Raman identification and characterization of chemical components included in simulated nuclear fuel debris synthesized from uranium, stainless steel, and zirconium

ABSTRACT

Raman spectroscopy is a powerful technique for studying nuclear materials. However, it has been scarcely utilized for nuclear fuel debris. Here, we present a Raman study of several types of simulated nuclear debris synthesized from uranium, stainless steel, and zirconium to identify and characterize chemical components included in the simulated debris. Raman spectroscopy sensitively identified many kinds of chemical components: cubic UO₂, U₃O₈, (Fe,Cr)UO₄ (iron–chromium uranate), spinel oxides, monoclinic ZrO₂, tetragonal ZrO₂, and Zr₃O. Some details concerning the chemical states of each component included in the simulated debris were obtained (e.g. spinel oxides were suggested to consist of Fe, Cr, Ni, Zr, and U). The results obtained here will be helpful in the Raman analysis of actual nuclear debris, such that in the Fukushima Daiichi nuclear power plants.

Graphical Abstract

KEYWORDS:

• Raman
• nuclear fuel debris
• chemical states

1. Introduction

In the decommissioning of the Fukushima Daiichi nuclear power plants, the retrieval of the nuclear fuel debris from the reactors and their storage, processing, and disposal are central missions. The nuclear fuel debris is considered to be complicated, heterogeneous, rugged, and solid mixtures consisting of various chemical components because they were generated through melting and cooling of nuclear fuels, zircaloy cladding, control rods, and structural materials in a wide range of temperatures and atmospheric conditions [1]. So far, the isotopic and physicochemical properties of the fuel debris have been characterized through particulates released into the environment (e.g. [2–5]) and some particles sampled from the reactors [6]. However, the nuclear fuel debris is mostly veiled because the amount of the actual sample is little, and the analytical methods applied to the nuclear fuel debris are very limited. Further analyses are desired to uncover what nuclear fuel debris is, and it will be helpful for future storage, processing, and disposal of the debris.

Studying simulated nuclear fuel debris is an effective approach to understanding the chemical properties of actual nuclear debris. This is because simulated debris systematically synthesized in controlled experimental conditions is considered to be model materials of a part of the actual complicated debris, and experimental data obtained from simulated debris are helpful in interpreting the data of the actual debris. Because the chemical stability of the debris is one of the most important properties to understand for the future treatments of the actual debris, a number of leaching and dissolution studies have been conducted using simulated debris [7–17]. In these studies and other characterization researches on simulated debris [18–24], the chemical states of simulated debris have been mostly analyzed using X-ray diffraction (XRD) and electron microscope techniques, which revealed crystal phases and microscopic morphologies with spatial elemental distributions. Indeed, studying simulated debris using other analytical methods will advance understanding of the physicochemical features of debris.

Raman spectroscopy is a powerful analytical technique to reveal micro-scale structures and detailed chemical states and is widely applied to various materials, including nuclear materials (e.g. [25–51]). However, the Raman analyses of actual [27,39,42,47] and simulated [14,16,17] nuclear debris are scarce, and very limited aspects of nuclear fuel debris have been studied so far. In our recent studies [16,17], several types of simulated debris were synthesized from uranium, stainless steel (SUS), and zirconium (i.e. U-SUS and U-SUS-Zr simulated debris) to understand the debris of the Fukushima Daiichi nuclear power plants, and Raman spectroscopy was utilized only to discuss the chemical alteration of the simulated debris through reactions with H₂O₂ in an aqueous solution [17] and whether Fe is dissolved in cubic UO₂ matrix solid solutions [16]. Herein, we identified and characterized various chemical components included in the U-SUS and U-SUS-Zr simulated debris using Raman spectroscopy.

