials are used (substrate/bond coat/topcoat). The first mode of degradation that is widely studied is the delamination of the topcoat by the growth of an oxide layer of aluminum, also called the thermally grown oxide (TGO), at the interface between the metallic bond coat (generally MCr AlY, M¼ Ni, Co) and the ceramic topcoat. The native Al2O3 layer protects the substrate against corrosion at high temperatures from oxygen diffusion through the ceramic layer. However, its growth induces compressive stresses in the coating that are responsible for the formation and propagation of cracks [12–14]. The second mode of degradation is due to the high temperature densification of the ceramic coating. This results in volume shrinkage and an increase in Young modulus of the material. The rise in these phenomena increase the stresses that occur in the topcoat, where the formation and propagation of the cracks responsible for the component damage occur [15–18]. A previous study [19] was carried out on the synthesis of La2Zr2O7 compounds using different preparation techniques and zirconium precursors. The results showed that the nature of the zirconium counter-cation impacts whether a single phase structure or a mixture of phases of pyrochlore is formed. Larger particles were observed for the phase mixtures compared with the single phase compounds at the same calcination temperature. As such, it would be interesting to study the thermal behavior of the different compounds by dilatometry because few studies have been conducted on the densification process. Indeed, only the final volume shrinkage and relative densities have been determined by the Archimedes method for La2Zr2O7 synthesized by hydrothermal [20,21], and coprecipitation methods [9,22] have been reported. This study focuses on the influence of the nature of the zirconium counter-cation on the La2Zr2O7 densification synthesized via the citrate route and by coprecipitation. The results of dilatometry experiments are correlated with the structural and microstructural analysis of the elaborated compounds.
Temperature ( C)
1198
2. Experimental procedure 2.1. Compounds synthesis The lanthanum zirconate compounds were synthesized using lanthanum nitrate (La(NO3)3, 6H2O) and three zirconium precursors characterized by different counter-cations: nitrate (ZrO(NO3)2, 6H2O), chloride (ZrOCl2, 8H2O) and acetate (Zr(OH)3CO2CH3). Lanthanum and zirconium precursors were dissolved in stoichiometric amounts in water. The two synthetic modes used were the citrate route and the coprecipitation, as described previously [19].
1000
800
600 0
20
40
60
80
100
120
140
160
180
200
Thermal treatment time (min)
Fig. 1. Thermal treatment time (□: calcination dwell time and ■: TMA time from calcination temperature to 1400 1C) for ZrN_VC calcinated at different temperatures.
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
(a)
CO (/50)
2,0
98
Weight (%)
600°C
1,0
94 92
Heat flow (W/g)
1,5 96
0,5
Relative intensity (a.u.)
100
88 200
400
600
800
1000
0,0 1200
25
200
Temperature (°C) (a)
800
2,0
1000
1200
CO (/50)
1,0
94 92
Heat flow (W/g)
1,5 96 Weight (%)
600
Temperature (°C)
98
800°C
400
0,5
R e la t iv e in t e n s it y ( a .u .)
100
NO
COOH
90 (b)
25
1199
NO COOH
90 (b)
88 25
200
400
600
800
1000
0,0 1200
25
200
Temperature (°C) (a)
1,0
94 92
Heat flow (W/g)
1,5 96 Weight (%)
800
1000
0,5
NO CO (/50)
90
COOH
(b)
88 25
200
400
600
800
1000
0,0 1200
25
200
(a)
800
92 0,5
R elative in ten sity (a.u .)
Weight (%)
1,0
94
Heat flow (W/g)
1,5 96
90
600
1000
NO CO (/50)
(b)
COOH
88 25
1200
2,0
98
1200°C
400
Temperature (°C)
Temperature (°C)
100
1200
2,0
98
1000°C
600
Temperature (°C)
R elative intensity (a.u .)
100
400
200
400
600
800
Temperature (°C)
1000
0,0 1200
25
200
400
600
800
1000
1200
Temperature (°C)
Fig. 2. (A) Thermal analysis curves ((a) weight loss and (b) heat flow) as a function of temperature and (B) relative intensity m/z¼ 18(H2O), 30(NO), 44(CO2), 45 (COOH) and 46(NO2) for ZrN_VC compounds calcinated at (1, 1′) 600 (ZrN_VC_600), (2, 2′) 800 (ZrN_VC_800), (3, 3′) 1000 (ZrN_VC_1000) and (4, 4′) 1200 1C (ZrN_VC_1200) (some intensities were divided/50).
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
1200 1C, ZrCl_VC, ZrAc_VC, and two compounds elaborated by coprecipitation, ZrN_CP and ZrCl_CP, all calcinated at 600 and 1000 1C. After the dilatometric experiments, the compounds that were calcinated at different temperatures were used in the experiments at different heat treatment times, as shown in Fig. 1. The heat treatment times included 2 h of calcination at the desired temperature (600, 800, 1000 or 1200 1C) and the time required for the temperature to increase to 1400 1C during the dilatometric experiments. With a heating rate of 10 1C/min for dilatometric analysis, this time varied from 20 to 80 min. Differential thermal analysis (DTA) and thermogravimetric analysis (TGA) coupled with mass spectrometry (MS) were performed on an SDT Q600 apparatus from TA Instruments coupled with a heated capillary column and a mass spectrometer from Hiden. Thermal analysis of the ZrN_VC powders was performed up to 1200 1C at a heating rate of 20 1C/min to 550 1C and then at a rate of 5 1C/min under air in platinum crucibles. Signal masses, m/z, of 18 (H2O), 44 (CO2), 45 (COOH) and 46 (NO2) were recorded. Thermal analyzes were carried out on the ZrN_CP and ZrCl_CP samples up to 1400 1C at a heating rate of 10 1C/min under air in platinum crucibles. Powder X-ray diffraction analysis was performed on a D8 Advance instrument from Bruker with a Bragg–Brentano geometry using copper Kα radiation (λ¼ 1,5406 Å).
