- Email: [email protected]
Phase evolution and microstructure analysis of CaZrTi2O7 zirconolite in glass
Ceramics International xxx (xxxx) xxx–xxx
Contents lists available at ScienceDirect
Ceramics International journal homepage: www.elsevier.com/locate/ceramint
Phase evolution and microstructure analysis of CaZrTi2O7 zirconolite in glass ⁎
Linggen Kong , Tao Wei, Yingjie Zhang, Inna Karatchevtseva Nuclear Fuel Cycle Research Theme, Australian Nuclear Science and Technology Organisation, Locked Bag 2001, Kirrawee DC, NSW 2232, Australia
A R T I C L E I N F O
A B S T R A C T
Keywords: CaZrTi2O7 Zirconolite Glass-ceramic Microstructure analysis
The internal crystallization of CaZrTi2O7 zirconolite in a sodium alumino-borosilicate glass has been investigated with powder X-ray diffraction (XRD), scanning electron microscopy and energy dispersion spectroscopy (SEMEDS), transmission electron microscopy (TEM) and selected area electron diffraction (SAED) techniques. The samples have been prepared using a soft chemistry route and ceramic phase evolution has been observed with sintering time. Zirconolite as the dominant phase at 1 h sintering gradually changes to baddeleyite structured materials for longer sintering times. XRD shows that one dominant phase belongs to zirconolite at 3 h sintering, however, SEM-EDS reveals that the dominant ceramic phase is actually baddeleyite phase, which is enclosed by zirconolite phase. TEM and SAED patterns also confirm the crystallization of zirconolite phase in glass. The addition of CaO enhances the formation of zirconolite (i.e. impedes baddeleyite phase) with CaO to glass weight ratio ≤ 35:100.
1. Introduction A glass-ceramic system, which combines the advantages of both ceramic and glass, has attracted recent attention as one of the potential nuclear waste forms for the immobilization of actinide-rich radioactive wastes [1–7]. The crystalline phases are able to host actinides into their stable crystal structures while the residual glass acts as a secondary barrier for the radioactive elements incorporated in the crystalline material as well as a versatile matrix to accommodate processing chemicals and impurities in the actinide-rich waste streams. Zirconolite (CaZrTi2O7) is one of the major compositions in Synroc (synthetic rock) formulation composed of zirconolite, pyrochlore, perovskite, hollandite and rutile structured phases [8–10]. The monoclinic structure of CaZrTi2O7 is formed by stacking of layers of edge-shared TiO6 octahedra, TiO5 trigonal pyramids and layers of Ca2+ and Zr4+ ions. This particular crystal structure is able to accommodate metal ions having widely different oxidation states and ionic radii and also can crystallize in a number of polytypes without significantly changing its structural framework. As a result, the zirconolite-related structures are promising to accommodate larger proportion of actinides and possessing properties such as higher leaching resistance, higher chemical durability and radiation tolerance [8–10]. Zirconolite based glass ceramics (GCs) have been developed as potential waste forms for the immobilization of nuclear wastes, especially the actinide residue wastes, and also have potential for the
⁎
immobilization of separated minor actinides [1–4]. However, unlike lanthanide titanate pyrochlores (Ln2Ti2O7) which can be internally crystallized inside a glass matrix [11,12], the formation of zirconolite phases inside a glass matrix is much more complicated [13–23]. One of the reasons is that zirconolite is thermodynamically instable, especially in a higher silica content glass [4,7]. In a two-step thermal synthesis [13–23], zirconolite glass-ceramic is produced by melting an oxide mixture (the parent glass) at elevated temperatures (≥ 1500 °C), then submitting it to a heat treatment (annealing) at lower temperatures (~ 800 to 1500 °C) to evaluate the degree of zirconolite crystallization. Neodymium is used as an actinide surrogate. In a one-step synthesis [4,7], ceramic and glass components are mixed together in one-pot and the GC samples are prepared. However, an intermediate hot isostatic pressing (HIP) stage is normally utilized to fabricate physically integrated large samples. Maddrell et al. [7] have revealed that the formation of crystalline phases in GC waste forms is dependent on the composition of the glass matrix. Zirconolite forms as the sole crystalline phase only for the most aluminous glass, that is, NaAlSi3O8. Thermodynamically, zirconolite becomes the preferred crystalline phase at low silica activities. In a separate work [4], the authors have used the same glass to study the evolution of crystalline phases as a function of temperature (1000–1200 °C) and time (30–90 min). The work has shown that perovskite and sphene form as transient phases before the formation of zirconolite as the final crystalline phase at 1200 °C. HIPing process has been used to prepare low porosity samples while cold
Corresponding author. E-mail address: [email protected] (L. Kong).
https://doi.org/10.1016/j.ceramint.2018.01.017 Received 22 November 2017; Received in revised form 2 January 2018; Accepted 3 January 2018 0272-8842/ Crown Copyright © 2018 Published by Elsevier Ltd. All rights reserved.
Please cite this article as: Kong, L., Ceramics International (2018), https://doi.org/10.1016/j.ceramint.2018.01.017
Ceramics International xxx (xxxx) xxx–xxx
L. Kong et al.
80 mm2 SDD X-ray microanalysis system. For pelletized samples, they were cut and mounted in an epoxy resin (with cross section on top) and polished to 1 µm diamond finish. A thin carbon film (~ 5 nm) was deposited onto the polished surface. Semi-quantitative analysis was performed on the Oxford Instruments INCA software. A copper metal standard was used for the Quant Optimization to ensure the best fit of the stored profiles within the software. Transmission electron microscopy (TEM) investigation was performed on a JEOL 2200FS (JEOL Ltd., Akishima, Tokyo, Japan) instrument, equipped with a field emission gun (FEG) electron source operated at 200 kV to record selected area electron diffraction patterns (SAED). The phase composition of samples was analyzed by energy dispersive X-ray spectrometry (EDS) under TEM using an Oxford Instruments® EDS detector. The specimens were prepared by finely grinding GC sample, suspending the ground powder in ethanol and depositing several drops onto a carbon coated 100-mesh copper grid.
pressed powder is too porous for microstructure analyses using micrographic imaging [4]. In the present work, a soft chemistry approach is used to prepare the ceramic phase to promote the intimate reaction of the ceramic components and the homogeneity of the product. The mixture of ceramic and glass precursor powder (NaAl0.8B0.2Si3O8) is cold pressed and sintered pressurelessly at 1250 °C in air for 1–16 h, in order to investigate the effect of long time sintering on ceramic phase equilibrium in glass. XRD is used to analyze the crystalline phases of the bulk materials, whereas SEM-EDS and HRTEM-SAED are utilized to investigate the microstructures of the ceramic phases. The results reveal that the combination of both the bulk material and the microstructure analysis techniques lead to a much better understanding of zirconolite phase evolution in a glass matrix. The effect of CaO addition on the formation of crystalline phases is also investigated. The formation of zirconolite phase is enhanced by addition of CaO in glass. 2. Experimental procedure
3. Results and discussion
2.1. Materials and method
3.1. XRD analysis
Tyzor® LA, titanium(IV) bis(ammonium lactato)dihydroxide solution (50 wt% in water), calcium(II) nitrate tetrahydrate (99%+), sodium carbonate (99.5%+), aluminum oxide (99.5%+), boric acid (99.5%+), silicon dioxide (99.9%+) were purchased from SigmaAldrich; zirconium(IV) nitrate pentahydrate was purchased from Fluka and used as received. Milli-Q water was used in all experimental procedures. The titanium content in Tyzor LA was quantitatively determined by gravimetric analysis. Glass (formula as NaAl0.8B0.2Si3O8) precursor powder was prepared by mixing calculated amounts of Na2CO3, Al2O3, H3BO3, and SiO2 with designed molar ratios, and heat treated in a furnace in air at 550 °C for 3 h. The oxide compositions were calculated to be Na2O (11.97 wt%), Al2O3 (15.75 wt%), B2O3 (2.69 wt%), and SiO2 (69.60 wt%). For a typical synthesis, 10 mmol of Tyzor LA (an aqueous solution) was weighed in a 100 mL round alumina crucible, 50 mL of water was added. 5 mmol of calcium nitrate and 5 mmol of zirconium nitrate were mixed thoroughly with the above aqueous solution. A clear aqueous solution was formed upon stirring, which was dried in an oven at 90 °C overnight. The dried product was heated in a furnace in air at 550 °C for 6 h, with temperature ramp rate of 2 °C/min. The majority of the organic species were removed upon heat treatment and a metal-composite (CaZrTi2-oxide) in a powder form was produced, with light yellow pale color. The composite powder was calcined at 1250 °C for 15 h followed by XRD analysis to confirm the formation of zirconolite. For GC sample preparations, the equal weight of glass precursor powder (0.500 g) and composite powder (0.500 g), both being heated at 550 °C, were thoroughly mixed, pelletized using a uniaxial press at approximately 180 MPa, followed by sintering at 1250 °C for different times in air, with heating rate at 3 °C/min and cooling rate at 10 °C/min. The pelletized glass-ceramic samples were termed as GC samples.