2. Experimental

2.1. Synthesis of simulated nuclear fuel debris

The simulated debris analyzed in this study was same as that used in our previous studies [16,17]. The synthesis conditions are summarized in Table 1. Briefly, the starting materials were a mixture of UO₂, stainless steel (SUS304, <100 mesh powder, Niraco, Japan), Zr metal (>95%, FUJIFILM Wako Pure Chemical Ltd., Japan), and ZrO₂ (>98%, FUJIFILM Wako Pure Chemical). The elemental composition of the SUS304 alloy (in wt%) included Fe: 68.77, Cr: 18.71, Ni: 11.09, Si: 0.86, Mn: 0.18, C: 0.15, and others: 0.24. The mixture of the samples was well ground using an agate mortar and heated on an alumina dish in the central part of a reaction tube made of alumina surrounded by an electric furnace at 1200°C or 1600°C with a gas flow of ultrapure Ar (Grade 1) or 98 vol.% Ar + 2 vol.% O₂ for 1, 8, or 12 h. For samples containing Zr synthesized by heating at 1600°C in a 2% O₂ atmosphere, the heat treatments were conducted on a platinum foil placed on an alumina dish to ease the collection of the samples. The products were pulverized using an agate mortar.

Table 1. Synthesis conditions of the U-SUS and U-SUS-Zr simulated debris and the chemical components identified by XRD and Raman spectroscopy.

Simulated debris (ID)	Synthesis conditions				Chemical components identified
	Starting materials and mol ratio	Temperature (°C)	Atmosphere	Heating duration (h)	XRD	Raman
(1) US-1200-O		1200	2%O2	1	UO2(s.s.) U3O8 (Fe,Cr)UO4 Fe3O4	U3O8 (Fe,Cr)UO4 Spinel oxides(s.s.)
(2) US-1600-O	UO2 SUS304U:(Fe+Cr+Ni) = 1:1		2%O2	1	UO2(s.s.)	UO2(s.s.) U3O8 Spinel oxides(s.s.)
(3) US-1600-A		1600	Ar	1	UO2	UO2(s.s.) U3O8 Spinel oxides(s.s.)
(4) US-1600-O-12			2%O2	12	UO2(s.s.) U3O8	U3O8 Spinel oxides(s.s.)
(5) USZr-1200-O		1200	2%O2	1	U3O8 (Fe,Cr)UO4 t-ZrO2	UO2(s.s.) U3O8 (Fe,Cr)UO4 Spinel oxides(s.s.) m-ZrO2
(6) USZr-1600-O	UO2SUS304Zr		2%O2	1	UO2(s.s.) t-ZrO2	UO2(s.s.) Spinel oxides(s.s.) m-ZrO2 t-ZrO2
(7) USZr-1600-O-8	U:(Fe+Cr+Ni):Zr = 1:1:1		2%O2	8	UO2(s.s.) t-ZrO2	Spinel oxides(s.s.) t-ZrO2
(8) USZr-1600-A		1600	Ar	1	UO2	UO2(s.s.) U3O8 m-ZrO2 Zr3O
(9) USZr-1600-A-12			Ar	12	UO2(s.s.)	Spinel oxides(s.s.) m-ZrO2 t-ZrO2
(10) USZrO2 -1600-O	UO2 SUS304 ZrO2U:(Fe+Cr+Ni):Zr   = 1:1:1	1600	2%O2	1	UO2(s.s.) t-ZrO2	UO2(s.s.) Spinel oxides(s.s.) t-ZrO2
Heated SUS	SUS304	1600	2%O2	12	Spinel oxides(s.s.)	Spinel oxides(s.s.)
Heated SUS-Zr	SUS304 Zr (Fe+Cr+Ni):Zr =1:1				Spinel oxides(s.s.) m-ZrO2	Spinel oxides(s.s.) m-ZrO2
Heated Zr	Zr				m-ZrO2	m-ZrO2
Heated Zr-1				1	Zr3O m-ZrO2	Zr3O m-ZrO2

Identification by XRD for uranium-containing simulated debris is from ref [16,17], and main components identified by XRD are only indicated in this table. (s.s.): solid solution. (Fe,Cr)UO₄: iron–chromium uranate. Two crystal forms of ZrO₂: monoclinic (m-ZrO₂) and tetragonal (t-ZrO₂). Fe₂O₃ was identified as a minor component by Raman in 1, 2, and 6. A single, sharp Raman band at ~565 cm⁻¹ assignable to Cr₂O₃ was observed as a minor component in 8.

