Quantification of the Partitioning Ratio of Minor Actinide Surrogates

Aug 7, 2017 - Guangdong Key Laboratory of Integrated Agro-environmental Pollution Control and Management, Guangdong Institute of Eco-environmental Sci...
1 downloads 10 Views 4MB Size
Download PDF
Article pubs.acs.org/IC

Quantification of the Partitioning Ratio of Minor Actinide Surrogates between Zirconolite and Glass in Glass-Ceramic for Nuclear Waste Disposal Chang-Zhong LIAO,†,‡ Chengshuai LIU,† Minhua SU,§ and Kaimin SHIH*,‡ †

Guangdong Key Laboratory of Integrated Agro-environmental Pollution Control and Management, Guangdong Institute of Eco-environmental Science & Technology, Guangzhou, China ‡ Department of Civil Engineering, The University of Hong Kong, Hong Kong SAR, China § School of Environmental Science and Engineering, Guangzhou University, Guangzhou, China S Supporting Information *

ABSTRACT: Zirconolite-based glass-ceramic is considered a promising wasteform for conditioning minor actinide-rich nuclear wastes. Recent studies on this wasteform have sought to enhance the partitioning ratio (PR) of minor actinides in zirconolite crystal. To optimize the PR in the SiO2−Al2O3−CaO−TiO2−ZrO2 system, a novel conceptual approach, which can be derived from the chemical composition and quantity of zirconolite crystal in glass-ceramic, was introduced based on the results of Rietveld quantitative X-ray diffraction analysis and transmission electron microscopy energy dispersive X-ray spectroscopy. To verify this new conceptual approach, the influences of the crystallization temperature, the concentration of additives, and ionic radii on the PR of various surrogates (Ce, Nd, Gd, and Yb) in zirconolite were examined. The results reveal that the PR of Nd3+ in zirconolite can be as high as 41%, but it decreases as the crystallization temperature increases. The quantities of all phases (including crystalline and amorphous) remained nearly constant when increasing the loading of Nd2O3 in glass-ceramic products crystallized at 1050 °C for 2 h. Correspondingly, the PR of Nd3+ decreases in a linear fashion with the loading contents of Nd2O3. The radius of ions also has a great influence on the PR, and an increase in the ionic radius leads to a decrease in the PR. This new approach will be an important tool to facilitate the exploration of a glass-ceramic matrix for the disposal of minor actinide-rich nuclear wastes.

■

INTRODUCTION Disposal of high-level radioactive waste is a challenging environmental issue throughout the world due to its characteristics of non-biodegradation and radiotoxicity in the biosphere. In high-level radioactive waste, minor actinides (Np, Am, Cm) are long-lived radionuclides and are major contributors to longterm radiotoxicity after a few hundred years of decay.1−3 Therefore, it is highly desirable to immobilize these long-lived radionuclides into durable matrices with long-term performance that can isolate these radionuclides from the living environment for hundreds of thousands of years. Potential matrices include glass, ceramic, and glass-ceramic.4 Glass matrices, such as borosilicate and phosphate glasses, have been used for nuclear waste immobilization on a commercial scale. However, these glasses possess very low solubility of minor actinides.5 Although ceramic matrices present excellent performance in chemical resistance, radiation damage, and loading capacity, the difficulty of preparing single-phase ceramics and the risk of forming water-soluble phases make their use in industry debatable.4 The glass-ceramic matrix offers the possibility of avoiding the drawbacks of both glass and ceramic matrices in isolation and combines their advantages by © 2017 American Chemical Society

means of the double-barrier containment principle to immobilize minor actinides.4 A zirconolite-based glass-ceramic matrix has been demonstrated to be a promising candidate for the disposal of minor actinides,6−12 but the partitioning ratio (PR) of the minor actinide surrogates (Ce, Nd, Eu, Gd, Yb, and Th) incorporated into zirconolite crystals is only 15% to 45%, depending on the ionic radii, when heat-treated at 1200 °C.3 This indicates that considerable amounts of minor actinide surrogates exist in the residual glass. Notice that a better zirconolite-based glass-ceramic matrix would incorporate more minor actinides into zirconolite crystals. Thus, it is essential to improve the PR of the minor actinides in zirconolite crystals in the glass-ceramic matrix. Because the zirconolite phase is designed as a crystalline phase to host minor actinides in the glass-ceramic matrix, enhancing the formation of the zirconolite phase is an alternative strategy to increase the PR of minor actinides in zirconolite crystal. Results reported by Loiseau et al. showed ∼9 to 11 vol % of zirconolite crystal in the bulk sample heat-treated Received: June 4, 2017 Published: August 7, 2017 9913

DOI: 10.1021/acs.inorgchem.7b01425 Inorg. Chem. 2017, 56, 9913−9921

Article

Inorganic Chemistry

Table 1. Chemical Compositions (in Oxide Weight and Molar Percentage) of Parent Glasses with Different Minor Actinides Surrogates (Ce, Nd, Gd, and Yb) sample ID

composition

SiO2

Al2O3

CaO

TiO2

ZrO2

lanthanide oxide

Na2O

GC-Nd2 (Nd2O3) GC-Nd4 (Nd2O3) GC-Nd6 (Nd2O3) GC-Nd8 (Nd2O3) GC-Nd10 (Nd2O3) GF1-Ce (CeO2)

wt % wt % wt % wt % wt % mol % wt % mol % wt % mol % wt % mol % wt %

42.293 41.43 40.567 39.703 38.84 48.226 40.51 48.226 40.566 48.226 40.379 48.226 40.154