∑ ni d 3i
(b)
0.0
25
200
400
600
800
1000
1200
(1’)
(a)
1.5
98
0.5
90
(2 )
1.0
96 0.5 94
(b)
0.0
92
90
-0.5 1400
25
200
400
Temperature (°C)
(2)
(a)
1,5
100
96 0,5 94 (b)
0,0
92 90 25
200
400
600
800
1000
Temperature (°C)
800
1000
1200
-0.5 1400
1200
-0,5 1400
(2’)
(a)
1,5
98
1,0
Weight (%)
Weight (%)
98
1000°C
600
Temperature (°C)
Heat flow (W/g)
100
Heat flow (W/g)
100
1.0
92
(1 )
∑ ni d 4i
1.5
96
94
∑ ni d2i
Weight (%)
Weight (%)
Γv=
∑ ni d 3i
(1)
(a)
98
600°C
Γs=
Heat flow (W/g)
100
Diffraction patterns were collected from 101 to 601, with a step size of 0.0121 and a dwell time of 0.86 s. Diffraction peaks were indexed using JCPDS files and the DiffracPlusEVA software (Bruker). Raman spectra were collected with a T64000 spectrometer from Horiba-Jobin Yvan using a Ar þ /Kr þ laser with an excitation wavelength of 514 nm and operating at a power of 30 mW. Scattered light was collected in backscattering mode using a long working distance objective ( 50) with a triple diffraction grating (1800 lines/mm). The spectral range was 50–900 cm 1. The acquisition time was 10 s. The microstructure of the samples was observed using a JSM 7400F field emission scanning electron microscope from Jeol. The particle size was measured using the ImageJ software. From these measurements, the values of surface size Γs (Eq. (1)) and volume size Γv (Eq. (2)) [23,24] were determined using the n values of the diameter d.
1,0
96 0,5 94 (b)
0,0
92 90 25
200
400
600
800
1000
1200
Heat flow (W/g)
1200
-0,5 1400
Temperature (°C)
Fig. 3. Thermal analysis curves ((a) weight loss and (b) heat flow) as a function of temperature for compounds (A) ZrN_CP and (B) ZrCl_CP calcinated at (1) 600 1C (ZrN_CP_600, ZrCl_CP_600) and (2) 1000 1C (ZrN_CP_1000, ZrCl_CP_1000).
2
2
0
0
-2
-2
-4
800°C
∇
∇
-6
L/L (%)
600°C
L/L (%)
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
1201
-4 -6
-8
-8
-10
-10 -12
-12 0
200
400
600
800
1000
1200
1400
0
200
400
2
2
0
0
-2
-2
-4
1200°C
∇
∇
-6
800
1000
1200
1400
1200
1400
-4 -6
-8
-8
-10
-10 -12
-12 0
200
400
600
800
1000
1200
0
1400
200
Temperature (°C)
2
2
0
0
-2
-2
-4
1000°C
600
800
1000
-4 -6
∇
∇
-6
400
Temperature (°C)
L/L (%)
L/L (%)
600
Temperature (°C)
L/L (%)
1000°C
L/L (%)
Temperature (°C)
-8
-8
-10
-10
600°C
-12 0
200
400
600
800
1000
1200
1000°C 600°C
-12 1400
Temperature (°C)
0
200
400
600
800
1000
1200
1400
Temperature (°C)
Fig. 4. Dilatometric curves of the compounds at various temperatures (A) ZrN_VC_TMA (a) 600, (b) 800, (c) 1000 and (d) 12001C and (B) (a) ZrCl_VC_TMA 600 (─) and 10001C (- -) and (b) ZrAc_VC_TMA 600 (─) and 10001C (- -).
3. Results and discussion 3.1. Effect of thermal treatment of synthesized compounds studied by thermal analyzes 3.1.1. Detailed study of a compound prepared by citrate route (ZrN_VC) Compounds elaborated by citrate route with zirconium nitrate precursor (ZrN_VC) and calcinated at various temperatures (600, 800, 1000 and 1200 1C) were analyzed by TG– DTA–MS to identify the chemical reaction occurring during thermal treatment of the powders and the effect of the calcination temperature. Fig. 2 presents the thermal analysis curves (Fig. 2(A)) and the relative intensity curves (Fig. 2(B)) of the collected gas from the decomposition of ZrN_VC
compounds that were used in the dilatometric study as a function of temperature for a m/z¼ 44 (CO2), 45 (COOH) and 46 (NO2) under air. Below 1000 1C, intense peaks appear within the temperature range. No chemical species are detected for compounds calcinated at temperatures higher than 1000 1C. The thermal analysis curves of the compounds calcinated at 600 1C (ZrN_VC_600) (Fig. 2 (A.1)) show two small weight losses at 66 and 300 1C that are specific to the desorption of water and surface rearrangement, which were not observed for the other compounds. Two important weight losses at 604 and 836 1C, which are associated with two endothermic peaks, are observed. The peak at 604 1C is attributed to the thermal decomposition of the residual precursors due to the presence of intense peaks, with a m/z¼ 44, 45 and 46, which are associated with carbon dioxide (CO2), a carboxyl group (COOH) and
1202
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
nitrogen dioxide (NO2), respectively (Fig. 2(B.1′)). The peak at 836 1C may be due to the thermal decomposition of the residual precursors because intense peaks with a m/z¼ 44, 45 and 46 are observed, but they may also be due to the initiation of the compounds crystallization into the fluorite structure [25,26]. The weight loss at 1051 1C is associated with a weak peak that has a m/z¼ 44 and is involved in the decomposition of carbon dioxide due to the formation of lanthanum carbonate [27]. The ZrN_VC_800 compound presents a weight loss at 840 1C that is associated with an exothermic peak (Fig. 2 (A.2)). Again, the collected gases have peaks at m/z¼ 44, 45 and 46 (Fig. 2(B.2′)) and are intense, which reveals that the residual precursors trapped within the compound are decomposed. The compounds calcinated at higher temperatures, ZrN_VC_1000 and ZrN_VC_1200, as shown in Fig. 2(A.3, A.4) and (B.3′, B.4′), do not show a weight loss, and the intensities of the collected gases are negligible. These results highlight that a small part of the precursors are trapped inside the compound and decompose below 1000 1C. Thanks to these data two calcination temperature were selected (600 and 1000 1C) to the study of compounds elaborated by coprecipitation method with two different zirconium precursors.
