Natural magmatic granite as matrix for immobilizing simulated An4+ waste

Lingshuang Li Southwest University of Science and Technology Xiaoyan Shu Southwest University of Science and Technology Hexi Tang Southwest University of Science and Technology Shunzhang Chen Sichuan University Wenxiao Huang Ministry of Land and Resources Guilin Wei Southwest University of Science and Technology Dadong Shao Nanjing University of Science and Technology Yi Xie University of Science and Technology of China Xirui Lu (  luxiruimvp116@163.com ) Southwest University of Science and Technology https://orcid.org/0000-0003-4751-8408


Introduction
With the development of nuclear power and national defense industry, the amount of radioactive waste increased. Actinides in high-level waste (HLW) presents high speci c activity, long half-life and high toxicity. If not properly handled, it will seriously endanger the living environment of human beings. At present, deep geological disposal has been widely recognized [1][2][3] . To ensure long-term safety during disposal, one of the keys is to immobilize waste in a stable matrix to prevent nuclides migration or leakage. As the rst barrier between high-level radioactive waste and biosphere, the safety and stability of the main matrix in a long time are particularly important.
Hatch et al. [4] rst proposed minerals to x radionuclides so that nuclear waste should return to nature like natural radionuclides, and to maintain long-term safety and stability. Now, Synroc is known as the second generation of matrix for immobilizing high-level radioactive waste. It has the characteristics of strong inclusiveness, high water resistance, low nuclides leaching rate, considerable thermal stability and good radiation resistance [5][6][7][8] . Up to now, high-alunite phases, titanium-based phases, silicon-based ore phases and its composite phase are recognized as candidate solidi ed bodies for long-lived radioactive waste [9][10][11][12][13] . With the aim of solidifying radioactive waste to nature, most of these researches concerned Synroc. Until 2007, B. I. Omel'yanenko et al. [14] proved the feasibility of using natural minerals to return radioactive waste to nature by studying both natural minerals and Synroc. However, a large amount of work, from material selection to performance evaluation, is still required.
This work chose natural magmatic granite as host matrix to simulated An 4+ waste, since the rock integrity of granite is highly recognized by geologists all over the world. Granite is mainly composed of quartz, plagioclase, K-feldspar, biotite and hornblende etc. [15] . And it has the characteristics of small porosity, low water content, poor water permeability, large elongation, and good stability, which contributes to prevent or delay the migration of radionuclides [16] . In addition, granite belongs to a glassceramic matrix, which has the advantages of both glass and ceramics. The basic principle shows in Fig. 1. In this experiment, blank granite powder was sintered to nd a suitable sintering temperature, and then Ce 4+ was doped as simulated tetravalent actinide oxides to study the solidi cation behavior of natural magmatic granite [17][18][19] . The phase, micro-morphology and mechanical properties of the immobilization were characterized.

Fabrication
The rock sample was collected at Tong'an in northeast of Guangxi. Zhu et al. [20] determined the geological age of the Tong'an quartz monzonite as 160 ± 4 Ma. The collected rock was mechanical broken up. Then the broken rock was fully washed with water in an ultrasonic cleaner and alcohol in turn to remove surface sludge. The clean rock was dried at 90 ℃ for 12 h in a drier and then crushed a high speed universal crusher (TASITE FW100, Tianjin ) to produce powder through 200 mesh. The crushed sample was subjected to further re ning by grinding with adding alcohol. The X-ray Fluorescence (XRF) result of the pure granite power was listed in Table 1, it presented characteristic of higher alumina element content. The powder was dried again and compressed into tablets (φ = 13 mm) under 16 MPa. The tablets were sintered in a mu e furnace at 800 ℃, 900 ℃, 1000 ℃, 1100 ℃, 1200 ℃, 1300 ℃ and 1400 ℃ respectively for 60 minutes. Sintered temperature was controlled by heating program, which was set to a heating rate of 5 ℃/min. After being kept at the highest temperature for 60 min, the sintered samples were naturally cooled down to room temperature. The whole sintered process was conducted in air atmosphere.
The rock and CeO 2 powder (Tianjin Kermel Co. Ltd., purity ≥ 99.99%) were weighed according to the content of CeO 2 was 5 wt.%, 10 wt.%, 15 wt.% and 20 wt.% respectively. The mixed powder was further ground and compressed into tablets under a pressure of 14 MPa. Then the samples were sintered in a mu e furnace at 1300 ℃ for 60 minutes and naturally cooled down to room temperature. The whole experiment process is as shown in Fig. 2. The morphology did not change signi cantly until the sintering temperature reached 1100 ℃. Melting was observed beyond 1200 ℃.

Characterization
The composition of the rock powder was analyzed by X-ray uorescence (XRF Axios, Netherlands θ/2θ, 2.4 kW). The results are showed in Table 1. To identify the phase structure of the samples, X-ray diffractometer (XRD D/MAX-1400, Rigaku Corporation) with Cu Kα radiation was applied with scanning range of 10 ° to 80 ° at scanning speed of 2 °/min. To observe the micro-morphology of the samples, scanning electron microscopy coupled with energy dispersive X-ray spectroscopy (SEM-EDS, TM4000, Hitachi, Japan) was used. In addition, the samples were also characterized by Raman spectroscopy (InVia laser Raman spectrometer manufactured by Renishaw Company, UK) to study molecular structure.
Specimens were carefully polished to get a at indentation face for the Vickers hardness test (TMVS-1S, China). 4.9 N was applied for indentation for three time on each sample to calculate the average value.

