1 Introduction

As a general nuclear fuel, uranium dioxide has been studied for 50 years, being concentrated on oxidation under different temperatures, oxygen pressure and so on ^[1]. Temperature is the main condition of oxidation process. Properties of the oxidation products change distinctively with temperature, such as the chemical valence, electronic structure and crystal structure. As temperature increases, UO₂ with a florite crystal type forms a series of valence mixed oxidation products: U₃O₇/U₄O₉,U₃O₈^[2]. This process occurs in the nuclear fuel cycle, transport and waste disposal, with a risk of the volume increase. For example, The final oxidation product U₃O₈ has a 36% net volume increase over UO₂ due to remarkable difference of crystalline structure and the decreased U/O ratio. To avoid this risk, the uranium dioxidation processes under different tempertures should be studied in advance. X-ray absorption near-edge spectroscopy (XANES) in fluorscence mode is an advanced technique for acquiring information about the chemical valence state and electronic structure of actinide elements. XANES require not much on the probed samples, and its high penetrating power allows bulk detection and in-situ measurement ^[3]. For detecting valence state of L edge of actinide elements, however, conventianl XANES is limited by the excessively broad spectra width originating from the shorte 2p core hole lifetime.

By collecting the Lα emission line, hhigh energy resolution fluorescence detection for X-ray near-edge absorption spectroscopy (HERFD-XAS) performs much better in energy resolution than conventional XAS. HERFD-XAS stations have been established worldwide in increasing numbers^[4] during the past two decades, and in the recent years, several groups began their efforts of using this method to study chemical states and electronic structure of actinide elements ^{[5, 6]}. The broad width of L₃ edge HERFD-XAS at Lα line originating from the core hole lifetime of 3d is longer than 2p, so some details of the spectra could be extracted^[7]. The white line of XAS could provide the information of unoccupied 6d states. The shift of white line is limited to the spectrum width, and narrow white line of HERFD-XAS could obtain more precise structure or the small variation of 6d states^{[8, 9]}. Furthermore, most of the chemical properties of actinide is related to the 5f states^[10]. The pre-edge peaks of actinide elements HERFD-XAS originating from transition from 2p to 5f could provide the information of valence state of 5f ^[11]. T.Vitova and co-workers studied different valence of urinium oxides^[12]. The HERFD-XAS of three standard samples illustrated the differerce among U⁴⁺, U⁵⁺andU⁶⁺. The energy resolution at U L₃ of conventional XAS is about 14 eV, while it was 9 eV for HERFD-XAS. This improved energy resolution provided precise analysis about valence changes.

The energy resolution power of U L₃ HERFD-XAS is limited by energy resoluton of the incident beam, the spectrometer and the nature broadening related to the lifetime of 3d corehole. The nature broadening of uranium M_4,5 edge is about 3 eV, so the energy resolution of U L₃ HERFD-XAS must be better than 3 eV. Under these conditions, the chemical shift (<3 eV) cannot be detected by L₃ edge HERFD-XAS. Therefore, the HERFD-XAS of U M₄-edge with high energy resolution^[13] was developed. The X-ray of U M₄-edge is too soft to penetrate the gas atmosphere in in-situ study^[14], so the hard X-ray of U L₃-edge, of good penetrability, is suitable for the HERFD-XAS technique in in-situ environment. A high resolution X-ray fluorescence spectrometer based on Rowland circle geometry on BL14W1 at SSRF has been established^[15] for X-ray emission spectroscopy and HERFD-XANES.

In this paper, HERFD-XANES data of U L₃-edge with high signal-to-noise for UO₂ oxidation are obtained, by using the in-situ high-energy resolution X-ray absorption spectroscopy with an in-situ heating cell.