2.2. Raman spectroscopy

For convenient Raman measurements, most of the simulated debris powders were molded into 7-mm diameter disks by hydraulic pressure using a tungsten carbide die and fixed on an acrylic resin holder using an epoxy adhesive (Figure 1a). A confocal Raman microscope setup (NRS-4500, JASCO) was used to obtain the Raman spectra of the simulated debris. A 532-nm laser with 0.5-mW power was introduced to a sample surface on a sample stage through a ×100 objective lens. By moving the motor stages, Raman spectra at many different positions were collected. In the optical microscope ×100 image in Figure 1a, the measurement points are represented by red-marked points. Some XY scans were conducted in several different regions optionally selected for each sample. The total measurement points for each sample were ~100–1000. The spatial measurement region at each single measurement position was about 1 μm³. The wavenumbers (cm⁻¹) in the Raman spectra were calibrated by measuring the Si Raman band at 520.5 cm⁻¹.

Figure 1. (a) Schematic illustrations of the Raman microscopy measurement and a picture of a molded sample and its optical microscope images with different magnifications. The red-marked points on the optical microscope ×100 image are sample measurement points. (b) Backscattered electron (BSE) SEM image of the 4 U-SUS simulated debris sample and elemental mapping data obtained using EDX spectroscopy.

In this study, averaged and nonaveraged Raman spectra were presented to show representative main features of each sample and heterogeneous features of each sample depending on spatial positions, respectively. The averaged Raman spectra were the ones obtained by arithmetic average of all Raman spectra measured for a simulated debris sample, whereas the nonaveraged Raman spectra were the ones measured at a single measurement position.

2.3. XRD and SEM-EDX spectroscopy

The XRD patterns of some samples made from stainless steel and Zr with no uranium were measured to validate the assignments of Raman signals for this study using a Rigaku MiniFlex 600 diffractometer equipped with a D/teX Ultra2 detector with Cu-Kα radiation at 40 kV and 15 mA. Diffraction data were collected from 20° to 140° in 2θ with a step interval of 0.02° at a scan rate of 20°/min. The XRD patterns of the uranium-containing simulated debris investigated in this work were already reported in our previous studies [16,17] but are presented in Figure S1. Scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM-EDX; JCM-7000, JEOL) was conducted to demonstrate elemental heterogeneity for one of the simulated debris (Figure 1b).

3. Results and discussion

3.1. U-SUS simulated debris

Figure 1b shows a SEM image of a U-SUS simulated debris sample with U, Fe, Cr, and Ni elemental mapping. These images indicate the spatial heterogeneity of the elements in the prepared samples. Because of the spatial heterogeneity of the samples, the Raman spectral features varied depending on the measurement positions on the surface of each simulated debris. Hence, in the present study, ~100–1000 Raman spectra were measured at different positions optionally selected on the surface of each sample, and arithmetic average was performed. The averaged Raman spectra are shown in Figure 2a as representatives of each sample, reflecting the main features of the simulated debris. Each averaged spectrum shows a spectral pattern different from those of other spectra. This indicates that each simulated debris is in a different chemical state, which is formed depending on the synthesis conditions. As discussed in the following, all Raman signals in Figure 2a can be assigned to four components: cubic UO₂(s.s.), U₃O₈, (Fe,Cr)UO₄ (iron–chromium uranate), and spinel oxides(s.s.), where (s.s.) stands for a solid solution with some metal ions.

Figure 2. (a) Averaged Raman spectra of the U-SUS simulated debris (1–4). (b) A typical Raman spectrum of cubic UO_2.

3.1.1. Cubic UO₂

The fluorite crystal form of UO₂ is in a cubic arrangement and typically shows three Raman bands at ~445, ~560, and ~1150 cm⁻¹, as shown in Figure 2b. The corresponding three bands due to the UO₂ component were observed in the 2 and 3 simulated debris samples. The ~1150-cm⁻¹ bands observed in 2 and 3 are broader, and their intensities are weaker than that of pure UO₂ (Figure 2b). This indicates that the chemical states of the UO₂ components in 2 and 3 are somewhat different from that of the pure UO₂. The greater intensity of the ~560-cm⁻¹ bands in 2 and 3 relative to that of pure UO₂ is also indicative of the difference in the UO₂ components included in these simulated debris. Figure S2 presents that the wavenumber position and broadness of the nonaveraged (single-point) Raman bands depend on the measurement positions, and hence, the chemical states of the UO₂ components included in 2 and 3 are not homogeneous.