12.455 12.201 11.946 11.692 11.438 8.369 11.93 8.369 11.946 8.369 11.891 8.369 11.825

20.466 20.049 19.631 19.213 18.796 25.005 19.604 25.005 19.631 25.005 19.54 25.005 19.431

12.984 12.719 12.454 12.189 11.924 11.138 12.436 11.138 12.454 11.138 12.396 11.138 12.327

8.822 8.642 8.462 8.282 8.102 4.905 8.45 4.905 8.462 4.905 8.423 4.905 8.376

2.00 4.00 6.00 8.00 10.00 2.548 6.131 1.274 6 1.274 6.436 1.274 6.957

0.98 0.96 0.94 0.92 0.90 1.083 0.938 1.083 0.94 1.083 0.935 1.083 0.93

GF2-Nd (Nd2O3) GF3-Gd (Gd2O3) GF4-Yb (Yb2O3)

at 1200 °C, as revealed by scanning electron microscopy (SEM) images.11 Our previous studies showed that the amount of zirconolite was ∼19 wt %, regardless of the crystallization temperature or time.2 It has been reported that changing the contents of CaO, TiO2, and ZrO2 in the parent glass can increase the quantity of zirconolite crystal in the glass-ceramic product.8,13 However, the optimum content of zirconolite crystal in the CaO−SiO2−Al2O3−TiO2−ZrO2 system has not yet been determined. When changing chemical compositions in glass-ceramic products, it is quite important to quantify all fractions (including all crystalline and amorphous/glass phases) and then optimize the parameters to achieve a higher zirconolite content. However, the literature lacks an effective method to achieve this goal when analyzing the zirconolitebased glass-ceramic matrix. Loiseau et al. quantified the amount of zirconolite phase with SEM,8,11,13 which is an indirect method of obtaining the quantities of phases and only reflects the results from small areas. In this study, we introduce a direct method to quantify the composition of all crystalline and amorphous phases in the glass-ceramic matrix via Rietveld quantitative X-ray diffraction analysis. In this method, the relative weight fractions of the crystalline phases can be evaluated from the scaling factors, which can be generated when adjusting the calculated patterns with the observed profiles with whole pattern-fitting routines. Adding a suitable internal standard (such as NIST676a with known crystallinity) with a certain percentage provides an opportunity of observing all the weight fractions of crystalline phases (including the weight fraction of internal standard) derived from Rietveld analysis. If the weight fraction of the internal standard derived from the Rietveld analysis is different from the doped amount, the amount of the amorphous/glass phase(s) in the sample can be thus calculated based on those data.14,15 This technique has been used successfully in many other applications due to its reliability, precision, and reproducibility.16−21 The incorporation amount of minor actinides in zirconolite is another important factor that leads to the different PRs of minor actinides in glass-ceramic products. Generally, energydispersive X-ray spectroscopy (EDX) is used to determine a crystal’s chemical composition. Only Loiseau and coauthors used SEM-EDX to estimate the chemical composition of zirconolite crystals heat-treated at 1200 °C for 2 h.11 However, their studies did not investigate the chemical composition of zirconolite crystals grown at temperatures lower than 1200 °C, because the crystals were too small (less than 1 μm). In

addition, the zirconolite crystals were encapsulated by residual glass in the zirconolite-based glass-ceramic matrix. Using SEMEDX to evaluate the chemical compositions of the crystals inevitably includes the signal reflected from the glass matrix, which can lead to questionable EDX results because the traveling length of the electrons in the materials is usually longer than 1 μm, depending on the energy of the X-ray photons.22 Unlike SEM-EDX, transmission electron microscopy energy-dispersive X-ray spectroscopy (TEM-EDX) can be used to obtain the chemical compositions of areas at the nanometer level, avoiding interference from the other areas. In our previous study, TEM-EDX was used to precisely determine the chemical compositions of zirconolite crystal grown at 1050 °C in a CaO−SiO2−Al2O3−TiO2−ZrO2−Na2O−Nd2O3 system.6 We also successfully used this method to study the solid solution of Mg(Al,Fe,Cr)2O4 spinel crystals in glass-ceramic products.20,21 Although Loiseau et al. evaluated the PR of Ln3+ (Ce3+, Nd3+, Eu3+, Gd3+, and Yb3+) ions incorporated into zirconolite phase using electron paramagnetic resonance (EPR),8,11,13,23 this technique only detected the species with unpaired electrons, so it is not applicable for all species, particularly for actinide ions with paired electrons. In addition, the signals of Ln3+ ions from crystals and residual glass in the zirconolite-based glass-ceramic matrix overlap significantly, leading to great difficulty in accurate quantification. Furthermore, it is also difficult to estimate the PRs with an insufficient amount of crystals.24 To further maximize the PRs of minor actinides incorporated into zirconolite, we introduce a new quantification method to resolve this issue, in which the PRs of minor actinides (surrogates) in zirconolite in the glass-ceramic matrix can be obtained. In this study, the lanthanides (Ce, Nd, Gd, and Yb) were used as nonradioactive actinide surrogates due to their similar ionic radii and oxidation states. TEM-EDX was used to obtain the chemical compositions of nanometer-size zirconolite crystals in the glass-ceramic matrix, in which only the zirconolite crystals located at the edge of the thinned sample not covered by the glass were probed. Rietveld quantitative Xray diffraction analysis (QXRD) with an internal standard (Al 2O 3 , SRM 676a) was used to quantify the phase compositions (including crystalline and amorphous). The PR, which is equal to the molar fraction of lanthanide ions incorporated into the zirconolite phase versus the total amount of simulants, was defined based on the results of TEM-EDX and QXRD. The influences of the crystallization temperature, 9914