Calcination temperature for fabricating GC samples is critical considering the factors including formation of the desired ceramic phase in glass, minimization of the impurities (undesired minor ceramic phases), glass melting and evaporation, and production cost, etc. A significant amount of zircon (ZrSiO4) is present in a GC pelletized sample sintered at 1200 °C for 3 h (data not shown) in this work. As a result, 1250 °C is used in subsequent studies. Higher temperatures are not practical considering the production cost, the evaporation of glass phase and its melting. To confirm the formation of zirconolite at 1250 °C, the (CaZrTi2-oxide) composite, which has been heated at 550 °C for 6 h, is calcined at 1250 °C for 15 h. The XRD peaks with Miller indices in Fig. 1 confirm the zirconolite-2M crystal structure [19]. Almost phase pure zirconolite-2M (monoclinic polytype) is produced at this processing condition. XRD patterns of CaO, ZrO2, and TiO2 powders calcined at 1300 °C for 6 h are shown in Fig. S1 with the intense peaks for CaO (2θ = 27.6, 36.2, 54.5°), ZrO2 (2θ = 28.2, 31.5°) and TiO2 (2θ = 32.3, 37.5, 53.9°). No single metal oxides are observed in zirconolite powder (Fig. 1) suggesting that the ceramic components are completely reacted. Fig. 2 shows the XRD patterns of GC samples calcined at 1250 °C for 1–15 h. The main diagnostic peaks for each phase indicate that samples contain zirconolite (CaZrTi2O7), zircon (ZrSiO4); baddeleyite structured materials (ZrO2 and/or ZrxTi1−xO2 with x ≥ 0.5); perovskite (CaTiO3) and rutile (TiO2), mainly Synroc phases. Small amount of sphene (CaTiSiO5) (2θ = 29.8) is observed especially at shorter sintering times. Fig. S2 displays the XRD patterns of corresponding minerals derived from open source, which clearly show the major peak positions and the relative intensities for each mineral. The XRD data show that in all
2.2. Characterization X-ray diffraction (XRD) patterns were recorded using a PANalytical X′Pert Pro diffractometer (Almelo, the Netherlands) with Cu Kα radiation (λ = 1.5418 Å) at 45 kV and 40 mA. XRD data were obtained using an angular range of 10–65° two theta in a continuous mode with a step size of 0.02° (2θ) with an acquisition time of 2 s per step. GC samples were finely ground manually before analyzing. Scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS) were used to analyze the microstructures and phase compositions. Samples were examined in a Zeiss Ultra Plus scanning electron microscope (Carl Zeiss NTS GmbH, Oberkochen, Germany) operating at 15 kV equipped with an Oxford Instruments X-Max
Fig. 1. XRD pattern of CaZrTi2O7 powder sample calcined at 1250 °C for 15 h.
2
Ceramics International xxx (xxxx) xxx–xxx
L. Kong et al.
changes, the XRD patterns at 1, 3, and 15 h sintering times are replotted in Fig. 3. There is no significant difference between 1 and 3 h for all ceramic phases which display sharp XRD peaks. In contrast, the peaks for perovskite (2θ = 32.9°) and baddeleyite (2θ = 49.3°) are intensified, and peaks for zirconolite become less intense and almost disappear within 14–17° at 15 h sintering (Fig. 3c). At longer sintering time, we cannot eliminate other minor phases formed such as anorthite (CaAl2Si2O8), nevertheless, either the quantity of these minor phases is too small to be detected by XRD or their grains are too small to be identified by SEM. 3.2. SEM analysis The backscattered SEM images of the samples calcined at 1250 °C for 1–15 h are shown in Fig. 4. The corresponding low magnification images are displayed in Fig. S3 to show the uniformity of the samples (distribution of ceramic grains in glass matrix). In general, heterogeneous textures of ceramic phases are observed when sintering time is less than 10 h. The textures of the ceramic phases become relatively homogeneous at sintering time more than 10 h. At short sintering time, most ceramic grains are in elongated shapes with sharp edges. In contrast, the elongation demission becomes shorter, and the edges become blunt at longer sintering times. This phenomenon may suggest that the crystalline phases undergo a dissolution and deposition process on the crystal surfaces, i.e. Ostwald ripening (small crystals dissolve and redeposit onto larger crystals). At 1 h sintering (Fig. 4a), two major phases are observed with SEM. The majority is zirconolite, and the second major phase is baddeleyite (high Ti-bearing zirconia) determined by EDS. In contrast, three minor phases (zircon, perovskite and rutile) are captured by XRD (Fig. 3a), possibly due to the small grain sizes of these minor phases or very similar contrast to residue glass to be captured by SEM. When sintering time increases to 3 h, the major phase changes to baddeleyite (high Tibearing zirconia) while the second phase is zirconolite (Fig. 4b). More importantly, the baddeleyite phase is sitting as cores surrounded/enclosed by zirconolite in most grains which are clearly seen in Fig. S4. This discrepancy between XRD and SEM could be due to the limitation of the penetration depth of X-ray. Such phenomenon has been observed previously [24] in which a powder mixture with a stoichiometric composition of Gd-bearing zirconolite, (Ca0.9Gd0.1)Zr(Ti1.9Al0.1)2O7, is prepared by melting the corresponding mixed oxides at 1600 °C, then cooling to room temperature at rate ~ 100 °C/h. The prepared sample displays a typical phase zonation with ZrTiO4 as cores surrounded by zirconolite, which is in turn enclosed by rutile and perovskite. As sintering time is further increased from 6 h (Fig. 4c) to 10 h (Fig. 4d), zirconolite surrounding/enclosing textures become less obvious, the major phase is still baddeleyite with compositions as ZrxTi1−xO2 with x ≥ 0.5. When sintering time is 15 h (Fig. 4e), all other ceramics become minor except baddeleyite (high Ti-bearing zirconia: ZrxTi1−xO2 with x = 0.5–0.65) as the major phase. For a selected baddeleyite phase (high Ti-bearing zirconia) grain enclosed by zirconolite (Fig. 5a), the EDS line scans are performed. The scanning elemental profiles for Ca, Ti and Zr are showed in Fig. 5b, c and d, respectively. In the glass phase (approximately the first and the last 2 µm of the scan line), significant Ca is detected suggesting that certain amounts of Ca are dissolved/incorporated in glass. For the same reason, small amounts of Ti are observed, but the Zr content is negligible. At the zirconolite layers which are approximately 1 µm thick, the Ca content is increased obviously compared with that in glass matrix. However, both Ti and Zr are increased significantly. In the baddeleyite phase (high Ti-bearing zirconia) core, very limited Ca is detected which could be due to the presence of a thin layer of glass. Both Ti and Zr reach the maximum and plateau in the core area, and the Ti/Zr molar ratio is roughly at 1. For comparison, another ceramic grain (baddeleyite phase with low Ti-bearing zirconia) is selected for EDS line scanning (Fig. S5-a). More Zr is inside core with composition as
Fig. 2. XRD patterns of GC samples sintered at 1250 °C for (a) 1 h, (b) 2 h, (c) 3 h, (d) 4 h, (e) 6 h, (f) 10 h, (g) 15 h. All samples were finely ground before analyzing. The main diagnostic peaks for each phase are indicated: zirconolite (Z); zircon (N); baddeleyite structured materials (B); rutile (R); perovskite (P).
Fig. 3. XRD patterns of GC samples calcined at 1250 °C for (a) 1 h, (b) 3 h, (c) 15 h. All samples were finely ground before analyzing. The main diagnostic peaks for each phase are indicated: zirconolite (Z); zircon (N); baddeleyite structured materials (B); rutile (R); perovskite (P).
sintering times two major phases [zirconolite and zirconium titanate (ZrxTi1−xO2 with x ≥ 0.5)] are present, especially between 1 to 6 h. During this period, the zirconolite peaks at lower 2θ range (14.2, 15.6 and 17.0°) are clearly observed. The minor zircon phase is present at all times, with no obvious intensity changes (2θ = 20.0°). The perovskite content is increased slightly with increase of sintering time, which can be seen from the peak at 32.9°. More rutile is observed with longer sintering time, which is apparent from the peak at 54.1°. The typical peaks at 31.7° and 34.2° for baddeleyite structured materials, i.e., zirconium titanate (ZrxTi1−xO2 with x ≥ 0.5), become more intense from 1 h to 3 h, then less intense from 4 h to 15 h. For a better view of these 3
Ceramics International xxx (xxxx) xxx–xxx
L. Kong et al.
Fig. 5. SEM-EDS line scan images of GC sample (high Ti-bearing zirconia enclosed by CaZrTi2O7) sintered at 1250 °C for 3 h. (a) SEM image, and scan profile of (b) Ca, (c) Ti, (d) Zr.
sample, the GC product is not intact in a pelletized form due to high fluid of glass phase. As a result, significant phase separation occurs. However, adding calcium oxide in glass in the range of ~ 10–35:100 (w/w) not only promotes the formation of zirconolite, but also increases the uniformity of the GC sample (Fig. S7).