4
4
2
2
0
0
L/L (%)
600°C
L/L (%)
3.1.2. Compounds synthesized by coprecipitation Coprecipitated compounds were also investigated by TG– DTA to confirm the thermal decomposition of the trapped
precursors in the 600–900 1C range as previously observed in the citrate based powders. Fig. 3 presents the thermal analysis curves as a function of temperature for ZrN_CP and ZrCl_CP calcinated at 600 and 1000 1C. The heat flow curve of ZrN_CP_600 (Fig. 3(A.1.b)) shows five exothermic peaks at 310, 650, 890, 1095 and 1210 1C. The first three peaks are associated with important weight losses (Fig. 3(A.1.a)) and can be attributed to the decomposition of the precursors and the beginning of the fluorite crystallization of the compound. The peak at 1095 1C could be associated to the degradation of lanthanum carbonate [27]. The last peak at 1210 1C is likely associated with the complete crystallization into the pyrochlore structure. For the ZrN_CP_1000 compound, two weak exothermic peaks are present at 949 and 1124 1C (Fig. 3 (A.2.b)). The peak at 949 1C is not associated with a weight loss, so it is less likely due to the complete fluorite crystallization of the sample. In contrast, the peak at 1124 1C is associated with a slight weight loss. Its origin may be the decomposition of lanthanum carbonate [27]. Similar behavior is observed for the chloride-based compound ZrCl_CP. Five exothermic peaks at 352, 698, 870, 1020 and 1240 1C can also be observed for the ZrCl_CP_600 sample (Fig. 3(B.1′.b)). Peaks at a temperature below 900 1C are characteristic of the decomposition reactions of the precursors. Conversely to the previous compound, the peak at 1020 1C also correlates to the weight loss due to the
∇
∇
-2
-4
-2
-4
-6
-6
0
200
400
600
800
1000
1200
1400
0
200
400
600
800
4
4
2
2
0
0
-2
∇
∇
1000
1200
1400
1000
1200
1400
Temperature (°C)
L/L (%)
1000°C
L/L (%)
Temperature (°C)
-4
-2
-4
-6
-6
0
200
400
600
800
1000
Temperature (°C)
1200
1400
0
200
400
600
800
Temperature (°C)
Fig. 5. Dilatometric curves of the compounds (A) ZrN_CP and (B) ZrCl_CP calcinated at (a) 6001C (ZrN_CP_TMA 600, ZrCl_CP_TMA 600) and (b) 1000 1C (ZrN_CP_TMA 1000, ZrCl_CP_TMA 1000).
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209 4 2
L/L (%)
0
∇
decomposition of the carbonate lanthanum. Finally, the peak at 1240 1C, which is not associated with a weight loss, is due to the completion of the pyrochlore crystallization of the compound. Thermal analysis curves of the compound calcinated at 1000 1C (ZrCl_CP_1000) (Fig. 3(B.2′)) shows a very weak weight loss above 1100 1C that is associated with an exothermic peak at 1135 1C, which could be due to the decomposition of the carbonate lanthanum species. As previously observed with samples that were prepared using the citrate route, precursors are trapped inside the coprecipitated compounds and are decomposed at a temperature below 1000 1C.
1203
R 0.160
0.165
0.170
0.175
0.180 (Å)
-2 1200
-4
1000
-6
800
-8
600
-10 -12
3.2. Dilatometric study
Fig. 6. Final shrinkage values as a function of the counter-cation radius for different compounds (■: citrate route, ▲: coprecipitation) calcinated at various temperature.