The UO₂ components observed in 2 and 3 are probably in hyperstoichiometric states (UO_2+x) because UO_2+x also shows spectral patterns similar to those of UO₂ but with a smaller intensity at ~1150 cm⁻¹ and a larger intensity at ~560 cm⁻¹ [38]. Because another Raman band at ~630 cm⁻¹ was observed for UO_2+x with x values of more than ~0.1 [38] and not clearly observed in 2 and 3, the x value of UO_2+x in 2 and 3 can be estimated to be less than ~0.1. On another note, the spectral features regarding the shoulder signals around 1100 cm⁻¹ of 2 and 3 (the dashed line in Figure 2 shows the shoulder signals) were not observed in UO_2+x [26,29,38]. Therefore, it was suggested that the UO₂ components included in 2 and 3 form UO₂(s.s.), which are UO₂ solid solutions mainly with Fe deriving from the SUS starting material, as already pointed out in our previous papers [16,17].

The UO₂ component was not observed in 1 and 4, although UO₂(s.s.) and UO₂ were identified using XRD as a main component in 1–4 (Table 1 and Figure S1). The UO₂(s.s.) components of 1 and 4 are considered dissolve a higher degree of Fe because their lattice parameters of UO₂(s.s.) are lower than those of 2 and 3 [16,17]. Accordingly, this suggests that a higher degree of Fe dissolution in UO₂ would make its Raman signals inactive, probably because its UO₂(s.s.) crystal structure becomes far from that of pure UO₂ by Fe dissolution. The missing of the UO₂ Raman bands in 1 and 4 cannot be explained by UO_2+x [29,38], supporting that Fe is dissolved in UO₂.

3.1.2. U₃O₈

In the Raman spectra of the U-SUS simulated debris (Figure 2a), the signals at ~800 cm⁻¹ and in the region of less than ~500 cm⁻¹ were attributed to U₃O₈. This assignment was confirmed by comparing the nonaveraged (single-point) Raman spectra of the U₃O₈ component observed in the simulated debris (Figure S3) with the Raman spectra in the literature [25,26,28,30,31,34,36,44] and of a U₃O₈ sample we prepared (Figure S4). These spectral patterns are clearly different from those of other uranium oxides whose oxidation state is close to U₃O₈, such as U₃O₇ [25] or U₄O₉ [32]. Because U₃O₈ and U₃O_8−x show similar Raman spectra [44], the slight difference in the oxidation states of the U₃O₈ components is difficult to discuss. Similarly, discussing the formation of U₃O₈ solid solutions like UO₂ in detail is difficult because the standard spectra of the U₃O₈ sample are heterogeneous (Figure S4), and the spectral signature induced by the solid solution formation is obscure.

The spectral feature of the U₃O₈ components observed in 2 and 3 was different from those observed in 1 and 4: the sharp bands in the region less than ~500 cm⁻¹ were not observed in 2 and 3 (Figure 2a). This feature can be clearly seen in the nonaveraged spectra in Figure S3, and all Raman spectra of the U₃O₈ components observed in 2 and 3 were broad and structureless. Because U₃O₈ was not detected by XRD in 2 and 3 (Table 1, Figure S1), the broad Raman signal of the U₃O₈ component is due to a U₃O₈ component undetected by XRD. Such U₃O₈ components have not been reported so far, to the best of our knowledge. Although the broad U₃O₈ component is also observed in 1 and 4 (Figure S3), XRD identified U₃O₈ in 1 and 4 likely because the U₃O₈ components showing sharp Raman bands can be sensitively detected by XRD. The broadening of vibrational Raman bands could result from the existence of multiple vibrational modes with slightly different frequencies that are induced by some structural heterogeneity of materials on a microscopic scale [41].