3.4. Addition of CaO in glass As determined from Ca line scanning (Fig. 5b), some Ca stays in the residue glass, which is a detrimental factor for producing zirconolite. As the stoichiometric Ca, Zr and Ti are introduced, the more Ca dissolves in glass, the more undesired phases are to be formed. Fig. 7 shows the XRD patterns of GC samples calcined at 1250 °C for 3 h with additional CaO added. With increase of CaO to glass weight ratios from 10:100 (Fig. 7a), to 20:100 (Fig. 7b) and 35:100 (Fig. 7c), the typical peaks of baddeleyite (2θ = 31.7° and 34.2°) become less intense, suggesting less baddeleyite phase is formed. When CaO to glass weight ratio increases to 50:100 (Fig. 7d), no typical baddeleyite peaks are present. However, strong perovskite peaks (2θ = 33.1° and 47.4°) are present instead. As discussed previously, at 3 h sintering without adding extra CaO (Fig. 4b), the major phase is baddeleyite. The second major phase is zirconolite, which encloses the baddeleyite in most grains. After adding small amounts of CaO (CaO:glass = 10:100 w/w), the major phase is zirconolite with main minor phase as baddeleyite as shown in the SEM image (Fig. 8a). No core-shell microstructured grains are observed which can be seen from the high magnification SEM image (Fig. S6a). Further increasing CaO in glass (weight ratio at 20:100 and 35:100) leads to the enhancement of zirconolite formation and small amounts of baddeleyite (ZrxTi1−xO2 with x ≥ 0.5) observed (Fig. 8b–c and Figs. S6b-c). When CaO to glass weight ratio increases to 50:100, two major phases including zirconolite and perovskite (CaTiO3) are observed, as well as other minor phases such as CaO in glass (Fig. 8d and Fig. S6d). This observation is consistent with XRD results. It is interesting to note that no apparent baddeleyite phase is observed for this sample by both SEM and XRD. It is found that with increase of CaO content in glass, the viscosity of the glass matrix is decreased. For the highest CaO amount
3.5. Discussion For a successful formation of a zirconolite crystal in an amorphous glass matrix, both thermodynamic and kinetic factors should be considered [4,7]. The recent study [4] reveals that in the sintering time range of 30–90 min at 1200 °C, zirconolite is probed to be the only crystalline phase in an alumino-silicate glass with XRD characterization. To form zirconolite via metal oxides: CaO + ZrO2 + 2TiO2 → CaZrTi2O7
(1)
the standard Gibbs free energy (ΔG) for the formation of zirconolite is −92.0 kJ mol–1, suggesting the formation of zirconolite is thermodynamically favorable. The standard molar Gibbs free energies (ΔG°) of CaO, ZrO2, TiO2, CaZrTi2O7 are −603.1, −1042.9, −888.8, −3515.6 kJ mol−1 at 298.15 K, respectively [4,7]. It is reasonable to assume that these data will remain essentially invariant with temperature [7]. However, we believe longer sintering time should be pursued to study the thermodynamic stability of zirconolite in glass. In certain circumstances, kinetics is overwhelmed due to the high variation of solubility of metal oxides in glass at some processing conditions such as longer phase equilibrium time and high fluid of the glass matrix. With increasing sintering time at 1250 °C, more baddeleyite structured material is observed in this work. As a result, the following decomposition reaction may occur: 5
Ceramics International xxx (xxxx) xxx–xxx
L. Kong et al.
Fig. 6. (a) TEM image of glass and B-phase, (b) TEM image of CaZrTi2O7, (c) HRTEM image of CaZrTi2O7 with inset as live fast fouier transform (FFT) of view [001], (d) SAED pattern in [001] zone axis. GC sample was calcined at 1250 °C for 3 h. B-phase is baddeleyite phase (high Ti-bearing zirconia).
CaZrTi2O7 → ZrTiO4 + CaO + TiO2
(2)
the ΔG° of ZrTiO4 is −1935.7 kJ mol−1 at 298.15 K [25]. The standard Gibbs free energy (ΔG) for the equilibrium reaction will be: ΔG = ΔG° + RT ln Qr, i.e. ΔG = 88.0 + RT ln Qr, where Qr is the reaction quotient defined here by Qr = [ZrTiO4][CaO][TiO2]/[CaZrTi2O7], with [CaO] representing the chemical activity of CaO, etc. Even though the ΔG° is not favorable for the decomposition of zirconolite, the changes of the chemical activities of CaO and TiO2 with time play a major role kinetically. There is no sufficient data showing the metal oxide solubility in variety types of glasses. Ojovan and Lee list the approximate solubility limits of elements in silicate glasses [26]. Ca has 15–25 wt% solubility limit whereas Ti and Zr have 5–15 wt%. Our SEM-EDS line scan experiments confirm that significant amounts of CaO and small amounts of TiO2 are dissolved in the residue glass. The major ceramic phase is ZrxTi1−xO2 with x ≥ 0.5 in longer sintering time. As such only trace amounts of Zr is observed in glass. In addition, little sphene (CaTiSiO5) is detected but with perovskite (CaTiO3) instead. With the chemical activity of CaO decreases dramatically due to its
Fig. 7. XRD patterns of GC samples calcined at 1250 °C for 3 h with CaO added in glass. CaO to glass weight ratio is (a) 10:100, (b) 20:100, (c) 35:100, (d) 50:100. All samples were finely ground before analyzing. The main diagnostic peaks for each phase are indicated: zirconolite (Z); baddeleyite structured materials (B); rutile (R); perovskite (P).
6
Ceramics International xxx (xxxx) xxx–xxx
L. Kong et al.
Fig. 8. Backscattered SEM images of GC samples calcined at 1250 °C for 3 h with CaO added in glass formulation. CaO to glass weight ratio is (a) 10:100, (b) 20:100, (c) 35:100, (d) 50:100. Dark grey (phase 1): CaZrTi2O7, light grey (phase 2): high Ti-bearing zirconia, bright grey (phase 3): low Ti-bearing zirconia, dim grey (phase 4): CaTiO3, determined by EDS. Black area is residual glass.
4. Conclusions
high solubility in glass, the reaction shifts to right side with time. In contrary, if extra CaO is added in glass phase, chemical activity of CaO is increased, which leads to the chemical reaction (2) shifting to left side. As a result, zirconolite phase is kinetically stabilized. It has been concluded [24] that the formation of ceramic phases follows the crystallization sequence of Ti-bearing zirconia → ZrTiO4 phase → Zr-rich zirconolite → Zr-poor zirconolite → rutile/perovskite from the melt with a zirconolite composition. This sequence is induced by a fractional crystallization process, in which Zr-rich phases tend to crystallize first, resulting in continuous depletion of Zr in the melt. Consistent with this melt compositional evolution, the Zr content in zirconolite decreases from the area next to the ZrTiO4 phase to areas next to rutile or perovskite. The results in the present work show the opposite crystallization sequence with longer sintering times at fixed temperature in a glass matrix. It is generally viewed that adding nucleating agents like TiO2 and ZrO2 to the glass composition will lead to the formation of ZrTiO4 nuclei, followed by the formation of zirconolite [27]. In addition, it has been reported [28] that baddeleyite (m-ZrO2) crystals form at the expense of zirconolite in the bulk of glass-ceramics heat treated at ≥ 1250 °C. In this study, zirconolite forms directly from (CaZrTi2-oxide) composite at short sintering time (≤ 1 h) as thermodynamically favorable phase, then it changes to mainly baddeleyite phase with time. Even though the thermodynamic data favorites the formation of zirconolite, the baddeleyite becomes the dominating phase kinetically with time due to the Ostwald ripening, and more importantly the high solubility of CaO in the melting glass. Our experimental results are consistent with the literature [4,28], and support the claims from both thermodynamic and kinetic point of views.
The internal crystallization of CaZrTi2O7 zirconolite in a sodium alumino-borosilicate glass has been investigated under pressureless sintering at 1250 °C in air. The ceramic precursor is prepared using a soft chemistry route to ensure ultimate mixing of the reactants in molecular level, thus higher homogeneity of the product. Neither organic solvent nor mechanical milling procedure is used during processing. Ceramic crystalline phase evolution in glass is observed with sintering time, with zirconolite as the dominant phase at short sintering and changes to baddeleyite phases for long sintering time. When sintering at 1250 °C for 3 h, XRD shows that zirconolite is the dominant crystalline phase, however, SEM and EDS line scan results reveal that the dominant ceramic phase is actually baddeleyite (high Ti-bearing zirconia, ZrxTi1−xO2 with x ≥ 0.5), which is enclosed by zirconolite phase forming a core-shell microstructured grains. HRTEM images and SAED patterns confirm that the zirconolite grains possess a high level of crystallographic perfection at the atomic scale suggesting the complete crystallization. These observations lead to the conclusion that both analytical techniques for bulk samples (XRD) and microstructure of micro domains (SEM, SEM-EDS, TEM, SAED) should be utilized in complementary to study the structures and the ceramic phase evolution in glass. In addition, both thermodynamic and kinetic factors should be considered for a successful synthesis of the desired ceramic phase in glass. The addition of CaO with CaO to glass weight ratios at ~ 10–35:100 stabilizes the zirconolite phase kinetically, thus enhances the zirconolite formation.