3.2.1. Effect of the precursor nature on citrate route compounds Dilatometric experiments were carried out on the ZrN_VC compounds to study the effect that the calcination temperature had on the La2Zr2O7 densification. Fig. 4(A) presents the dilatometric curves as a function of temperature for the compounds that were synthesized using the citrate route (ZrN_VC) from an aqueous solution and calcinated at different temperatures. Regardless of the calcination temperature, the dilatometric curves presented a slight expansion at lower temperatures, then significant shrinkage occurred until the cooling was complete. For a calcination temperature of 600 1C (ZrN_VC_TMA 600) (Fig. 4(A.a)), the curve shows a slight expansion between 30 and 540 1C due to the desorption of adsorbed water and surface rearrangements [28]. Then, different shrinkage events are observed. The first shrinkage (540–820 1C) is attributed to the thermal decomposition of residual precursors that were trapped within the material after calcination but were not eliminated as previously shown by thermal analysis coupled with mass spectrometry of the powder. Between 820 and 870 1C, a second shrinkage appears and is corroborated by the transformation of weakly crystallized structure into fluorite. The appearance of the fluorite structure in the La2Zr2O7 compound occurs at a temperature lower than 900 1C [29,30]. The fluorite to pyrochlore transformation would occur at a temperature lower than 1000 1C. Various studies [20,21,26] have indicated that it is difficult to determine the polymorphic transformation temperature for powders calcinated between 900 and 1100 1C by XRD because of the very weak intensity of the (331) and (511) characteristic diffraction peaks of the pyrochlore structure. Beyond 1080 1C, shrinkages are specific to the material densification. During the cooling, the shrinkage is linear and reaches a final value of -9.4%. The coefficient of thermal expansion (CTE) of ZrN_VC_TMA 600 between 1300 and 500 1C is 9.8.10 6 K 1. The ZrN_VC_TMA 800 compound calcinated at 800 1C (Fig. 4(A.b)) presents an expansion followed by a first shrinkage at a temperature of 870 1C (þ 0.3%) due to surface rearrangements [28]. A shrinkage characteristic of the densification is observed above 1000 1C. The final shrinkage value is 6.8%, and the CTE between 1300 and 500 1C is
9.5.10 6 K 1. For a calcination temperature of 1000 1C (ZrN_VC_TMA 1000) (Fig. 4(A.c)), an expansion (þ 0.8%) is observed until 1020 1C, and a significant shrinkage then occurs due to the densification phenomena. The final shrinkage is 6.7%, and the CTE is constant at 9.5.10 6 K 1. The ZrN_VC_TMA 1200 compound calcinated at the maximum temperature of 1200 1C (Fig. 4(A.d)) shows an expansion (þ 1.2%) at 1130 1C, followed by the final shrinkage. The shrinkage value is the lowest ( 5.2%), but the CTE remains constant at 9.5.10 6 K 1. The increase in the calcination temperature leads to a decreased of the final shrinkage but also removes the residual precursors trapped within the compound that are responsible for the significant shrinkage. Fig. 4(B) shows the dilatometric curves as a function of temperature for compounds that were synthesized using the citrate route in an aqueous medium with different zirconium counter-cations of chloride (ZrOCl2, 8H2O) or acetate (Zr(OH)3CH3CO2). The powders are then calcinated at two different temperatures, 600 and 1000 1C. The compounds that were calcinated at 600 1C, ZrCl_VC_TMA 600 (Fig. 4(B.a)) and ZrAc_VC_TMA 600 (Fig. 4(B.b)), behave similarly to the previously studied compound (ZrN_VC_TMA 600) with significant shrinkages at temperatures below 870 1C due to the thermal decomposition of the residual precursors that are still trapped within the compounds after calcination at a low temperature. However, ZrCl_VC_TMA 600 shows a higher shrinkage at 870 1C ( 6.9%) than ZrAc_VC_TMA 600 ( 3.1%). This difference in the shrinkage could be explained by the different thermal decompositions of the residual precursors. The thermal decomposition of the acetate- and nitrate-based precursors can occur through the formation of hydrocarbons, nitrogen carbonates and oxides [31–33]. However, the chloride-based precursor can only decompose into the chloride ion or dichloride [34]. Nevertheless, both compounds have a similar final shrinkage and CTE, which values of of 10.4% and 9.8.10 6 K 1 for ZrCl_VC_TMA 600, and 10.4% and 9.8.10 6 K 1 for ZrAc_VC_TMA 600. The importance of calcination at high temperature to avoid shrinkage due to the decomposition of residual precursors is confirmed by ZrCl_VC_TMA 1000 (Fig. 4(B.a)) and ZrAc_VC_TMA 1000 (Fig. 4(B.b)) with an expansion of þ 0.80% at 1000 1C and þ 0.7% at 870 1C, followed by a
1204
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
Fig. 7. (A) XRD patterns and (B) Raman spectra of different compounds synthesized by citrate route from ZrN-LaN (a) as calcinated and (b) after dilatometric test (JCPDS file| P (La2Zr2O7 pyrochlore): 00-017-0450).