3.1.3. (Fe,Cr)UO₄

In the Raman spectra of Figure 2a, the (Fe,Cr)UO₄ uranate component was identified by referring to the Raman spectra of (Fe,Cr)UO₄ synthesized from starting materials with different Fe/(Fe+Cr) ratios, as reported in our recent study [52]. The synthesized (Fe,Cr)UO₄ showed a single main band around 700 cm⁻¹ regardless of the Fe/(Fe+Cr) ratio, while spectral patterns consisting of several bands in the region less than 500 cm⁻¹ were strongly dependent on the Fe/(Fe+Cr) ratio. Figure 3 shows the nonaveraged spectra identified as the (Fe,Cr)UO₄ components in 1. The Fe/(Fe+Cr) ratio of the (Fe,Cr)UO₄ components was estimated to be between 0.75 and 1 by comparing with the spectral features of the synthesized (Fe,Cr)UO₄ [52]. The high Fe/(Fe+Cr) ratio of the (Fe,Cr)UO₄ components observed in the simulated debris corresponds to the Fe-rich elemental composition (68.77 wt%) of the SUS starting material. The spectral patterns due to the (Fe,Cr)UO₄ components as shown in Figure 3 were only observed in 1 but not in 2–4, consistent with the XRD result (Table 1, Figure S1). The reason why the bands around 700 cm⁻¹ were also observed in 2–4 (Figure 2a) is that other chemical components also show bands at ~700 cm⁻¹, which is discussed in the following section.

Figure 3. Nonaveraged (single-point) Raman spectra of the (Fe,Cr)UO₄ uranate components observed in the 1 U-SUS debris.

3.1.4. Spinel oxides

Figure 4 shows representative examples of nonaveraged Raman spectra of chemical components showing a band around 700 cm⁻¹ but not assignable to (Fe,Cr)UO₄. These spectra did not show signals characteristic of (Fe,Cr)UO₄: several bands in the region less than 500 cm⁻¹ were observed in the higher Fe/(Fe+Cr) ratio (Figure 3) and a single band around 425 cm⁻¹ was observed in the lower Fe/(Fe+Cr) ratio [52]. Therefore, the components of Figure 4 are not due to (Fe,Cr)UO₄. In the spectra of Figure 4, the wavenumber positions of the ~700-cm⁻¹ bands and their bandwidths are slightly different from each other, suggesting that these components are highly heterogeneous and can form various chemical states that are very similar to each other. In the averaged Raman spectra of Figure 2a, all ~700-cm⁻¹ bands consist of these components, and the ~700-cm⁻¹ band of 1 has some contributions of (Fe,Cr)UO₄.

Figure 4. Representative some examples of the nonaveraged (single-point) Raman spectra of chemical components assignable to spinel oxides, observed in the 1–4 U-SUS simulated debris.

It has been reported that spinel oxides, denoted generally as M₃O₄, form various kinds of solid solutions. Among them, ferrite and chromate spinel oxides commonly show a single main Raman band at ~700 cm⁻¹, and the wavenumber positions of the ~700-cm⁻¹ bands are sensitively shifted depending on the kind and ratio of metal elements M dissolved in the spinel oxides [53,54]. In particular, the Raman spectra of NiFe_xCr_2−xO₄ (0 < x < 2) spinel oxides are very similar to those in Figure 4: single main bands of the spinel oxides appear in the 686–702 cm⁻¹ region depending on the x value [53]. The spinel oxides Fe₃O₄, NiFe₂O₄ (x = 2), and NiCr₂O₄ (x = 0) were readily excluded as candidates for the components of Figure 4 because Fe₃O₄ shows a single main band at ~670 cm⁻¹ (rather lower than the ~700-cm⁻¹ bands observed in Figure 4) and NiFe₂O₄ (x = 2) and NiCr₂O₄ (x = 0) show strong sidebands at ~500 cm⁻¹ [53]. Therefore, the components of Figure 4 are assignable to components close to NiFe_xCr_2−xO₄ (0 < x < 2) spinel oxides. The Raman spectrum of 3 showing bands at ~545 and ~680 cm⁻¹ is close to that of FeCr₂O₄ [53]. Some iron oxides, such as Fe₂O₃, FeO, and FeOOH, were excluded as the components of Figure 4 because of their different Raman spectral features [55].