7
Ceramics International xxx (xxxx) xxx–xxx
L. Kong et al.
Acknowledgments
J. Nucl. Mater. 494 (2017) 29–36. [13] C.-Z. Liao, C. Liu, P.-H. Lee, M.C. Stennett, N.C. Hyatt, K. Shih, Combined quantitative X-ray diffraction, scanning electron microscopy, and transmission electron microscopy investigations of crystal evolution in CaO–Al2O3–SiO2–TiO2–ZrO2–Nd2O3–Na2O system, Cryst. Growth Des. 17 (2017) 1079–1087. [14] C.-Z. Liao, C. Liu, M. Su, K. Shih, Quantification of the partitioning ratio of minor actinide surrogates between zirconolite and glass in glass-ceramic for nuclear waste disposal, Inorg. Chem. 56 (2017) 9913–9921. [15] C.-Z. Liao, K. Shih, W.E. Lee, Crystal structures of Al−Nd codoped zirconolite derived from glass matrix and powder sintering, Inorg. Chem. 54 (2015) 7353–7361. [16] P. Loiseau, D. Caurant, Glass–ceramic nuclear waste forms obtained by crystallization of SiO2–Al2O3–CaO–ZrO2–TiO2 glasses containing lanthanides (Ce, Nd, Eu, Gd, Yb) and actinides (Th): study of the crystallization from the surface, J. Nucl. Mater. 402 (2010) 38–54. [17] C. Martin, I. Ribet, P. Frugier, S. Gin, Alteration kinetics of the glass-ceramic zirconolite and role of the alteration film – comparison with the SON68 glass, J. Nucl. Mater. 366 (2007) 277–287. [18] D. Caurant, I. Bardez, P. Loiseau, Crystallization of CaHf1−xZrxTi2O7 (0 < x < 1) zirconolite in SiO2–Al2O3–CaO–Na2O–TiO2–HfO2–ZrO2–Nd2O3 glasses, J. Mater. Sci. 42 (2007) 10203–10218. [19] D. Caurant, P. Loiseau, I. Bardez, C. Gervais, Effect of Al2O3 concentration on zirconolite (Ca(Zr,Hf)Ti2O7) crystallization in (TiO2,ZrO2,HfO2)-rich SiO2–Al2O3–CaO–Na2O glasses, J. Mater. Sci. 42 (2007) 8558–8570. [20] D. Caurant, O. Majerus, P. Loiseau, I. Bardez, N. Baffier, J.L. Dussossoy, Crystallization of neodymium-rich phases in silicate glasses developed for nuclear waste immobilization, J. Nucl. Mater. 354 (2006) 143–162. [21] P. Loiseau, D. Caurant, N. Baffier, L. Mazerolles, C. Fillet, , Glass–ceramic nuclear waste forms obtained from SiO2–Al2O3–CaO–ZrO2–TiO2 glasses containing lanthanides (Ce, Nd, Eu, Gd, Yb) and actinides (Th): study of internal crystallization, J. Nucl. Mater. 335 (2004) 14–32. [22] P. Loiseau, D. Caurant, O. Majerus, N. Baffier, C. Fillet, , Crystallization study of (TiO2,ZrO2)-rich SiO2-Al2O3-CaO glasses Part II Surface and internal crystallization processes investigated by differential thermal analysis (DTA), J. Mater. Sci. 38 (2003) 853–864. [23] P. Loiseau, D. Caurant, O. Majerus, N. Baffier, C. Fillet, Crystallization study of (TiO2, ZrO2)-rich SiO2-Al2O3-CaO glasses Part I Preparation and characterization of zirconolite-based glass-ceramics, J. Mater. Sci. 38 (2003) 843–852. [24] H. Xu, Y. Wang, Crystallization sequence and microstructure evolution of Synroc samples crystallized from CaZrTi2O7 melts, J. Nucl. Mater. 279 (2000) 100–106. [25] R.L. Putnam, A. Navrotsky, Thermodynamics of formation for zirconolite (CaZrTi2O7) from T = 298.15 K to T = 1500 K, J. Chem. Thermodyn. 31 (1999) 229–243. [26] M.I. Ojovan, W.E. Lee, An Introduction to Nuclear Waste Immobilisation, 2nd ed., Elsevier Ltd, Amsterdam, etc., 2014, p. 251 (Chapt. 17). [27] M. Chavoutier, D. Caurant, O. Majérus, R. Boulesteix, P. Loiseau, C. Jousseaume, E. Brunet, E. Lecomte, Effect of TiO2 content on the crystallization and the color of (ZrO2,TiO2)-doped Li2O–Al2O3–SiO2 glasses, J. Non-Cryst. Solids 384 (2014) 15–24. [28] P. Loiseau, D. Caurant, N. Baffier, L. Mazerolles, C. Fillet, Development of zirconolite-based glass-ceramics for the conditioning of actinides, Mat. Res. Soc. Symp. Proc. 663 (2000) 179–187.
We would like to thank the Nuclear Materials Development and Characterization Platform at ANSTO for providing facility access to carry out the material synthesis and characterizations. Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ceramint.2018.01.017 References [1] Y. Zhang, D.J. Gregg, L. Kong, M. Jovanovich, G. Triani, Zirconolite glass-ceramics for plutonium immobilization: the effects of processing redox conditions and plutonium waste loadings, J. Nucl. Mater. 490 (2017) 238–241. [2] M.W.A. Stewart, S.A. Moricca, Y. Zhang, R.A. Day, B.D. Begg, C.R. Scales, E.R. Maddrell, J. Hobbs, Glass-Ceramic Waste Forms for Uranium and Plutonium Residues Wastes – 13164, WM2013 Conference, Phoenix USA, 2013. [3] E.R. Vance, S. Moricca, B.D. Begg, M.W.A. Stewart, Y. Zhang, M.L. Carter, Advantages hot isostatically pressed ceramic and glass-ceramic waste forms bring to the immobilization of challenging intermediate- and high-level nuclear wastes, Adv. Sci. Technol. 73 (2010) 130–135. [4] E.R. Maddrell, H.C. Paterson, S.E. May, K.M. Burns, Phase evolution in zirconolite glass-ceramic wasteforms, J. Nucl. Mater. 493 (2017) 380–387. [5] L. Wu, X. Wang, H. Li, Y. Teng, L. Peng, The effects of sulfate content on crystalline phase, microstructure, and chemical durability of zirconolite−barium borosilicate glass-ceramics, J. Nucl. Mater. 478 (2016) 303–309. [6] H. Li, L. Wu, D. Xu, X. Wang, Y. Teng, Y. Li, Structure and chemical durability of barium borosilicate glass-ceramics containing zirconolite and titanite crystalline phases, J. Nucl. Mater. 466 (2015) 484–490. [7] E. Maddrell, S. Thornber, N.C. Hyatt, The influence of glass composition on crystalline phase stability in glass-ceramic wasteforms, J. Nucl. Mater. 456 (2015) 461–466. [8] M. Jafar, S. Phapale, S.N. Achary, R. Mishra, A.K. Tyagi, High-temperature crystallographic and thermodynamic investigations on synthetic zirconolite (CaZrTi2O7), J. Therm. Anal. Calorim. (2017), http://dx.doi.org/10.1007/s10973017-6816-0. [9] M. Jafar, P. Sengupta, S.N. Achary, A.K. Tyagi, Phase evolution and microstructural studies in CaZrTi2O7–Nd2Ti2O7 system, J. Am. Ceram. Soc. 97 (2014) 609–616. [10] M. Jafar, P. Sengupta, S.N. Achary, A.K. Tyagi, Phase evolution and microstructural studies in CaZrTi2O7 (zirconolite)–Sm2Ti2O7 (pyrochlore) system, J. Eur. Ceram. Soc. 34 (2014) 4373–4381. [11] L. Kong, I. Karatchevtseva, Y. Zhang, A new method for production of glassLn2Ti2O7 pyrochlore (Ln = Gd, Tb, Er, Yb), J. Eur. Ceram. Soc. 37 (2017) 4963–4972. [12] L. Kong, Y. Zhang, I. Karatchevtseva, Preparation of Y2Ti2O7 pyrochlore glassceramics as potential waste forms for actinides: the effects of processing conditions,
8
Contents lists available at ScienceDirect
Ceramics International journal homepage: www.elsevier.com/locate/ceramint
Phase evolution and microstructure analysis of CaZrTi2O7 zirconolite in glass ⁎
Linggen Kong , Tao Wei, Yingjie Zhang, Inna Karatchevtseva Nuclear Fuel Cycle Research Theme, Australian Nuclear Science and Technology Organisation, Locked Bag 2001, Kirrawee DC, NSW 2232, Australia
A R T I C L E I N F O
A B S T R A C T
Keywords: CaZrTi2O7 Zirconolite Glass-ceramic Microstructure analysis
The internal crystallization of CaZrTi2O7 zirconolite in a sodium alumino-borosilicate glass has been investigated with powder X-ray diffraction (XRD), scanning electron microscopy and energy dispersion spectroscopy (SEMEDS), transmission electron microscopy (TEM) and selected area electron diffraction (SAED) techniques. The samples have been prepared using a soft chemistry route and ceramic phase evolution has been observed with sintering time. Zirconolite as the dominant phase at 1 h sintering gradually changes to baddeleyite structured materials for longer sintering times. XRD shows that one dominant phase belongs to zirconolite at 3 h sintering, however, SEM-EDS reveals that the dominant ceramic phase is actually baddeleyite phase, which is enclosed by zirconolite phase. TEM and SAED patterns also confirm the crystallization of zirconolite phase in glass. The addition of CaO enhances the formation of zirconolite (i.e. impedes baddeleyite phase) with CaO to glass weight ratio ≤ 35:100.