DOI: 10.1021/acs.inorgchem.7b01425 Inorg. Chem. 2017, 56, 9913−9921

Article

Inorganic Chemistry

Figure 1. Glass-ceramics prepared after nucleation at 810 °C for 2 h and crystal growth at various temperatures for 2 h with their (a) XRD phase identification results, (b) lattice parameters of zirconolite crystals, (c) phase quantification results derived from Rietveld refinement through Al2O3 internal standard, and (d) partitioning ratios of Nd3+ in zirconolite crystals. was preheated to 775 °C and annealed at this temperature for 2 h to relieve the internal stresses before cooling the sample. X-ray Diffraction Characterization. The X-ray powder diffraction spectra were collected by a D8 Advanced Diffractometer (Bruker AXS) equipped with a Cu Kα radiation source (40 kV, 40 mA) and LynxEye detector. For phase identification, the 2θ scan range was from 10° to 80°, and the step size was 0.02° with a scan speed of 1 s step−1. For phase quantification, the 2θ scan range was from 10° to 110° with a step size of 0.02° and a scan speed of 1.5 s step−1. QXRD analysis was performed with TOPAS 4.2 software (Bruker AXS GmbH, Germany) with a fundamental parameters approach. To quantify the contents of both crystalline and amorphous (glass) phases, all obtained glass-ceramic samples were spiked with 20 wt % of Al2O3 (NIST, SRM 676a) as the standard reference in the QXRD analysis. TEM-EDX and Selected-Area Electron Diffraction (SAED) Characterization. The glass-ceramic was cut into thin sections and further thinned by mechanical grinding. Ion beam milling (Fischione Model 1010 Ion Beam Milling) was used at the last thinning step. TEM-EDX and SAED analysis was performed on an FEI Tecnai G2 20 S-TWIN operated at 200 kV. TEM-EDX was used to determine the chemical compositions of the nanometer-size crystals, where only the crystals at the edge of the sample not covered by glass were analyzed. SEM-EDX Characterization. The samples were ground by decreasing the grain sizes of diamond paste and polished with polishing cloths loaded with lubricant and alumina abrasives. A Hitachi S-3400N with variable pressure was used for SEM-EDX analysis. PR. The PR is defined as the molar fraction of lanthanide ions incorporated into the zirconolite crystal over the total amount of lanthanide ions as shown below: For Ln2O3 (Ln= Nd, Gd, or Yb):

the concentrations of additives, and the ionic radii on the PR of different lanthanides (Ce, Nd, Gd, and Yb) incorporated into the zirconolite were investigated.

■

EXPERIMENTAL SECTION

Glass-Ceramic Sample Preparation. The parent glass was prepared first with a mixture of SiO2−Al2O3−CaO−TiO2−ZrO2−Ln-oxide−Na2O (Ln, lanthanide). The chemical compositions of the starting mixture (Table 1) were chosen based on those described in the literature.8,9 The limitation of Nd2O3 concentration (10 wt %) was set due to the following reasons: (1) to reduce the risk of early mechanical breakage and decomposition of wasteform due to high concentrations of radionuclides;25 (2) to avoid the strong heating in the wasteform.3 The chemical compositions of GC-Nd are the same as those of GC-Nd6. Lanthanides (Ce, Nd, Gd, and Yb) were used as the surrogates for actinides because of their similar radii and valences. To study the ionic radius effect, the same atomic ratio of Si/Al/Ca/Ti/Zr/ Ln/Na (Ln = Ce, Nd, Gd, or Yb) was used in the raw materials. Mixtures of reagent-grade oxides (SiO2, Al2O3, TiO2, ZrO2, Ln-oxides) and carbonates (CaCO3, Na2CO3·10H2O) were melted in a platinum crucible at 1500 °C for 9 h and poured into an alumina crucible in air. To fabricate homogeneous glass, the obtained glass was further ground into powder, remelted at 1500 °C for 5 h, and poured into the alumina crucible in air to form the glass. The cast parent glass was annealed at 775 °C for 2 h to relieve the internal stresses and then cooled to room temperature. The glass-ceramic product was obtained via a two-step heat treatment as described in the literature:11 (1) placing the parent glass into an 810 °C preheated furnace for 2 h of nucleation and (2) transferring the sample to another furnace that was preheated to the target temperatures (900, 950, 1000, 1050, and 1100 °C) and then held for 2 h. Finally, the sample was transferred to a third furnace that 9915

DOI: 10.1021/acs.inorgchem.7b01425 Inorg. Chem. 2017, 56, 9913−9921

Article

Inorganic Chemistry

Figure 2. TEM-EDX results of glass-ceramic sample prepared after nucleation at 810 °C for 2 h and crystal growth at 1050 °C for 2 h. The SAED pattern confirms that Area 1 corresponds to zirconolite crystals. The TEM-EDX spectra reflect the results of the crystals (Area 1) and the residual glass (Area 2).