Results And Discussion
3.1 Phase analysis through XRD 3.1.1 Effect of temperature on the structure of pure granite Fig. 3 (a) shows XRD patterns of virgin and sintered pure granite samples obtained below 1200 ℃. There are mainly quartz, feldspar, and a little amount of chromite in the virgin granite. As the sintering temperature ranged from 800 ℃ to 1200 ℃, all the diffraction peaks existed with varied intensity. The intensity of quartz and feldspar gradually weakened with enhanced temperatures, while the intensity of chromite became gradually strong. This is due to the decomposition of feldspar as the temperature rose and the quartz gradually vitri ed [21] . Fig. 3 (b) presents the pure granite sintered at 1300 ℃ and 1400 ℃. Feldspar and quartz almost disappeared after heat treated beyond 1300 ℃, and chromite became the main phase. In addition, diffuse scattering peak ranging from 15 ° to 33 ° emerged, re ecting the amorphous phase of glass structure. Co-existence of amorphous and chromite phase indicated a glassceramic matrix of sintered samples. Moreover, a more disordered glass network structure under increased temperatures was disclosed from the weakening and broadening of diffuse scattering peak [22] . Fig. 4 presents the crystallinity of the pure granite sintered at different temperatures. The sample sintered at 800 ℃ showed the highest crystallinity of about 80 %. As the temperature increased, the crystallinity of the sample decreases gradually. At 1300 ℃, the sample exhibited the minimum crystallinity of about 52 %. Then, the crystallinity turned to increase with sintering temperature by improving the proportion of chromite. According to Fig. 3, the crystallinity reduced due to the phase transformation of granite from crystal to an amorphous state. At 1300 ℃, there is a obvious characteristic peak of glassy amorphization, and the crystallinity reached the lowest level. Subsequently, the intensity of diffuse scattering peak reduced, the crystallinity of the sample increased.

Effect of the CeO 2 content
According to the XRD results obtained from blank granite, magmatic rocks was selected as host material to immobilize simulated tetravalent actinides at 1300 ℃ for 60 minutes. Figure 5 shows the XRD patterns of samples doped with 0 ~ 20 wt.% of CeO 2 . When the content is 5 wt.%, no CeO 2 related peaks can be detected. As the doping amount reached 10 wt.%, some peaks related to CeO 2 appeared on the XRD pattern, indicating that granite cannot immobilize CeO 2 over this content. It re ects that the ultimate solubility of the simulated tetravalent actinide oxides in granite was between 5 ~ 10 wt.%. In order to further study the ultimate solubility, CeO 2 was doped into pure granite powder with the content of 6 wt.%, 7 wt.%, 8 wt.%, and 9 wt.%, and sintered at 1300 ℃ for 60 min. Figure 6 presents the XRD results of the samples with the doping amounts of CeO 2 from 5 wt.% to 10 wt.%. As the doping content of CeO 2 below 8 wt.%, only the diffuse scattering peak and the diffraction peaks of chromite were detected, while the diffraction peaks related to CeO 2 are absent. When the amount of CeO 2 reach 9 wt.%, the diffraction peaks of CeO 2 were observed. It indicates that the ultimate solubility of simulated tetravalent actinide oxides in granite is about 8 wt.%. In the solidi ed body, amorphous phase exists together with chromite crystals, which suggests a glass-ceramic structure. Fig. 7 shows the Raman spectra (300~900 cm -1 ) of fabricated samples holding different contents of CeO 2 . All the samples with different doping amount of CeO 2 present a high intensity peak at around 680 cm -1 , which belongs to Si-O-Si bending vibration [23] . There is a characteristic band appears in the range of 500-600 cm -1 , which corresponded to Al-O-Al linkages [24,25] . The weak peak near 462 cm -1 belongs to Ce-O vibration, according to literatures [26,27] . Fig. 7 shows the intensity of the peak at 680 cm -1 gradually increases as the doping amounts below 7 wt.%. However, the peak turns to becomes weak and wide as the doping amount beyond 8 wt.%. In addition, the Raman vibration spectra peak at 680 cm -1 shifted towards higher frequencies with increased CeO 2 content, which can be explained by the substitution of Si atom by increased Al in the symmetric stretching mode of Si-O-Si (Q 4 ) [23] . The schematic process in glass is presented in Fig. 8. It shows the structural disordering tends to increase with increased CeO 2 and the main Raman spectroscopy inclines to be homogeneous [28,29] . CeO 2 effects on the microstructure evolution. It can be seen from Fig. 9 (a 1 ), (b 1 ) and (c 1 ) that the surface of the samples was smooth and bright, which is essentially consistent with the macro-pro le and the XRD results that these samples are mainly glass. However, obvious crystals precipitate on the sample surface with over enhanced CeO 2 content, just as shown in Fig. 9 (d 1 ) and (e 1 ).

Microtopography analysis
Energy-dispersive spectroscopy (EDS) was also used to evaluate compositional distribution of the samples holding different doping content of CeO 2 [30,31] , and the element mapping of Si, O, Ce and Al are illustrated in Fig. 9. It can be seen that all the elements in the solidi ed bodies were relatively uniform without enrichment, which indicates that the simulated nuclides Ce is uniformly solidi ed into granite. Combined Fig. 9    The Raman spectra of the samples with the doping amounts of CeO2 from 5 wt.% to 10 wt.%  wt.% (e) 10 wt.% Figure 10 Vickers hardness of sample with holding CeO2 from 0 wt.% to 20 wt.%