1. Introduction A glass-ceramic system, which combines the advantages of both ceramic and glass, has attracted recent attention as one of the potential nuclear waste forms for the immobilization of actinide-rich radioactive wastes [1–7]. The crystalline phases are able to host actinides into their stable crystal structures while the residual glass acts as a secondary barrier for the radioactive elements incorporated in the crystalline material as well as a versatile matrix to accommodate processing chemicals and impurities in the actinide-rich waste streams. Zirconolite (CaZrTi2O7) is one of the major compositions in Synroc (synthetic rock) formulation composed of zirconolite, pyrochlore, perovskite, hollandite and rutile structured phases [8–10]. The monoclinic structure of CaZrTi2O7 is formed by stacking of layers of edge-shared TiO6 octahedra, TiO5 trigonal pyramids and layers of Ca2+ and Zr4+ ions. This particular crystal structure is able to accommodate metal ions having widely different oxidation states and ionic radii and also can crystallize in a number of polytypes without significantly changing its structural framework. As a result, the zirconolite-related structures are promising to accommodate larger proportion of actinides and possessing properties such as higher leaching resistance, higher chemical durability and radiation tolerance [8–10]. Zirconolite based glass ceramics (GCs) have been developed as potential waste forms for the immobilization of nuclear wastes, especially the actinide residue wastes, and also have potential for the
⁎
immobilization of separated minor actinides [1–4]. However, unlike lanthanide titanate pyrochlores (Ln2Ti2O7) which can be internally crystallized inside a glass matrix [11,12], the formation of zirconolite phases inside a glass matrix is much more complicated [13–23]. One of the reasons is that zirconolite is thermodynamically instable, especially in a higher silica content glass [4,7]. In a two-step thermal synthesis [13–23], zirconolite glass-ceramic is produced by melting an oxide mixture (the parent glass) at elevated temperatures (≥ 1500 °C), then submitting it to a heat treatment (annealing) at lower temperatures (~ 800 to 1500 °C) to evaluate the degree of zirconolite crystallization. Neodymium is used as an actinide surrogate. In a one-step synthesis [4,7], ceramic and glass components are mixed together in one-pot and the GC samples are prepared. However, an intermediate hot isostatic pressing (HIP) stage is normally utilized to fabricate physically integrated large samples. Maddrell et al. [7] have revealed that the formation of crystalline phases in GC waste forms is dependent on the composition of the glass matrix. Zirconolite forms as the sole crystalline phase only for the most aluminous glass, that is, NaAlSi3O8. Thermodynamically, zirconolite becomes the preferred crystalline phase at low silica activities. In a separate work [4], the authors have used the same glass to study the evolution of crystalline phases as a function of temperature (1000–1200 °C) and time (30–90 min). The work has shown that perovskite and sphene form as transient phases before the formation of zirconolite as the final crystalline phase at 1200 °C. HIPing process has been used to prepare low porosity samples while cold
Corresponding author. E-mail address: [email protected] (L. Kong).
https://doi.org/10.1016/j.ceramint.2018.01.017 Received 22 November 2017; Received in revised form 2 January 2018; Accepted 3 January 2018 0272-8842/ Crown Copyright © 2018 Published by Elsevier Ltd. All rights reserved.
Please cite this article as: Kong, L., Ceramics International (2018), https://doi.org/10.1016/j.ceramint.2018.01.017
Ceramics International xxx (xxxx) xxx–xxx
L. Kong et al.
80 mm2 SDD X-ray microanalysis system. For pelletized samples, they were cut and mounted in an epoxy resin (with cross section on top) and polished to 1 µm diamond finish. A thin carbon film (~ 5 nm) was deposited onto the polished surface. Semi-quantitative analysis was performed on the Oxford Instruments INCA software. A copper metal standard was used for the Quant Optimization to ensure the best fit of the stored profiles within the software. Transmission electron microscopy (TEM) investigation was performed on a JEOL 2200FS (JEOL Ltd., Akishima, Tokyo, Japan) instrument, equipped with a field emission gun (FEG) electron source operated at 200 kV to record selected area electron diffraction patterns (SAED). The phase composition of samples was analyzed by energy dispersive X-ray spectrometry (EDS) under TEM using an Oxford Instruments® EDS detector. The specimens were prepared by finely grinding GC sample, suspending the ground powder in ethanol and depositing several drops onto a carbon coated 100-mesh copper grid.
pressed powder is too porous for microstructure analyses using micrographic imaging [4]. In the present work, a soft chemistry approach is used to prepare the ceramic phase to promote the intimate reaction of the ceramic components and the homogeneity of the product. The mixture of ceramic and glass precursor powder (NaAl0.8B0.2Si3O8) is cold pressed and sintered pressurelessly at 1250 °C in air for 1–16 h, in order to investigate the effect of long time sintering on ceramic phase equilibrium in glass. XRD is used to analyze the crystalline phases of the bulk materials, whereas SEM-EDS and HRTEM-SAED are utilized to investigate the microstructures of the ceramic phases. The results reveal that the combination of both the bulk material and the microstructure analysis techniques lead to a much better understanding of zirconolite phase evolution in a glass matrix. The effect of CaO addition on the formation of crystalline phases is also investigated. The formation of zirconolite phase is enhanced by addition of CaO in glass. 2. Experimental procedure
3. Results and discussion
2.1. Materials and method
3.1. XRD analysis
Tyzor® LA, titanium(IV) bis(ammonium lactato)dihydroxide solution (50 wt% in water), calcium(II) nitrate tetrahydrate (99%+), sodium carbonate (99.5%+), aluminum oxide (99.5%+), boric acid (99.5%+), silicon dioxide (99.9%+) were purchased from SigmaAldrich; zirconium(IV) nitrate pentahydrate was purchased from Fluka and used as received. Milli-Q water was used in all experimental procedures. The titanium content in Tyzor LA was quantitatively determined by gravimetric analysis. Glass (formula as NaAl0.8B0.2Si3O8) precursor powder was prepared by mixing calculated amounts of Na2CO3, Al2O3, H3BO3, and SiO2 with designed molar ratios, and heat treated in a furnace in air at 550 °C for 3 h. The oxide compositions were calculated to be Na2O (11.97 wt%), Al2O3 (15.75 wt%), B2O3 (2.69 wt%), and SiO2 (69.60 wt%). For a typical synthesis, 10 mmol of Tyzor LA (an aqueous solution) was weighed in a 100 mL round alumina crucible, 50 mL of water was added. 5 mmol of calcium nitrate and 5 mmol of zirconium nitrate were mixed thoroughly with the above aqueous solution. A clear aqueous solution was formed upon stirring, which was dried in an oven at 90 °C overnight. The dried product was heated in a furnace in air at 550 °C for 6 h, with temperature ramp rate of 2 °C/min. The majority of the organic species were removed upon heat treatment and a metal-composite (CaZrTi2-oxide) in a powder form was produced, with light yellow pale color. The composite powder was calcined at 1250 °C for 15 h followed by XRD analysis to confirm the formation of zirconolite. For GC sample preparations, the equal weight of glass precursor powder (0.500 g) and composite powder (0.500 g), both being heated at 550 °C, were thoroughly mixed, pelletized using a uniaxial press at approximately 180 MPa, followed by sintering at 1250 °C for different times in air, with heating rate at 3 °C/min and cooling rate at 10 °C/min. The pelletized glass-ceramic samples were termed as GC samples.