wt%of Ca xZryTiαLnβAl γO7 × PR =

increased to 950 °C, two crystalline phases, namely, cubiczirconia and zirconolite, were detected in the sample. The zirconolite phase was obtained in the bulk samples crystallized at both 1000 and 1050 °C for 2 h. A trace amount of crystalline titanite phase was also observed in the XRD patterns of these two samples. This crystalline phase has been shown to grow from the sample surface, and Loiseau et al. also noted that the surface crystallization increased dramatically at 1100 °C.12 This phenomenon was also shown in our XRD patterns (Figure 1a), and titanite and anorthite were observed with the zirconolite phase. The intensities of these two phases were also stronger than those in the samples crystallized at 1000 or 1050 °C. By performing Rietveld quantitative XRD with an internal standard, the weight fractions of the components (including crystalline and glass phases) in the glass-ceramic samples were obtained and are shown in Figure 1c. Residual glass was the dominant component in all glass-ceramic samples and still composed ∼72.5 wt % even after crystallization at 1100 °C for 2 h. Although cubic-zirconia was detected by XRD, the amount was less than 1 wt %, and the sample crystallized at 900 °C was nearly all in glass phase. The maximum content of zirconolite phase was ∼19 wt % at 1000 °C. The amounts of both titanite and anorthite increased when the crystallization temperature was increased, because the thickness of the surface crystallization also increased.12

MW of LnβO1.5β MW of CaxZryTiαLnβAl γO7

wt%of Ln2O3 in sample

(1)

For CeO2: wt% of Ca xZryTiαCeβ Al γO7 × PR =

MW of Ceβ O2β MW of CaxZryTiαCeβ Al γO7

wt% of CeO2 in sample

(2)

The chemical formula of CaxZryTiαLnβAlγO7 or CaxZryTiαCeβAlγO7 (zirconolite phase) is determined by TEM-EDX, and the wt % of CaxZryTiαLnβAlγO7 or CaxZryTiαCeβAlγO7 (zirconolite phase) is determined by the results of XRD quantification. MW is the molecular weight. A PR of 100% denotes the complete incorporation of Ln/Ce into the zirconolite phase, whereas 0% means that no Ln/Ce was incorporated into the zirconolite crystals.

■

RESULTS AND DISCUSSION Effect of Temperature. The GC-Nd6 parent glass was chosen to study the temperature effect on the PR of radionuclides. After nucleation at 810 °C for 2 h, the samples were crystallized at various temperatures (900, 950, 1000, 1050, and 1100 °C) with a constant dwell time (2 h). The phase compositions of the obtained glass-ceramic were revealed by powder XRD patterns in Figure1a. Only cubic-zirconia was observed in the sample crystallized at 900 °C, and its intensity was very low. When the crystallization temperature was 9916

DOI: 10.1021/acs.inorgchem.7b01425 Inorg. Chem. 2017, 56, 9913−9921

Article

Inorganic Chemistry

Table 2. TEM-EDX Results of Zirconolite Crystals in the Samples (GC-Nd6: Nd2O3) Prepared after Nucleation at 810 °C for 2 h and Crystal Growth at Different Temperatures for 2 h temperature

Al (atomic %)

Ca (atomic %)

Nd (atomic %)

Ti (atomic %)

Zr (atomic %)

1000 °C 1050 °C 1100 °C

2.70 (0.70) 2.42(0.41) 2.79(0.51)

20.42(3.25) 20.55(0.65) 18.73(0.86)

6.98(0.57) 7.07(0.53) 6.15(0.56)

46.18(1.69) 46.21(0.68) 44.45(0.29)

22.58(3.8) 23.75(1.14) 27.88(1.16)

Figure 3. Results of glass-ceramics with various weight percentages of Nd2O3 (2, 4, 6, 8, and 10 wt %), prepared after nucleation at 810 °C for 2 h and crystal growth at 1050 °C for 2 h. The results show (a) the identified phases by XRD, (b) lattice parameters of zirconolite crystals, (c) phase quantification results derived from Rietveld refinement through Al2O3 internal standard, and (d) partitioning ratios of Nd3+ in zirconolite crystals.

TEM-EDX was used to obtain more precise chemical compositions of the zirconolite crystal. The SAED pattern in the inset of Figure 2 indicates that Area 1 is of zirconolite crystal. The chemical composition of the zirconolite crystal was obtained from the crystals at the edge of the thinned sample without being covered by glass. This strategy can avoid the interference of signal from the glass. As shown in Figure 2, the EDX spectra indicate only Ca, Al, Ti, Zr, Nd, and O in Area 1, whereas Si, Ca, Al, Ti, Zr, Nd, and O were detected in Area 2. Loiseau et al. indicated that Si can be excluded from the zirconolite crystals grown from a CaO−Al2O3−SiO2−ZrO2− TiO2−Nd2O3−Na2O glass matrix.11 Therefore, the results above confirm that Area 1 has only zirconolite crystal, whereas Area 2 has the residual glass. The chemical compositions of zirconolite crystals grown from the same glass matrix at various temperatures are summarized in Table 2. These results indicate the same chemical compositions for zirconolite crystals even when the crystallization temperature was increased. The lattice parameters of the zirconolite phase calculated from the XRD patterns also support this result. Figure 1b shows that the lattice parameters of the zirconolite phase such as a, b, c, and volume changed only slightly when the temperature was increased from 1000 to 1100 °C. These results indicate that the chemical composition and the amount of incorporated actinides in a