shrinkage due to densification. However, the final shrinkage and CTE are different. ZrCl_VC_TMA 1000 shows the lowest final shrinkage ( 4.2%) compared with ZrAc_VC_TMA 1000 ( 8.2%) and ZrN_VC_TMA 1000 ( 6.7%). In contrast, ZrAc_VC_TMA 1000 presents the highest CTE of 11.3.10 6 K 1 against 10.3.10 6 K 1 for ZrCl_VC_TMA 1000 and 9.5.10 6 K 1 for ZrN_VC_TMA 1000. For the citrate route, the nature of the zirconium countercation has a minimal effect on the general behavior observed in the dilatometric analysis, but the mineral counter-cations ( NO−3 , Cl ) seem to promote a lower final shrinkage than does the organic counter-cation (CH3COO ). 3.2.2. Effect of the synthesized method: coprecipitated compounds Fig. 5 presents the dilatometric curves as a function of temperature for compounds that were elaborated by coprecipitation in an aqueous medium with nitrate- and chloride-based precursors and were then calcinated at 600 and 1000 1C. The compound ZrN_CP_TMA 600 (Fig. 5(A.a)) shows a slight expansion (þ 0.3%) between 30 and 408 1C due to the
desorption of adsorbed water and surface rearrangements [28]. Two phenomena are observed: a shrinkage (408– 890 1C), followed by an expansion that intensifies from 1150 1C. The first phenomenon can be attributed to the thermal decomposition of the precursors and the fluorite transformation, as previously described. The second phenomenon may be due to the appearance of a mixture of crystalline phases. During the cooling step, the shrinkage is linear ( 0.36%), and the CTE is close to that of the citrate-based compounds (9.5.10 6 K 1). The ZrN_CP_TMA 1000 (Fig. 5(A.b)) presents an expansion during the temperature rise that sharply increases from 1150 1C and attains a maximum value of þ 3.2% at 1400 1C. The expansion linearly decreases during the cooling (þ 1.9%), and the CTE is similar (9.9.10 6 K 1). The expansion of the ZrN_CP_TMA 1000 compound was confirmed by experiments carried out on other samples. Conversely, the ZrCl_CP shows similar behavior to those of the compounds elaborated from citrate route. The ZrCl_CP_TMA 600 samples (Fig. 5(B.a)) present a first shrinkage ( 2.2%) at 880 1C, with a final value of 5.0%. The CTE remains close to 9.5.10 6 K 1. When the compound
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
nsP
1000°C (1400°C-24h) TMA 1000°C
1000°C (1400°C-24h)
nsP
P
TMA 1000°C
P
20
30
40
50
60
10
20
nsP
P
P
P
200
30
40
50
60
P
P
50
nsP
1000°C
1000°C 10
1205
400
P
P
P
600
800
900
-1
Raman shift (cm )
50
200
400
PP
600
800
900
-1
Raman shift (cm )
Fig. 8. (A) XRD patterns as calcinated and after dilatometric test and (B) Raman spectra of calcinated compounds synthesized by coprecipitation (a) ZrN_CP_1000 and (b) ZrCl_CP_1000 (JCPDS files | P (La2Zr2O7 pyrochlore): 00-017-0450; nsP (La1.88Zr2.09O7): 04-018-5483, ● (ZrO2 tetragonal) :00-050-1089).
is calcinated at a temperature of 1000 1C (ZrCl_CP_TMA 1000) (Fig. 5(B.b)), an expansion is observed ( þ 0.9%) until 1060 1C, followed by shrinkage due to densification. The final shrinkage value is 2.9%, and the CTE is slightly higher (10.2.10 6 K 1). Unlike the citrate route, the effect of the zirconium counter-cation is important in coprecipitation. Indeed, the use of nitrate leads to completely different thermal behavior with a definitive material expansion that is not observed for the coprecipitated chloride compound or citrate based powders.
3.2.3. Comparison of different preparations Fig. 6 shows the final shrinkage for the different compounds that were prepared using the citrate route and by coprecipitation at various temperatures as a function of the atomic radius of the zirconium counter-cation (acetate, chloride and nitrate at, respectively, 0.162, 0.172 and 0.178 Å). These representations, considering all of the shrinkages, exacerbate the effect of the
preparation mode relative to the counter-cation effect. Citratebased compounds exhibit similar behavior for each calcination temperature. Low-temperature compounds present shrinkage values that are approximately 9% higher, regardless of the precursor used. This may occur because of the higher porosity that is caused by the combustion synthesis and the residual precursors revealed by mass spectrometry. Increasing the calcination temperature results in a reduced shrinkage due to a higher densification rate and avoiding residual precursors. Coprecipitated compounds exhibit low shrinkage due to their higher density and lower porosity. Expansion is even observed for the ZrN_CP_TMA 1000 sample. The countercation appears to impact the shrinkage for coprecipitated compounds. Indeed, for 600 and 1000 1C, the ZrCl_CP sample presents a higher final shrinkage than the ZrN_CP_TMA 1000; thus, the shrinkage decreases with the increase in the countercation radius. The difference in the thermal behavior of the two coprecipitated compounds is not well understood. However, the presence of different structures as a function of the
1206
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
precursor nature may affect the expansion or shrinkage of the compound. More experiments should be performed to understand this phenomenon more completely. 400
Particles size (nm)
300
200
100
0
140 min
160 min
ZrN_VC_1200 ZrN_VC_1000
180 min ZrN_VC_800
200 min ZrN_VC_600
Fig. 9. Particles size values (Γs: □ and Γv: ■) determined by SEM as a function of the heat treatment time (calcinationsþ TMA until 1400 1C) of ZrN_VC compounds.
3.3. Effect of various thermal treatments on structural and microstructural data 3.3.1. Crystalline phases investigations The XRD patterns and Raman spectra of the different ZrN_VC compounds as calcinated and after dilatometric analysis are shown in Fig. 7. The ZrN_VC_600 and ZrN_VC_800 (Fig. 7(A.a)) samples show characteristic XRD patterns of a weakly crystallized structure. At 1000 1C (ZrN_VC_1000), the fluorite structure is identified, but at 1200 1C (ZrN_VC_1200), (331) and (511) diffraction peaks appear that are characteristic of the pyrochlore structure. These results are confirmed by the Raman spectra (Fig. 7(B.a)), which show that at a local level, the pyrochlore structure is the unique phase in the ZrN_VC_1200 compounds due to the presence of characteristic Raman bands (299 (F2g), 395 (Eg) and 520 cm 1 (F2g)). After dilatometric tests at 1400 1C, both ZrN_VC_TMA 600 and ZrN_VC_TMA 1000 samples present a single pyrochlore structure by XRD and Raman spectroscopy (Fig. 7(A.b and B.b)). Thus, the presence of unreacted precursors trapped within the compound calcinated at 600 1C
100 nm
100 nm
100 nm
100 nm
Fig. 10. SEM photographs after dilatometric experiments of compounds synthesized by (A) citrate route (a) ZrN_VC_TMA 1200 and (b) ZrN_VC_TMA 600, and (B) coprecipitation (a) ZrN_CP_TMA 1000 and (b) ZrCl_CP_TMA 1000.