Calcination temperature for fabricating GC samples is critical considering the factors including formation of the desired ceramic phase in glass, minimization of the impurities (undesired minor ceramic phases), glass melting and evaporation, and production cost, etc. A significant amount of zircon (ZrSiO4) is present in a GC pelletized sample sintered at 1200 °C for 3 h (data not shown) in this work. As a result, 1250 °C is used in subsequent studies. Higher temperatures are not practical considering the production cost, the evaporation of glass phase and its melting. To confirm the formation of zirconolite at 1250 °C, the (CaZrTi2-oxide) composite, which has been heated at 550 °C for 6 h, is calcined at 1250 °C for 15 h. The XRD peaks with Miller indices in Fig. 1 confirm the zirconolite-2M crystal structure [19]. Almost phase pure zirconolite-2M (monoclinic polytype) is produced at this processing condition. XRD patterns of CaO, ZrO2, and TiO2 powders calcined at 1300 °C for 6 h are shown in Fig. S1 with the intense peaks for CaO (2θ = 27.6, 36.2, 54.5°), ZrO2 (2θ = 28.2, 31.5°) and TiO2 (2θ = 32.3, 37.5, 53.9°). No single metal oxides are observed in zirconolite powder (Fig. 1) suggesting that the ceramic components are completely reacted. Fig. 2 shows the XRD patterns of GC samples calcined at 1250 °C for 1–15 h. The main diagnostic peaks for each phase indicate that samples contain zirconolite (CaZrTi2O7), zircon (ZrSiO4); baddeleyite structured materials (ZrO2 and/or ZrxTi1−xO2 with x ≥ 0.5); perovskite (CaTiO3) and rutile (TiO2), mainly Synroc phases. Small amount of sphene (CaTiSiO5) (2θ = 29.8) is observed especially at shorter sintering times. Fig. S2 displays the XRD patterns of corresponding minerals derived from open source, which clearly show the major peak positions and the relative intensities for each mineral. The XRD data show that in all
2.2. Characterization X-ray diffraction (XRD) patterns were recorded using a PANalytical X′Pert Pro diffractometer (Almelo, the Netherlands) with Cu Kα radiation (λ = 1.5418 Å) at 45 kV and 40 mA. XRD data were obtained using an angular range of 10–65° two theta in a continuous mode with a step size of 0.02° (2θ) with an acquisition time of 2 s per step. GC samples were finely ground manually before analyzing. Scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS) were used to analyze the microstructures and phase compositions. Samples were examined in a Zeiss Ultra Plus scanning electron microscope (Carl Zeiss NTS GmbH, Oberkochen, Germany) operating at 15 kV equipped with an Oxford Instruments X-Max
Fig. 1. XRD pattern of CaZrTi2O7 powder sample calcined at 1250 °C for 15 h.
2
Ceramics International xxx (xxxx) xxx–xxx
L. Kong et al.
changes, the XRD patterns at 1, 3, and 15 h sintering times are replotted in Fig. 3. There is no significant difference between 1 and 3 h for all ceramic phases which display sharp XRD peaks. In contrast, the peaks for perovskite (2θ = 32.9°) and baddeleyite (2θ = 49.3°) are intensified, and peaks for zirconolite become less intense and almost disappear within 14–17° at 15 h sintering (Fig. 3c). At longer sintering time, we cannot eliminate other minor phases formed such as anorthite (CaAl2Si2O8), nevertheless, either the quantity of these minor phases is too small to be detected by XRD or their grains are too small to be identified by SEM. 3.2. SEM analysis The backscattered SEM images of the samples calcined at 1250 °C for 1–15 h are shown in Fig. 4. The corresponding low magnification images are displayed in Fig. S3 to show the uniformity of the samples (distribution of ceramic grains in glass matrix). In general, heterogeneous textures of ceramic phases are observed when sintering time is less than 10 h. The textures of the ceramic phases become relatively homogeneous at sintering time more than 10 h. At short sintering time, most ceramic grains are in elongated shapes with sharp edges. In contrast, the elongation demission becomes shorter, and the edges become blunt at longer sintering times. This phenomenon may suggest that the crystalline phases undergo a dissolution and deposition process on the crystal surfaces, i.e. Ostwald ripening (small crystals dissolve and redeposit onto larger crystals). At 1 h sintering (Fig. 4a), two major phases are observed with SEM. The majority is zirconolite, and the second major phase is baddeleyite (high Ti-bearing zirconia) determined by EDS. In contrast, three minor phases (zircon, perovskite and rutile) are captured by XRD (Fig. 3a), possibly due to the small grain sizes of these minor phases or very similar contrast to residue glass to be captured by SEM. When sintering time increases to 3 h, the major phase changes to baddeleyite (high Tibearing zirconia) while the second phase is zirconolite (Fig. 4b). More importantly, the baddeleyite phase is sitting as cores surrounded/enclosed by zirconolite in most grains which are clearly seen in Fig. S4. This discrepancy between XRD and SEM could be due to the limitation of the penetration depth of X-ray. Such phenomenon has been observed previously [24] in which a powder mixture with a stoichiometric composition of Gd-bearing zirconolite, (Ca0.9Gd0.1)Zr(Ti1.9Al0.1)2O7, is prepared by melting the corresponding mixed oxides at 1600 °C, then cooling to room temperature at rate ~ 100 °C/h. The prepared sample displays a typical phase zonation with ZrTiO4 as cores surrounded by zirconolite, which is in turn enclosed by rutile and perovskite. As sintering time is further increased from 6 h (Fig. 4c) to 10 h (Fig. 4d), zirconolite surrounding/enclosing textures become less obvious, the major phase is still baddeleyite with compositions as ZrxTi1−xO2 with x ≥ 0.5. When sintering time is 15 h (Fig. 4e), all other ceramics become minor except baddeleyite (high Ti-bearing zirconia: ZrxTi1−xO2 with x = 0.5–0.65) as the major phase. For a selected baddeleyite phase (high Ti-bearing zirconia) grain enclosed by zirconolite (Fig. 5a), the EDS line scans are performed. The scanning elemental profiles for Ca, Ti and Zr are showed in Fig. 5b, c and d, respectively. In the glass phase (approximately the first and the last 2 µm of the scan line), significant Ca is detected suggesting that certain amounts of Ca are dissolved/incorporated in glass. For the same reason, small amounts of Ti are observed, but the Zr content is negligible. At the zirconolite layers which are approximately 1 µm thick, the Ca content is increased obviously compared with that in glass matrix. However, both Ti and Zr are increased significantly. In the baddeleyite phase (high Ti-bearing zirconia) core, very limited Ca is detected which could be due to the presence of a thin layer of glass. Both Ti and Zr reach the maximum and plateau in the core area, and the Ti/Zr molar ratio is roughly at 1. For comparison, another ceramic grain (baddeleyite phase with low Ti-bearing zirconia) is selected for EDS line scanning (Fig. S5-a). More Zr is inside core with composition as
Fig. 2. XRD patterns of GC samples sintered at 1250 °C for (a) 1 h, (b) 2 h, (c) 3 h, (d) 4 h, (e) 6 h, (f) 10 h, (g) 15 h. All samples were finely ground before analyzing. The main diagnostic peaks for each phase are indicated: zirconolite (Z); zircon (N); baddeleyite structured materials (B); rutile (R); perovskite (P).
Fig. 3. XRD patterns of GC samples calcined at 1250 °C for (a) 1 h, (b) 3 h, (c) 15 h. All samples were finely ground before analyzing. The main diagnostic peaks for each phase are indicated: zirconolite (Z); zircon (N); baddeleyite structured materials (B); rutile (R); perovskite (P).
sintering times two major phases [zirconolite and zirconium titanate (ZrxTi1−xO2 with x ≥ 0.5)] are present, especially between 1 to 6 h. During this period, the zirconolite peaks at lower 2θ range (14.2, 15.6 and 17.0°) are clearly observed. The minor zircon phase is present at all times, with no obvious intensity changes (2θ = 20.0°). The perovskite content is increased slightly with increase of sintering time, which can be seen from the peak at 32.9°. More rutile is observed with longer sintering time, which is apparent from the peak at 54.1°. The typical peaks at 31.7° and 34.2° for baddeleyite structured materials, i.e., zirconium titanate (ZrxTi1−xO2 with x ≥ 0.5), become more intense from 1 h to 3 h, then less intense from 4 h to 15 h. For a better view of these 3
Ceramics International xxx (xxxx) xxx–xxx
L. Kong et al.
Fig. 5. SEM-EDS line scan images of GC sample (high Ti-bearing zirconia enclosed by CaZrTi2O7) sintered at 1250 °C for 3 h. (a) SEM image, and scan profile of (b) Ca, (c) Ti, (d) Zr.
sample, the GC product is not intact in a pelletized form due to high fluid of glass phase. As a result, significant phase separation occurs. However, adding calcium oxide in glass in the range of ~ 10–35:100 (w/w) not only promotes the formation of zirconolite, but also increases the uniformity of the GC sample (Fig. S7).