zirconolite crystal may not be affected by the crystallization temperature. On the basis of the QXRD and TEM-EDX results, the PR of the Nd into zirconolite phase in the glass-ceramic system can be calculated with eq 1 and is shown in Figure 1d. The PR of Nd into the zirconolite phase in the sample crystallized at 1000 °C was ∼41%. It then decreased as the crystallization temperature was increased, because the amounts of the zirconolite phase decreased as the temperature increased (Figure 1c). This result suggests that to increase the PR of radionuclides into a target crystalline phase (such as zirconolite), it is preferable to first increase the overall amounts of that target phase in the products. Effect of Dopant Concentration. Different weight percentages of Nd2O3 (2, 4, 6, 8, 10 wt %) were used to simulate the loadings of radionuclides in the glass-ceramic matrix. The 10 wt % is nearly the limit of minor actinide loading in the hosting matrix, due to the strong heating in the container during the α-decay of minor actinides.3 All parent glasses were heat-treated under the same conditions: nucleation at 810 °C for 2 h and crystallization at 1050 °C for 2 h. The XRD patterns of the obtained glass-ceramic with different concentrations of Nd2O3 are shown in Figure 3a. All XRD patterns are similar and show the dominant zirconolite phase, together with two other minor crystalline phases (titanite and 9917

DOI: 10.1021/acs.inorgchem.7b01425 Inorg. Chem. 2017, 56, 9913−9921

Article

Inorganic Chemistry

Table 3. TEM-EDX Results of Zirconolite Crystals in the Samples with Different Weight Percentages of Nd2O3 (2, 4, 6, 8, and 10 wt %), Prepared after Nucleation at 810 °C for 2 h and Crystal Growth at 1050 °C for 2 h sample ID

Al (atomic %)

Ca (atomic %)

Nd (atomic %)

Ti (atomic %)

Zr (atomic %)

Nd/Al ratio

GC-Nd2 GC-Nd4 GC-Nd6 GC-Nd8 GC-Nd10

1.62(0.20) 2.25(0.27) 2.42(0.41) 2.33(0.33) 2.73(0.67)

22.64(0.69) 20.34(1.63) 20.55(0.65) 19.47(0.57) 18.46(0.30)

2.77(0.17) 4.95(0.30) 7.07(0.53) 8.59(0.50) 10.91(0.94)

48.42(0.76) 48.62(1.49) 46.21(0.68) 46.31(0.69) 43.88(0.47)

24.56(0.36) 23.84(1.55) 23.75(1.14) 23.31(0.30) 24.03(0.64)

1.71 2.20 2.92 3.69 4.00

Figure 4. Line-scan profiles of Al, Ca, Nd, Si, Ti, and Zr elements with (a) TEM-EDX mode and (b) SEM-EDX mode taken across zirconolite crystals embedded in the residual glass.

shows that the Nd/Al ratio in the zirconolite crystal increases as the Nd2O3 in the system increases. Meanwhile, as the Nd2O3 increased, there was a decrease of Ca in the zirconolite crystals. Because Al3+ is used for the charge compensation of Nd3+ replacing Ca2+ and Zr4+ in the zirconolite phase,26 the result of a Nd/Al ratio greater than 1 further confirms the coexistence of two charge compensation mechanisms in zirconolite: (Nd3+, Al3+) ↔ (Ca2+, Ti4+) and (Nd3+, Nd3+) ↔ (Ca2+, Zr4+).27 In Figure 3b, the lattice parameters of zirconolite increased with the increase in the Nd2O3 content. Because the ionic radii of Nd3+ (r = 1.109 Å) were similar to those of Ca2+ (r = 1.12 Å) but larger than those of Zr4+ (r = 0.84 Å),28 the increase in both the lattice parameters and the Nd/Al ratio in the zirconolite phase reflected the increase in Nd3+ at Zr sites. Using the results of QXRD and TEM-EDX, the PR of Nd into zirconolite phase in the glass-ceramic was calculated according to eq 1 and is shown in Figure 3d. The PR of Nd had a linear relationship with the Nd2O3 content in the glass matrix and showed a decreasing trend as the Nd2O3 increased. With loading 2 wt % of Nd2O3, the PR was ∼42%, but this value decreased to ∼32% when Nd2O3 was added up to 10 wt %. This outcome is consistent with that reported by Loiseau et al., who used EPR to investigate the PR.23 Note that EPR only can detect species with unpaired electrons and that a significant overlap of EPR signals of Ln3+ exists in crystals and residual glass. The new approach developed in our study can detect all species, including nanometer-scale crystals, through the entire XRD pattern to minimize statistical errors. Therefore, the current method is much more practicable and precise than the EPR method. In addition, the results of PR with different

anorthite). To evaluate the effects of the Nd2O3 concentration, the quantitative analysis of phases in the glass-ceramic was determined by XRD data with the Rietveld refinement method. Figure 3c provides the phase compositions of amorphous (glass), zirconolite, titanite, and anorthite phases. Glass was the major component in all samples and made up appropriately 80 wt % regardless of the Nd2O3 concentration. The amount of zirconolite, which is the target crystalline phase to incorporate the surrogate (Nd) of minor actinides, did not change with an increase in Nd2O3. Because titanite and anorthite phases were reported to grow from the surface of glass, their amounts reflect the depth of surface crystallization and were not affected by the incorporated Nd2O3. TEM-EDX was performed to obtain the elemental compositions of the zirconolite crystals grown from glass matrix with different Nd 2 O 3 contents. The elemental compositions of zirconolite crystals and residual glass in the glass-ceramic samples are summarized in Table 3 and Table S1 (Supporting Information). The atomic percentage of Nd in the zirconolite phase was found to increase with the increase of Nd2O3 in the glass matrix. The Nd, Ti, and Zr were enriched in the zirconolite phase, whereas Al was concentrated in the residual glass. In Figure 4, the line-scan profiles of both the TEM-EDX and SEM-EDX modes were taken across zirconolite crystal embedded in the residual glass. The amounts of Si, Al, Nd, Ti, and Zr in zirconolite and residual glass were clearly different in the TEM-EDX result but were not much different as observed by SEM-EDX. This finding further confirms that TEM-EDX is more suitable to reveal the components of nanometer-size crystals in glass-ceramic products. Table 3 also 9918