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
do not affect the structure of the sample after a hightemperature treatment. Fig. 8 presents the XRD patterns and Raman spectra of the compounds that were prepared using coprecipitation, ZrN_CP_1000 and ZrCl_CP_1000, which were calcinated and heat treated for an additional 24 h at 1400 1C after the dilatometric analysis. The ZrN_CP_1000 and ZrCl_CP_1000 (Fig. 8(A.a and A.b)) samples show XRD patterns that are characteristic of the fluorite crystallized structure. After dilatometric tests at 1400 1C, the ZrN_CP_TMA 1000 compounds showed a single pyrochlore structure that was due to the appearance of the characteristic (331) and (511) diffraction peaks. These results are corroborated by the Raman spectrum (Fig. 8(B.a)), where the characteristic Raman bands associated with the pyrochlore structure (299 (F2g), 395 (Eg) and 520 cm 1 (F2g)) are present. It must be noted that the unpressed powder, after being heat treated at 1400 1C for 24 h (ZrN_CP_1000(1400 1C-24 h)), is also a single structure that is a non-stoichiometric pyrochlore (La1.88Zr2.09O7) (Fig. 8(A.a)). From previous work [19], unpressed ZrN_CP_1200 compounds without dilatometric tests show a phase mixture that is composed of pyrochlore and fluorite zirconia. Thus, it can be stated that the shaping and/or heat treatment at 1400 1C influences the crystallization of the compound. However, the ZrCl_CP_TMA 1000 compound presents a mixture of crystalline phases: pyrochlore and tetragonal zirconia. Indeed, the XRD pattern (Fig. 8(A.b)) is characteristic of the pyrochlore structure, but a peak at approximately 301 indicates that the tetragonal zirconia phase is also present. The Raman spectrum of the compound (Fig. 8(B.b)) confirmed the phase mixture at the local order because the characteristic Raman bands of the pyrochlore structure (299 (F2g), 395 (E2) and 520 cm 1 (F2g)) are observed, as are the tetragonal zirconia bands (146, 456 and 646 cm 1). A phase mixture is also observed by XRD for the compounds that were heat treated at 1400 1C for 24 h ((ZrCl_CP_1000(1400 1C-24 h)), with the presence of non-stoichiometric pyrochlore (La1.88Zr2.09O7) and tetragonal zirconia (Fig. 8 (A.b)). The brief thermal treatment of the powders at 1400 1C that occurred in the dilatometric study led to the full crystallization of the compounds from fluorite to pyrochlore. However, pressing the powder appears to favor a phase mixture with the formation of tetragonal zirconia in ZrCl_CP_TMA 1000, whereas the single pyrochlore structure is only observed for the ZrN_CP_TMA 1000 sample. Moreover, the heat treatment time also affects unpressed powders. Indeed, after 24 h at 1400 1C, the fluorite phase observed in calcinated powders (ZrN_CP_1000 and ZrCl_CP_1000) was crystallized into non-stoichiometric pyrochlore (La1.88Zr2.09O7), which is in accordance with the ZrO2–La2O3 binary diagram [35]. As observed for the previous compound (ZrN_CP), the unpressed ZrCl_CP_1200 compound exhibited a different crystallization because a single pyrochlore structure was found [19]. This result confirms that the pressure shaping and/or heat treatment at 1400 1C modified the crystallization of the coprecipitated compound. 3.3.2. Microstructure by SEM investigations Fig. 9 presents the particle size values (surface size (Γs) and volume sizes (Γs)) of the different ZrN_VC calcinated compounds as a function of the heat treatment time. For each
1207
sample, the volume size (Γs) is slightly higher than the surface size (Γs), but they remain relatively close. Comparing all of the samples, the particle size increases with the heat treatment time. High-temperature calcinated compounds (ZrN_VC_1200) are characterized by the smallest particle size. This can be explained by the fact that the particle growth grain size occurs when the temperature during the dilatometric test exceeds the calcination temperature, in this case 1200 1C. Thus, the ZrN_VC_1200 particles only have a very short time to grow, 20 min (Fig. 1). For the ZrN_VC_600 compound, the particles have more time to grow (80 min) because a lower calcination temperature was used (600 1C) (Fig. 1). The difference in particle size is highlighted clearly in the SEM images of the two citrate based compounds presented in Fig. 10(A). The ZrN_VC_TMA 600 compounds have larger particles (Γs¼ 284 nm and Γv¼ 323 nm) than do the ZrN_VC_TMA 1200 compounds (Γs ¼ 170 nm and Γv ¼ 187 nm). Calcination at 1200 1C favors the presence of small particles during the dilatometric process, which are of interest for decreasing the thermal conductivity due to the presence of more grain boundaries. However, high-temperature calcination also leads to material densification and has the opposite effect