3.4. Addition of CaO in glass As determined from Ca line scanning (Fig. 5b), some Ca stays in the residue glass, which is a detrimental factor for producing zirconolite. As the stoichiometric Ca, Zr and Ti are introduced, the more Ca dissolves in glass, the more undesired phases are to be formed. Fig. 7 shows the XRD patterns of GC samples calcined at 1250 °C for 3 h with additional CaO added. With increase of CaO to glass weight ratios from 10:100 (Fig. 7a), to 20:100 (Fig. 7b) and 35:100 (Fig. 7c), the typical peaks of baddeleyite (2θ = 31.7° and 34.2°) become less intense, suggesting less baddeleyite phase is formed. When CaO to glass weight ratio increases to 50:100 (Fig. 7d), no typical baddeleyite peaks are present. However, strong perovskite peaks (2θ = 33.1° and 47.4°) are present instead. As discussed previously, at 3 h sintering without adding extra CaO (Fig. 4b), the major phase is baddeleyite. The second major phase is zirconolite, which encloses the baddeleyite in most grains. After adding small amounts of CaO (CaO:glass = 10:100 w/w), the major phase is zirconolite with main minor phase as baddeleyite as shown in the SEM image (Fig. 8a). No core-shell microstructured grains are observed which can be seen from the high magnification SEM image (Fig. S6a). Further increasing CaO in glass (weight ratio at 20:100 and 35:100) leads to the enhancement of zirconolite formation and small amounts of baddeleyite (ZrxTi1−xO2 with x ≥ 0.5) observed (Fig. 8b–c and Figs. S6b-c). When CaO to glass weight ratio increases to 50:100, two major phases including zirconolite and perovskite (CaTiO3) are observed, as well as other minor phases such as CaO in glass (Fig. 8d and Fig. S6d). This observation is consistent with XRD results. It is interesting to note that no apparent baddeleyite phase is observed for this sample by both SEM and XRD. It is found that with increase of CaO content in glass, the viscosity of the glass matrix is decreased. For the highest CaO amount
3.5. Discussion For a successful formation of a zirconolite crystal in an amorphous glass matrix, both thermodynamic and kinetic factors should be considered [4,7]. The recent study [4] reveals that in the sintering time range of 30–90 min at 1200 °C, zirconolite is probed to be the only crystalline phase in an alumino-silicate glass with XRD characterization. To form zirconolite via metal oxides: CaO + ZrO2 + 2TiO2 → CaZrTi2O7
(1)
the standard Gibbs free energy (ΔG) for the formation of zirconolite is −92.0 kJ mol–1, suggesting the formation of zirconolite is thermodynamically favorable. The standard molar Gibbs free energies (ΔG°) of CaO, ZrO2, TiO2, CaZrTi2O7 are −603.1, −1042.9, −888.8, −3515.6 kJ mol−1 at 298.15 K, respectively [4,7]. It is reasonable to assume that these data will remain essentially invariant with temperature [7]. However, we believe longer sintering time should be pursued to study the thermodynamic stability of zirconolite in glass. In certain circumstances, kinetics is overwhelmed due to the high variation of solubility of metal oxides in glass at some processing conditions such as longer phase equilibrium time and high fluid of the glass matrix. With increasing sintering time at 1250 °C, more baddeleyite structured material is observed in this work. As a result, the following decomposition reaction may occur: 5
Ceramics International xxx (xxxx) xxx–xxx
L. Kong et al.
Fig. 6. (a) TEM image of glass and B-phase, (b) TEM image of CaZrTi2O7, (c) HRTEM image of CaZrTi2O7 with inset as live fast fouier transform (FFT) of view [001], (d) SAED pattern in [001] zone axis. GC sample was calcined at 1250 °C for 3 h. B-phase is baddeleyite phase (high Ti-bearing zirconia).
CaZrTi2O7 → ZrTiO4 + CaO + TiO2
(2)
the ΔG° of ZrTiO4 is −1935.7 kJ mol−1 at 298.15 K [25]. The standard Gibbs free energy (ΔG) for the equilibrium reaction will be: ΔG = ΔG° + RT ln Qr, i.e. ΔG = 88.0 + RT ln Qr, where Qr is the reaction quotient defined here by Qr = [ZrTiO4][CaO][TiO2]/[CaZrTi2O7], with [CaO] representing the chemical activity of CaO, etc. Even though the ΔG° is not favorable for the decomposition of zirconolite, the changes of the chemical activities of CaO and TiO2 with time play a major role kinetically. There is no sufficient data showing the metal oxide solubility in variety types of glasses. Ojovan and Lee list the approximate solubility limits of elements in silicate glasses [26]. Ca has 15–25 wt% solubility limit whereas Ti and Zr have 5–15 wt%. Our SEM-EDS line scan experiments confirm that significant amounts of CaO and small amounts of TiO2 are dissolved in the residue glass. The major ceramic phase is ZrxTi1−xO2 with x ≥ 0.5 in longer sintering time. As such only trace amounts of Zr is observed in glass. In addition, little sphene (CaTiSiO5) is detected but with perovskite (CaTiO3) instead. With the chemical activity of CaO decreases dramatically due to its
Fig. 7. XRD patterns of GC samples calcined at 1250 °C for 3 h with CaO added in glass. CaO to glass weight ratio is (a) 10:100, (b) 20:100, (c) 35:100, (d) 50:100. All samples were finely ground before analyzing. The main diagnostic peaks for each phase are indicated: zirconolite (Z); baddeleyite structured materials (B); rutile (R); perovskite (P).
6
Ceramics International xxx (xxxx) xxx–xxx
L. Kong et al.
Fig. 8. Backscattered SEM images of GC samples calcined at 1250 °C for 3 h with CaO added in glass formulation. CaO to glass weight ratio is (a) 10:100, (b) 20:100, (c) 35:100, (d) 50:100. Dark grey (phase 1): CaZrTi2O7, light grey (phase 2): high Ti-bearing zirconia, bright grey (phase 3): low Ti-bearing zirconia, dim grey (phase 4): CaTiO3, determined by EDS. Black area is residual glass.
4. Conclusions
high solubility in glass, the reaction shifts to right side with time. In contrary, if extra CaO is added in glass phase, chemical activity of CaO is increased, which leads to the chemical reaction (2) shifting to left side. As a result, zirconolite phase is kinetically stabilized. It has been concluded [24] that the formation of ceramic phases follows the crystallization sequence of Ti-bearing zirconia → ZrTiO4 phase → Zr-rich zirconolite → Zr-poor zirconolite → rutile/perovskite from the melt with a zirconolite composition. This sequence is induced by a fractional crystallization process, in which Zr-rich phases tend to crystallize first, resulting in continuous depletion of Zr in the melt. Consistent with this melt compositional evolution, the Zr content in zirconolite decreases from the area next to the ZrTiO4 phase to areas next to rutile or perovskite. The results in the present work show the opposite crystallization sequence with longer sintering times at fixed temperature in a glass matrix. It is generally viewed that adding nucleating agents like TiO2 and ZrO2 to the glass composition will lead to the formation of ZrTiO4 nuclei, followed by the formation of zirconolite [27]. In addition, it has been reported [28] that baddeleyite (m-ZrO2) crystals form at the expense of zirconolite in the bulk of glass-ceramics heat treated at ≥ 1250 °C. In this study, zirconolite forms directly from (CaZrTi2-oxide) composite at short sintering time (≤ 1 h) as thermodynamically favorable phase, then it changes to mainly baddeleyite phase with time. Even though the thermodynamic data favorites the formation of zirconolite, the baddeleyite becomes the dominating phase kinetically with time due to the Ostwald ripening, and more importantly the high solubility of CaO in the melting glass. Our experimental results are consistent with the literature [4,28], and support the claims from both thermodynamic and kinetic point of views.
The internal crystallization of CaZrTi2O7 zirconolite in a sodium alumino-borosilicate glass has been investigated under pressureless sintering at 1250 °C in air. The ceramic precursor is prepared using a soft chemistry route to ensure ultimate mixing of the reactants in molecular level, thus higher homogeneity of the product. Neither organic solvent nor mechanical milling procedure is used during processing. Ceramic crystalline phase evolution in glass is observed with sintering time, with zirconolite as the dominant phase at short sintering and changes to baddeleyite phases for long sintering time. When sintering at 1250 °C for 3 h, XRD shows that zirconolite is the dominant crystalline phase, however, SEM and EDS line scan results reveal that the dominant ceramic phase is actually baddeleyite (high Ti-bearing zirconia, ZrxTi1−xO2 with x ≥ 0.5), which is enclosed by zirconolite phase forming a core-shell microstructured grains. HRTEM images and SAED patterns confirm that the zirconolite grains possess a high level of crystallographic perfection at the atomic scale suggesting the complete crystallization. These observations lead to the conclusion that both analytical techniques for bulk samples (XRD) and microstructure of micro domains (SEM, SEM-EDS, TEM, SAED) should be utilized in complementary to study the structures and the ceramic phase evolution in glass. In addition, both thermodynamic and kinetic factors should be considered for a successful synthesis of the desired ceramic phase in glass. The addition of CaO with CaO to glass weight ratios at ~ 10–35:100 stabilizes the zirconolite phase kinetically, thus enhances the zirconolite formation.