DOI: 10.1021/acs.inorgchem.7b01425 Inorg. Chem. 2017, 56, 9913−9921

Article

Inorganic Chemistry

Figure 5. Results of glass-ceramics incorporated with different lanthanide oxides (GF1-Ce:CeO2, GF2-Nd:Nd2O3, GF3-Gd:Gd2O3, and GF4Yb:Yb2O3) and prepared after nucleation at 810 °C for 2 h and crystal growth at 1050 °C for 2 h. The results show the (a) identified phases by XRD, (b) lattice parameters of zirconolite crystals, (c) phase quantification results derived from Rietveld refinement through the Al2O3 internal standard, and (d) partitioning ratios of different ions (Ce4+, Nd3+, Gd3+, Yb3+) in zirconolite crystals.

Table 4. TEM-EDX Resultsa of Zirconolite Crystals in the Samples with Different Lanthanide Oxides (GF1-Ce:CeO2, GF2Nd:Nd2O3, GF3-Gd:Gd2O3, and GF4-Yb:Yb2O3), Prepared after Nucleation at 810 °C for 2 h and Crystal Growth at 1050 °C for 2 h

a

sample ID

Al (atomic %)

Ca (atomic %)

Ti (atomic %)

Zr (atomic %)

REE (atomic %)

GF1-Ce GF2-Nd GF3-Gd GF4-Yb

2.63(0.51) 2.42(0.41) 3.20(0.63) -

22.35(0.70) 20.55(0.65) 17.50(0.90) 18.08(1.27)

46.01(1.19) 46.21(0.68) 44.44(0.82) 45.75(0.64)

23.28(0.32) 23.75(1.14) 25.55(0.49) 24.24(0.52)

5.72(0.29) 7.07(0.53) 9.32(0.28) 11.93(1.05)

“-” represents not available. REE represents Ce, Nd, Gd, or Yb in the corresponding sample.

anorthite. The weight fractions of the phase components (including crystalline and amorphous/glass phases) in the samples were obtained by QXRD and are shown in Figure 5c. Glass was the predominant component in all samples and made up more than 73 wt %, depending on the ionic radii of lanthanides incorporated. The amount of zirconolite phase in the glass-ceramic samples was also affected by the ionic radius and decreased as the ionic radius increased. The quantities of both titanite and anorthite were ∼1 wt %, which means that the depth of surface crystallization was very limited. TEM-EDX was also performed for the zirconolite and residual glass in the samples with different lanthanides (Ce, Nd, Gd, and Yb), and the results are summarized in Table 4 and Table S2 (Supporting Information). In all of the glass-ceramic samples, the lanthanides (Ce, Nd, Gd, and Yb) were enriched in the zirconolite phase, whereas Al was concentrated in the residual glass. The distributions of both Ti and Zr differed greatly in the zirconolite and residual glass. In contrast, little difference was found for Ca in the zirconolite and residual glass.

dopant concentrations also indicate that greater loading of radionuclides may lead to more of them in the residual glass. This is not ideal for nuclear waste immobilization, because the loading capacity is one of the most concerning factors. Overall, zirconolite was found to enrich Nd, and the loading of Nd in each zirconolite increased with the increase of Nd2O3 in the glass matrix. To increase the PR of radionuclides in the target crystalline phase, it is beneficial to enhance the formation of the crystalline phase from the glass matrix. Effect of Ionic Radius. Different lanthanides (Ce, Nd, Gd, Yb) with various radii were used to observe their PRs. Figure 5a shows the XRD patterns of the obtained glass-ceramic with different lanthanides (Ce, Nd, Gd, Yb). All samples were prepared with a two-step crystallization route: 810 °C nucleation for 2 h and 1050 °C crystal growth for 2 h. These XRD patterns have similar diffraction peaks and intensities. The phase identification results show three crystalline phases in the samples. Zirconolite was the dominant crystalline phase and was accompanied by two minor phases, namely, titanite and 9919

DOI: 10.1021/acs.inorgchem.7b01425 Inorg. Chem. 2017, 56, 9913−9921

Article

Inorganic Chemistry

zirconolite crystal in the Yb2O3 doping glass-ceramic system. The overall results suggest this new characterization approach can effectively assist the design of optimized chemical compositions and processing parameters to obtain the higher incorporation efficiencies of minor actinides to the targeted durable phases.