to the small particle size, which increases the thermal conductivity. Thus, the choice of calcination temperature must be a compromise between thermal conductivity and densification. The SEM images of the samples after the dilatometric tests at 1400 1C (ZrN_CP_TMA 1000 and ZrCl_CP_TMA 1000) show similar microstructures that consist of a mixture of large and small particles (Fig. 10(B)), though they are different crystalline structures. Moreover, the particle size measurements do not show any difference between the compounds. Indeed, the surface size (Γs) and volume size (Γv) of the ZrN_CP_TMA 1000 sample of 263 and 286 nm, respectively, are close to those of the ZrCl_CP_TMA 1000 sample of 241 and 265 nm, respectively. 4. Conclusions Ceramic materials, which are used as topcoats in thermal barrier coatings, are subjected to densification during the operating step of an engine. This phenomenon is partly responsible for the severe degradation of TBC, but little information is available. The dilatometric study of La2Zr2O7 compounds that were synthesized by the citrate route or by coprecipitation revealed that the preparation mode has a significant impact on the shrinkage of the material. The compounds that were elaborated using the citrate route with different zirconium precursors showed significant shrinkage for all of the samples analyzed. A preliminary thermal treatment results in a reduced shrinkage. For instance, from 9.4% (ZrN_VC_TMA 600) to 5.2% (ZrN_VC_TMA 1200), The DTA-TGA-MS results highlighted that the shrinkage during low-temperature calcination (o 1000 1C) is due to the decomposition of residual precursors that are trapped within the compounds. Therefore, preliminary thermal treatments are important for controlling the shrinkage. Moreover,
1208
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209
the higher the calcination temperature is, the smaller the particle size of the citrate-based compound is after the dilatometric tests. The formation of small particles is interesting because of their potential thermal barrier applications due to the presence of more grain boundaries, which decreases the thermal conductivity. However, higher densification can also be achieved at a high temperature, which increases the thermal conductivity. Compounds elaborated by coprecipitation exhibited varying results compared with the citrate-based powders. Their thermal behaviors are dependent on the zirconium counter-cation, contrary to the citrate-based compounds. Chloride countercations favored shrinkage, whereas the nitrate sample with a calcination at 1000 1C resulted in an expansion of the dilatometric sample. Thus, the counter-cation radius affects the thermal behavior of the coprecipitated compounds. XRD and Raman spectroscopy revealed that the powder shaping by pressing combined with the high temperature (1400 1C) that was experienced during the dilatometric experiments modified the structure compared with the raw powder that was calcinated at 1200 1C. The counter-cation also influences the coprecipitated compounds, with a phase mixture of pyrochlore and tetragonal zirconia observed for the chloride counter-cation and a single pyrochlore structure observed for nitrate, as for the citrate-based compounds. The coefficient of thermal expansion for the various compounds prepared by either the citrate or coprecipitation method was approximately 10.10 6 K 1, which is consistent with the literature values. These results show that the thermal behavior of La2Zr2O7 powders can be modified by the preparation mode, the nature of the precursor, and the thermal treatment. To limit the shrinkage responsible for the degradation of the thermal barrier coating, coprecipitated compounds appear to be quite promising. Moreover it would be interesting to study the thermal behavior of synthesized powders after their treatment in a plasma jet. Acknowledgments The authors gratefully acknowledge the DGA (French Ministry of Defense) and CNES National R&T program for their financial support. References [1] T.M. Pollock, S. Tin, Nickel-based superalloys for advanced turbine engines: chemistry, microstructure, and properties, J. Propuls. Power 22 (2006) 361–374. [2] G.E. Fuchs, Solution heat treatment response of a third generation single crystal Ni-base superalloy, Mater. Sci. Eng. A 300 (2001) 52–60. [3] D.R. Clarke, M. Oechsner, N.P. Padture, Thermal barrier coatings for more efficient gas turbine engines, MRS Bull. 37 (2012) 891–898. [4] X.Q. Cao, R. Vassen, D. Stover, Ceramic materials for thermal barrier coatings, J. Eur. Ceram. Soc. 24 (2004) 1–10. [5] V. Lughi, D.R. Clarke, High temperature aging of YSZ coatings and subsequent transformation at low temperature, Surf. Coat. Technol. 200 (2005) 1287–1291.