7
Ceramics International xxx (xxxx) xxx–xxx
L. Kong et al.
Acknowledgments
J. Nucl. Mater. 494 (2017) 29–36. [13] C.-Z. Liao, C. Liu, P.-H. Lee, M.C. Stennett, N.C. Hyatt, K. Shih, Combined quantitative X-ray diffraction, scanning electron microscopy, and transmission electron microscopy investigations of crystal evolution in CaO–Al2O3–SiO2–TiO2–ZrO2–Nd2O3–Na2O system, Cryst. Growth Des. 17 (2017) 1079–1087. [14] C.-Z. Liao, C. Liu, M. Su, K. Shih, Quantification of the partitioning ratio of minor actinide surrogates between zirconolite and glass in glass-ceramic for nuclear waste disposal, Inorg. Chem. 56 (2017) 9913–9921. [15] C.-Z. Liao, K. Shih, W.E. Lee, Crystal structures of Al−Nd codoped zirconolite derived from glass matrix and powder sintering, Inorg. Chem. 54 (2015) 7353–7361. [16] P. Loiseau, D. Caurant, Glass–ceramic nuclear waste forms obtained by crystallization of SiO2–Al2O3–CaO–ZrO2–TiO2 glasses containing lanthanides (Ce, Nd, Eu, Gd, Yb) and actinides (Th): study of the crystallization from the surface, J. Nucl. Mater. 402 (2010) 38–54. [17] C. Martin, I. Ribet, P. Frugier, S. Gin, Alteration kinetics of the glass-ceramic zirconolite and role of the alteration film – comparison with the SON68 glass, J. Nucl. Mater. 366 (2007) 277–287. [18] D. Caurant, I. Bardez, P. Loiseau, Crystallization of CaHf1−xZrxTi2O7 (0 < x < 1) zirconolite in SiO2–Al2O3–CaO–Na2O–TiO2–HfO2–ZrO2–Nd2O3 glasses, J. Mater. Sci. 42 (2007) 10203–10218. [19] D. Caurant, P. Loiseau, I. Bardez, C. Gervais, Effect of Al2O3 concentration on zirconolite (Ca(Zr,Hf)Ti2O7) crystallization in (TiO2,ZrO2,HfO2)-rich SiO2–Al2O3–CaO–Na2O glasses, J. Mater. Sci. 42 (2007) 8558–8570. [20] D. Caurant, O. Majerus, P. Loiseau, I. Bardez, N. Baffier, J.L. Dussossoy, Crystallization of neodymium-rich phases in silicate glasses developed for nuclear waste immobilization, J. Nucl. Mater. 354 (2006) 143–162. [21] P. Loiseau, D. Caurant, N. Baffier, L. Mazerolles, C. Fillet, , Glass–ceramic nuclear waste forms obtained from SiO2–Al2O3–CaO–ZrO2–TiO2 glasses containing lanthanides (Ce, Nd, Eu, Gd, Yb) and actinides (Th): study of internal crystallization, J. Nucl. Mater. 335 (2004) 14–32. [22] P. Loiseau, D. Caurant, O. Majerus, N. Baffier, C. Fillet, , Crystallization study of (TiO2,ZrO2)-rich SiO2-Al2O3-CaO glasses Part II Surface and internal crystallization processes investigated by differential thermal analysis (DTA), J. Mater. Sci. 38 (2003) 853–864. [23] P. Loiseau, D. Caurant, O. Majerus, N. Baffier, C. Fillet, Crystallization study of (TiO2, ZrO2)-rich SiO2-Al2O3-CaO glasses Part I Preparation and characterization of zirconolite-based glass-ceramics, J. Mater. Sci. 38 (2003) 843–852. [24] H. Xu, Y. Wang, Crystallization sequence and microstructure evolution of Synroc samples crystallized from CaZrTi2O7 melts, J. Nucl. Mater. 279 (2000) 100–106. [25] R.L. Putnam, A. Navrotsky, Thermodynamics of formation for zirconolite (CaZrTi2O7) from T = 298.15 K to T = 1500 K, J. Chem. Thermodyn. 31 (1999) 229–243. [26] M.I. Ojovan, W.E. Lee, An Introduction to Nuclear Waste Immobilisation, 2nd ed., Elsevier Ltd, Amsterdam, etc., 2014, p. 251 (Chapt. 17). [27] M. Chavoutier, D. Caurant, O. Majérus, R. Boulesteix, P. Loiseau, C. Jousseaume, E. Brunet, E. Lecomte, Effect of TiO2 content on the crystallization and the color of (ZrO2,TiO2)-doped Li2O–Al2O3–SiO2 glasses, J. Non-Cryst. Solids 384 (2014) 15–24. [28] P. Loiseau, D. Caurant, N. Baffier, L. Mazerolles, C. Fillet, Development of zirconolite-based glass-ceramics for the conditioning of actinides, Mat. Res. Soc. Symp. Proc. 663 (2000) 179–187.
We would like to thank the Nuclear Materials Development and Characterization Platform at ANSTO for providing facility access to carry out the material synthesis and characterizations. Appendix A. Supplementary material Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.ceramint.2018.01.017 References [1] Y. Zhang, D.J. Gregg, L. Kong, M. Jovanovich, G. Triani, Zirconolite glass-ceramics for plutonium immobilization: the effects of processing redox conditions and plutonium waste loadings, J. Nucl. Mater. 490 (2017) 238–241. [2] M.W.A. Stewart, S.A. Moricca, Y. Zhang, R.A. Day, B.D. Begg, C.R. Scales, E.R. Maddrell, J. Hobbs, Glass-Ceramic Waste Forms for Uranium and Plutonium Residues Wastes – 13164, WM2013 Conference, Phoenix USA, 2013. [3] E.R. Vance, S. Moricca, B.D. Begg, M.W.A. Stewart, Y. Zhang, M.L. Carter, Advantages hot isostatically pressed ceramic and glass-ceramic waste forms bring to the immobilization of challenging intermediate- and high-level nuclear wastes, Adv. Sci. Technol. 73 (2010) 130–135. [4] E.R. Maddrell, H.C. Paterson, S.E. May, K.M. Burns, Phase evolution in zirconolite glass-ceramic wasteforms, J. Nucl. Mater. 493 (2017) 380–387. [5] L. Wu, X. Wang, H. Li, Y. Teng, L. Peng, The effects of sulfate content on crystalline phase, microstructure, and chemical durability of zirconolite−barium borosilicate glass-ceramics, J. Nucl. Mater. 478 (2016) 303–309. [6] H. Li, L. Wu, D. Xu, X. Wang, Y. Teng, Y. Li, Structure and chemical durability of barium borosilicate glass-ceramics containing zirconolite and titanite crystalline phases, J. Nucl. Mater. 466 (2015) 484–490. [7] E. Maddrell, S. Thornber, N.C. Hyatt, The influence of glass composition on crystalline phase stability in glass-ceramic wasteforms, J. Nucl. Mater. 456 (2015) 461–466. [8] M. Jafar, S. Phapale, S.N. Achary, R. Mishra, A.K. Tyagi, High-temperature crystallographic and thermodynamic investigations on synthetic zirconolite (CaZrTi2O7), J. Therm. Anal. Calorim. (2017), http://dx.doi.org/10.1007/s10973017-6816-0. [9] M. Jafar, P. Sengupta, S.N. Achary, A.K. Tyagi, Phase evolution and microstructural studies in CaZrTi2O7–Nd2Ti2O7 system, J. Am. Ceram. Soc. 97 (2014) 609–616. [10] M. Jafar, P. Sengupta, S.N. Achary, A.K. Tyagi, Phase evolution and microstructural studies in CaZrTi2O7 (zirconolite)–Sm2Ti2O7 (pyrochlore) system, J. Eur. Ceram. Soc. 34 (2014) 4373–4381. [11] L. Kong, I. Karatchevtseva, Y. Zhang, A new method for production of glassLn2Ti2O7 pyrochlore (Ln = Gd, Tb, Er, Yb), J. Eur. Ceram. Soc. 37 (2017) 4963–4972. [12] L. Kong, Y. Zhang, I. Karatchevtseva, Preparation of Y2Ti2O7 pyrochlore glassceramics as potential waste forms for actinides: the effects of processing conditions,
8