The Si was excluded in the zirconolite phase and existed in the residual glass. It is interesting that Al could not be detected in the zirconolite crystals of the Yb-doped glass-ceramic sample. This finding can be explained by the TEM-EDX spectra in Figure S1 (Supporting Information). With the coexistence of Al and Yb, the binding peak shifted to the lower energy side (EDX spectra for area 2 in Figure S1), compared to the existence of Yb without Al (EDX spectra for areas 1 and 3 in Figure S1). In addition, the smaller ionic radius of Yb3+ may have led to the exclusion of Al in the zirconolite structure. The lattice parameters (a and b) of the Yb-doped zirconolite phase are much smaller than those of Ce-doped, Nd-doped, and Gddoped zirconolite phases (Figure 5b). The smaller lattice parameters may have led to the contraction of the a and b planes and the cell volume in the zirconolite structure, which is likely due to the difficulty of incorporating Al into the zirconolite. The QXRD and TEM-EDX results show that the PRs of different lanthanides incorporated into the zirconolite phase in glass-ceramic products can be calculated according to eqs 1 and 2. Figure 5d clearly shows that PR has a significant relationship with the ionic radius and decreases as the ionic radius increases. This trend is similar to the results reported by Caurant et al., who used EPR to evaluate the PR in samples crystallized at 1200 °C.3 However, the PR values in our study are in general higher than those reported in the literature. For instance, the PR for Yb3+ was ∼82% in our work but was reported as ∼36% by Loiseau et al.11 For the largest ion, Ce4+, our result (∼32%) is more than double that (13%) reported in the former study.11 These differences may be due to the different amounts of zirconolite phase in the corresponding samples. The results from Figure 1c reveal that the amount of zirconolite decreases as the temperature increases. Accordingly, the amount of zirconolite crystallized at 1050 °C was much higher than the amount crystallized at 1200 °C.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.7b01425. Two tables present the TEM-EDX results of residual glass in the glass-ceramic samples, and one figure shows the TEM-EDX spectra of zirconolite crystal and residual glass in the Yb-doped glass-ceramic (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Phone: +852 28591973. Fax: +852 25595337. ORCID

Kaimin SHIH: 0000-0002-6461-3207 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors would like to thank F. Y. F. Chan for his help with the TEM experiments and his advice regarding the analysis. This study was funded by the Research Grants Council of Hong Kong (Project Nos. 17212015, C7044-14G, and T21771/16R), and GDAS’ Special Project of Science and Technology Development (2017GDASCX-0834).

■

■

CONCLUSIONS In this study, zirconolite-based glass-ceramic was successfully synthesized by SiO2−Al2O3−CaO−TiO2−ZrO2−Na2O−Ln− Oxide (Ln = Ce, Nd, Gd, Yb) system through a two-step heat treatment. Zirconolite is the only crystalline phase in the glassceramic products, except trace amounts of anorthite and titanite phases occurring on the product surfaces. With QXRD and TEM-EDX, a new and reliable approach to quantify the partitioning ratio (PR) of minor actinides surrogates into zirconolite crystal in the glass-ceramic product was presented in this work. TEM-EDX results indicate no considerable change in the chemical compositions of zirconolite crystals even when the crystallization temperature was increased. The PR of Nd3+ in the zirconolite was up to 41% but decreased at higher temperatures, due to the decomposition of zirconolite. QXRD results reveal the quantities of all phases (including crystalline and amorphous) nearly kept constant at different Nd2O3 loadings when treated at 1050 °C for 2 h. The PR of Nd3+ into zirconolite decreased linearly with the increase of Nd2O3, but the total Nd3+ incorporated into zirconolite phase per formula unit increased with the increase of Nd2O3. The increase of Nd/Al ratio in the zirconolite at higher Nd2O3 loadings reflects the coexistence of two charge compensation mechanisms in zirconolite, that is, (Nd3+, Al3+) ↔ (Ca2+, Ti4+) and (Nd3+, Nd3+) ↔ (Ca2+, Zr4+). Increasing ionic radii of the surrogates led to a decrease of PR from 82% for Yb to 32% for Ce. It is also noticed that Al3+ does not incorporate into

REFERENCES

(1) Salvatores, M. Nuclear Fuel Cycle Strategies Including Partitioning and Transmutation. Nucl. Eng. Des. 2005, 235, 805−816. (2) Liao, C.-Z.; Liu, C.; Lee, P.-H.; Stennett, M. C.; Hyatt, N. C.; Shih, K. 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. 2017, 17, 1079−1087. (3) Caurant, D.; Loiseau, P.; Majerus, O.; Aubin Chevaldonnet, V.; Bardez, I.; Quintas, A. Glasses, Glass-Ceramics and Ceramics for Immobilization of Highly Radioactive Nuclear Wastes; Nova Science Publishers: New York, 2009. (4) Lee, W. E.; Ojovan, M. I.; Stennett, M. C.; Hyatt, N. C. Immobilisation of Radioactive Waste in Glasses, Glass Composite Materials and Ceramics. Adv. Appl. Ceram. 2006, 105, 3−12. (5) Ojovan, M. I.; Lee, W. E. New Developments in Glassy Nuclear Wasteforms; Nova Publishers: New York, 2007. (6) Liao, C.-Z.; Shih, K.; Lee, W. E. Crystal Structures of Al−Nd Codoped Zirconolite Derived from Glass Matrix and Powder Sintering. Inorg. Chem. 2015, 54, 7353−7361. (7) Loiseau, P.; Caurant, D.; Baffier, N.; Mazerolles, L.; Fillet, C. Development of Zirconolite-Based Glass-Ceramics for the Conditioning of Actinides. MRS Online Proc. Libr. 2000, 663, 179−187. (8) Loiseau, P.; Caurant, D.; Majerus, O.; Baffier, N.; Fillet, C. Crystallization Study of (TiO2, ZrO2)-Rich SiO2-Al2O3-CaO Glasses. Part I. Preparation and Characterization of Zirconolite-Based GlassCeramics. J. Mater. Sci. 2003, 38, 843−852. (9) Loiseau, P.; Caurant, D.; Majerus, O.; Baffier, N.; Fillet, C. Crystallization Study of (TiO2, ZrO2)-Rich SiO2-Al2O3-CaO Glasses 9920