2 Experiments

The crystal structures of the UO₂ sample before and after in-situ oxidation were characterized with powder X-ray diffraction (XRD) on a Bruker D8 Advance X-ray diffractometer using the Ni-filtered Cu Kα radiation source at 40 kV and 40 mA, performed on the BL14W1 beamline at SSRF^[16]{Yu, 2015 #4}. The incident beams were from a pair of Si(311) reflection crystal monochromator. Little mismatch of the monochromator was performed for the rejection of higher harmonics. The flux at the sample position could reach to 3×10¹² photos/s. The energy calibration was achieved at Zr-L3 absorption edge (17998 eV). As shown in Fig 1(a), the sample was placed in the in-situ heating cell, and the centers of the cell, crystal and detector were on the Rowlan circle. The X-ray beams were from a bending crystal Ge (777) (1 m in curvature radius) and the energy resolution could be down to 2.4 eV with the convolution of incident energy. The XANES spectra were scanned from 17155 eV to 17250 eV in 0.3 eV steps around white line. The emission spectrometer was performed for HERFD-XAS and resonant X-ray emission spectrum. The conventional XAS (total fluorescence yield XAS TFY-XAS) was acquired by a gas chamber at transition mode. The UO₂ powder, fabricated by China North Nuclear Fuel Company Limited, was annealed for 6 h under 5% H₂ and Ar atmosphere at 1273 K to ensure that the composition of UO₂ powder was stoichiometric. The UO₂ powder was oxidated to U₃O₈ in air. Two standard samples (UO₂ and U₃O₈) were used as the reference. The sample UO₂ was powdered in order to accelerate the oxidation process. Two asbestine cylinder was used to clamp the powder sample and reduce absorption of the incident X-ray. The in-situ cell then was placed at the intersection of X-ray line and Rowland circle. The heating appratus contains two probes inside the cell and a temperature controller that altered and maintained the temperature at the sample position. The cell has two more probes to provide 100% O₂ gas and pure He gas as protective gas. The entire cell was water cooled (Fig 1b). The sample UO₂ was heated to 300, 470, 520, 620, 770 and 970 K at O₂ atmosphere and kept at each temperature for 1 hour. Then, the HERFD-XAS was acquired at pure He atmosphere to ensure the oxide reaction no longer proceeded.

Fig. 1.Full-screen View

Priciple of the HERFD-XAS spectrometer (a) and the in-situ cell (b).

3 Results and Discussion

The spectra of HERFD-XAS and conventional XAS (TFY) of UO₂ at room temperature are shown in Fig 2(a). The full width at half maximum (FWHM) of Peaks A and B for TFY XAS is 14 and 7 eV, respectively; while they are 9 and 6 eV for HERFD-XAS. Peak A is the white line. It is sensitive to the variation of valence state, and its shift provides information about valence state of oxidation products. Peak B originates from mutiple scatting, and the distance between A and B is inversely proportional to lattic spacing. So, the more narrow peaks, the more precisce lattic spacing will be obtained ^[17]. U₃O₈ forms via mixed valence oxidation, and Peaks A and B of U₃O₈ are wider than those of UO₂. The HERFD-XAS of UO₂ and U₃O₈, as standard samples, are shown in Fig 2(b). The peak intensities decreased gradually and the width increased. The intensity decrease of the main peaks could be due to the damaged symmetry of crystalline structure by the added O atoms ^{[18, 19]}., And the broadened main peaks could be originated from the combination of two valences of U₃O₈^[20]. The spectrum of U₃O₈ has a distinct shoulder peak at the higher energy side, due to mutiple scattering.

Fig.2.Full-screen View

The U-L₃ edge HERFD-XAS and TFY-XAS of UO₂ (a) and HERFD-XAS of UO₂ and U₃O₈ (b).

Fig3 shows the valence and structure changes of UO₂ powder samples heated in-situ in O₂ from 300 K to 970 K. The peak of HERFD-XAS at 300 K had the same shape as UO₂. At 470 K the peak became broadened, with the disappearance of the sub-peak just after the main peak. The HERFD-XAS of UO₂ at 520 K was basically the same as that of 470 K, indicating that the oxidation process could not be maintained at lower temperatures, though some authors found that UO₂ powder above 470 K was transformed to U₃O₇/U₄O₉ or similar structure. ^[21] At 620, 770 and 970 K, the HERFD-XAS spectra, without substantial changes, are similar to U₃O₈.

Fig.3.Full-screen View

The HERFD-XAS of UO₂ at different temperatures.

Powder diffraction patterns of the sample before and after oxidation are shown in Fig 4. In Fig 4(a), at room temperature, the diffraction patterns are of characteristic diffraction peaks corresponding to the UO₂ structure (PDF-65-0285), while after oxidation (Fig. 3b), the diffraction patterns show the same peak positions to U₃O₈ structure (PDF-47-1493), indicating the transform from UO₂ to U₃O₈. 2θ (°) ×

Fig. 4.Full-screen View

The powder XRD patterns of the sample at room temperature (a) and after the oxidation process (b).

4 Conclusion

The in-situ HERFD-XAS of UO₂ oxidation is helpful for analyzing complex products in oxidation process and providing precise XAS of U L₃ edge. The in-situ high energy-resolution X-ray absorption spectroscopy technology has been established successfully at SSRF. From the HERFD-XAS of UO₂ samples heated in-situ to 300–970 K and oxidized for one hour, the variation of valence states can be described and the changes of crystalline structure can be deduced. Temperature is basically the decisive factor in the oxidation process. The HERFD-XAS indicated structural changes agree with previous reports by other methods.