[6] J. Moon, H. Choi, H. Kim, C. Lee, The effects of the heat treatment on the pahse transformation behavior of plasma sprayed stabilized ZrO2 coatings, Surf. Coat. Technol. 155 (2002) 1–10. [7] C.U. Hardwicke, Y.C. Lau, Advances in thermal spray coatings for gas turbines and energy generation : a review, J. Therm. Spray Technol. 22 (2013) 564–576. [8] D. Zhu, R.A. Miller, Development of advanced low conductivity thermal barrier coatings, Int. J. Ceram. Technol. 1 (2004) 86–94. [9] R. Vassen, X. Cao, F. Tietz, D. Basu, D. Stöver, Zirconates as new materials for thermal barrier coatings, J. Am. Ceram. Soc. 83 (2000) 2023–2028. [10] H. Kido, S. Komarneni, R. Roy, Preparation of La2Zr2O7 by sol–gel route, J. Am. Ceram. Soc. 74 (1991) 422–424. [11] J. Nair, P. Nair, E.B.M. Doesburg, J.G. Van Ommen, J.R.H. Ross, A. J. Burggraff, F. Mizukami, Preparation and characterization of lanthanum zirconate, J. Mater. Sci. 33 (1998) 4517–4523. [12] A. Rabiei, A.G. Evans, Failure mechanisms associated with the thermally grown oxide in plasma sprayed thermal barrier coating, Acta Mater. 48 (2000) 3963–3976. [13] K.W. Schlichting, N.P. Padture, E.H. Jordan, M. Gell, Failures modes in plasma sprayed thermal barrier coatings, Mater. Sci. Eng. A 342 (2003) 120–130. [14] A.G. Evans, D.R. Mumm, J.W. Hutchinson, G.H. Meier, F.S. Petit, Mechanisms controlling the durability of thermal barrier coatings, Prog. Mater. Sci. 46 (2001) 505–553. [15] C. Chen, H. Guo, S. Gong, X. Zhao, P. Xiao, Sintering of electron beam physical vapor deposited thermal barrier coatings under flame shock, Ceram. Int. 39 (2013) 5093–5102. [16] D. Zhu, R.A. Miller, Sintering and creep behavior of plasma sprayed zirconia and hafnia based thermal barrier coating, Surf. Coat. Technol. 108–109 (1998) 114–120. [17] A. Cipitria, I.O. Golosnoy, T.W. Clyne, A sintering model for plasma sprayed zirconia thermal barrier coatings. Part II: Coatings bonded to a rigid substrate, Acta Mater. 57 (2009) 993–1003. [18] G. Thurn, G.A. Schneider, F. Aldinger, High temperature deformation of plasma sprayed ZrO2 thermal barrier coatings, Mater. Sci. Eng. A 233 (1997) 176–182. [19] W. Duarte, A. Meguekam, M. Colas, M. Vardelle, S. Rossignol, Effect of the counter-cation nature and preparation method on the structure of La2Zr2O7, J. Mater. Sci. 50 (2015) 463–475. [20] C. Wang, Y. Wang, Y. Cheng, W. Huang, Z.S. Khan, X. Fan, Y. Wang, B. Zou, X. Cao, Preparation and thermophysical properties of nano-sized Ln2Zr2O7 (Ln¼ La, Nd, Sm and Gd) ceramics with pyrochlore structure, J. Mater. Sci. 47 (2012) 4392–4399. [21] Y. Wang, C. Li, C. Wang, A. Zhang, Y. Chen, Synthesis of La2O3–ZrO2–Y2O3 with various Y2O3 contents via hydrothermal method : a phase evolution and sintering-resistance properties study, Int. J. Appl. Ceram. Technol. 12 (2014) 65–70. [22] J. Wang, S. Bai, H. Zhang, C. Zhang, The structure, thermal expansion coefficient and sintering behavior of Nd3 þ doped La2Zr2O7 for thermal barrier coatings, J. Alloy. Compd. 476 (2009) 89–91. [23] R.D. Cadle, Particle Size Theory and Industrial Applications, Reinhold Publishing Corp, New York, 1965. [24] T. Allen, Particle Size Measurement, third ed., Chapman and Hall, London, 1981. [25] A. Zhang, M. Lu, G. Zhou, S. Wang, Y. Zhou, Combustion synthesis and photoluminescence of Eu3 þ , Dy3 þ doped La2Zr2O7 nanocrystals, J. Phys. Chem. Solids 67 (2006) 2430–2434. [26] D. Prusty, A. Pathak, A. Chintha, B. Mukherjee, A. Chowdhury, Structural investigations on the compositional anomalies in lanthanum zirconate system synthesized by coprecipitation method, J. Am. Ceram. Soc. 97 (2014) 718–724. [27] J.C. Grivel, Y. Zhao, X. Tang, P.G.P.A. Pallewatta, A. Watenphul, M. V. Zimmermann, Thermal decomposition of lanthanum(III) butyrate in argon atmosphere, Thermochim. Acta 566 (2013) 112–117. [28] H. Chen, Y. Gao, Y. Liu, H. Luo, Coprecipitation synthesis and thermal conductivity of La2Zr2O7, J. Alloy. Compd. 480 (2009) 843–848. [29] A. Zhang, M. Lu, G. Zhou, S. Wang, Y. Zhou, Combustion synthesis and photoluminescence of Eu3 þ , Dy3 þ doped La2Zr2O7 nanocrystals, J. Phys. Chem. Solids 67 (2006) 2430–2434.
W. Duarte et al. / Ceramics International 42 (2016) 1197–1209 [30] D. Prusty, A. Pathak, A. Chintha, B. Mukherjee, A. Chowdhury, Structural investigations on the compositional anomalies in lanthanum zirconate system synthesized by coprecipitation method, J. Am. Ceram. Soc. 97 (2014) 718–724. [31] K. Tonsuaadu, A. Zalga, A. Beganskiene, A. Kareiva, Thermoanalytical study of the YSZ precurors prepared by aqueous sol-gel synthesis route, J. Therm. Anal. Calorim. 110 (2012) 77–83. [32] J.C. De Jesus, I. Gonzalez, A. Quevedo, T. Puerta, Thermal decomposition of nickel acetate tetrahydrate an integrated study by TGA, QMS and XPS techniques, J. Mol. Catal. A Chem. 228 (2005) 283–291.
1209
[33] G.A.M. Hussein, Rare earth metal oxides: formation, characterization and catalytic activity. Thermoanalytical and applied pyrolysis, J. Anal. Appl. Pyrolysis 37 (1996) 111–149. [34] T. Settu, Characterisation of MgO–ZrO2 precursor powders prepared by in-situ peptisation of coprecipitated oxalate gel, Ceram. Int. 26 (2000) 517–521. [35] A. Rouanet, Contribution à l’étude des systèmes zircone-oxydes des lanthanides au voisinage de la fusion, Rev. Int. Hautes Temp. Réfract. 8 (1971) 161–180.