DOI: 10.1021/acs.inorgchem.7b01425 Inorg. Chem. 2017, 56, 9913−9921

Article

Inorganic Chemistry Part II Surface and Internal Crystallization Processes Investigated by Differential Thermal Analysis (DTA). J. Mater. Sci. 2003, 38, 853−864. (10) Loiseau, P.; Caurant, D.; Majerus, O.; Baffier, N.; Fillet, C. Competition Between Internal and Surface Crystallization in GlassCeramics Developed for Actinides Immobilization. MRS Online Proc. Libr. 2003, 807, 333−338. (11) Loiseau, P.; Caurant, D.; Baffier, N.; Mazerolles, L.; Fillet, C. 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. 2004, 335, 14−32. (12) Loiseau, P.; Caurant, D. Glass−ceramic Nuclear Waste Forms Obtained by Crystallization of SiO 2−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. 2010, 402, 38−54. (13) Loiseau, P.; Caurant, D.; Bardez, I.; Majerus, O.; Baffier, N.; Fillet, C. Zirconolite-Based Glass-Ceramics for Actinides Immobilization: Effects of Glass Composition and of Actinides Simulant Nature. MRS Online Proc. Libr. 2002, 757, 281−287. (14) De La Torre, A. G.; Bruque, S.; Aranda, M. A. G. Rietveld Quantitative Amorphous Content Analysis. J. Appl. Crystallogr. 2001, 34, 196−202. (15) Orlhac, X.; Fillet, C.; Deniard, P.; Dulac, A. M.; Brec, R. Determination of the Crystallized Fractions of a Largely Amorphous Multiphase Material by the Rietveld Method. J. Appl. Crystallogr. 2001, 34, 114−118. (16) Bernardo, E.; Esposito, L.; Rambaldi, E.; Tucci, A.; Hreglich, S. Recycle of Waste Glass into “glass-Ceramic Stoneware,. J. Am. Ceram. Soc. 2008, 91, 2156−2162. (17) Goel, A.; Pascual, M. J.; Ferreira, J. M. F. Stable Glass-Ceramic Sealants for Solid Oxide Fuel Cells: Influence of Bi2O3 Doping. Int. J. Hydrogen Energy 2010, 35, 6911−6923. (18) Serbena, F. C.; Soares, V. O.; Peitl, O.; Pinto, H.; Muccillo, R.; Zanotto, E. D. Internal Residual Stresses in Sintered and Commercial Low Expansion Li2O-Al2O3-SiO2 Glass-Ceramics. J. Am. Ceram. Soc. 2011, 94, 1206−1214. (19) Dittmer, M.; Ritzberger, C.; Schweiger, M.; Rheinberger, V.; Wörle, M.; Höland, W. Phase and Microstructure Formation and Their Influence on the Strength of Two Types of Glass-Ceramics. J. Non-Cryst. Solids 2014, 384, 55−60. (20) Liao, C.; Tang, Y.; Liu, C.; Shih, K.; Li, F. Double-Barrier Mechanism for Chromium Immobilization: A Quantitative Study of Crystallization and Leachability. J. Hazard. Mater. 2016, 311, 246− 253. (21) Liao, C. Z.; Tang, Y.; Lee, P. H.; Liu, C.; Shih, K.; Li, F. Detoxification and Immobilization of Chromite Ore Processing Residue in Spinel-Based Glass-Ceramic. J. Hazard. Mater. 2017, 321, 449−455. (22) Goldstein, J. I.; Newbury, D. E.; Echlin, P.; Joy, D. C.; Romig, A. D., Jr.; Lyman, C. E.; Fiori, C.; Lifshin, E. Scanning Electron Microscopy and X-Ray Microanalysis: A Text for Biologists, Materials Scientists, and Geologists; Springer Science & Business Media, 2012. (23) Loiseau, P.; Caurant, D.; Baffier, N.; Fillet, C. Neodymium Incorporation in Zirconolite-Based Glass-Ceramics. MRS Online Proc. Libr. 2000, 663, 169−177. (24) Caurant, D.; Bardez, I.; Loiseau, P. Crystallization of CaHf1−xZrxTi2O7 (0 ≤ x ≤ 1) Zirconolite in SiO2−Al2O3−CaO− Na2O−TiO2−HfO2−ZrO2−Nd2O3 Glasses. J. Mater. Sci. 2007, 42, 10203−10218. (25) Ojovan, M. I.; Burakov, B. E.; Lee, W. E. Destruction of MicroCrystal Containing Wasteforms by Charge-Induced Crystal Shape Change on Self-Irradiation. MRS Adv. 2017, 2, 621−626. (26) Loiseau, P.; Caurant, D.; Baffier, N.; Fillet, C. Structural Characterization of Polycrystalline (Nd,Al)-Substituted Zirconolite. MRS Online Proc. Libr. 2002, 757, 243−250. (27) Loiseau, P.; Caurant, D.; Baffier, N.; Fillet, C. Lanthanide and Actinide Incorporation in Zirconolite Based Glass Ceramics. Phys. Chem. Glas. 2002, 43, 201−206.

(28) Shannon, R. D. Revised Effective Ionic Radii and Systematic Studies of Interatomic Distances in Halides and Chalcogenides. Acta Crystallogr., Sect. A: Cryst. Phys., Diffr., Theor. Gen. Crystallogr. 1976, 32, 751−767.

9921

DOI: 10.1021/acs.inorgchem.7b01425 Inorg. Chem. 2017, 56, 9913−9921