The influence of the existence of uranium on the formation of spinel oxides can be discussed by comparing the Raman spectrum of 4 (the bottom one of Figure 4) and that of a heated SUS sample (Figure 5a) because both samples were prepared at the same heating condition (1600°C, 2% O₂, and 12 h), and the only difference between the two samples is the presence or absence of uranium. The heated SUS sample was identified as spinel oxides dissolving Fe, Cr, and Ni because the XRD pattern (Figure 5b) is the closest to the standard XRD data of (Fe_0.8Cr_0.2)(NiFe_0.2Cr_0.8)O₄ among some spinel oxides available in the Inorganic Crystal Structure Database (ICSD) [56]. The Raman spectrum of the heated SUS sample (Figure 5a) shows a single band at ~700 cm⁻¹, and its spectrum feature is very similar to those observed in the simulated debris (Figure 4), supporting that the components of Figure 4 are spinel oxides dissolving Fe, Cr, and Ni elements. Importantly, the Raman bands due to the spinel oxides observed in 4 exhibit a single main band at ~735 cm⁻¹ (Figures 4 and 2a), higher than that of the heated SUS (Figure 5a), and the high wavenumber position cannot be explained by any spinel oxides formed from Fe, Cr, and Ni elements in the literature. Therefore, the high wavenumber band of 4 at ~735 cm⁻¹ strongly suggests the dissolution of U in the spinel oxides generated in the U-SUS simulated debris and the formation of (Fe,Cr,Ni,U)₃O₄. Although clear evidence was not observed, some degree of U dissolution is also expected in 1–3. It is generally recognized that uranium elements are incorporated in Fe₃O₄ (magnetite) by X-ray absorption spectroscopy [57–59] and theoretical studies [60–62]. To the best of our knowledge, the present study provides the first evidence suggesting that U can be dissolved even in composite spinel oxides. Furthermore, the present study demonstrated that Raman spectroscopy is a sensitive probe for such composite spinel oxides relative to XRD (Table 1, Figure S1) partly because the crystallinity of composite spinel oxides formed from several different elements would be low.

Figure 5. (a) Averaged Raman spectrum of a heated SUS sample. (b) XRD pattern of the heated SUS and some standard XRD data of spinel oxides available in the ICSD [56].

3.2. U-SUS-Zr simulated debris

Figure 6 shows the averaged Raman spectra of the U-SUS-Zr simulated debris. Similar to the U-SUS simulated debris, ~100–1000 numbers of Raman spectra were measured at different positions of each sample because of the spatial heterogeneity of the samples. The spectral features varied depending on the synthesis conditions, suggesting that various chemical components were observed in the U-SUS-Zr simulated debris. The spectral features are also different from those of the U-SUS simulated debris (Figure 2a), indicating that some chemical components originating from Zr were generated in the U-SUS-Zr simulated debris. On the other hand, Raman signals due to the four components identified in the U-SUS simulated debris (UO₂(s.s.), U₃O₈, (Fe,Cr)UO₄, and spinel oxides(s.s.)) were also observed. These four components are discussed in the next section. Figure S5 presents some representative examples of the nonaveraged Raman spectra of the four components.

Figure 6. Averaged Raman spectra of the U-SUS-Zr simulated debris (5–10).

3.2.1. UO₂(s.s.), U₃O₈, (Fe,Cr)UO₄, and spinel oxides(s.s.)

Because the band at ~1100 cm⁻¹ due to the UO₂ component was observed in 5, 6, 8, and 10 similar to that with the U-SUS simulated debris, the UO₂ components in the U-SUS-Zr simulated debris are considered to be UO₂(s.s.) with Fe. Zr is also expected to be dissolved in the UO₂(s.s.) component because of its much lower lattice parameters [16,17]. Contrary to the XRD result, UO₂(s.s.) was not observed in 7 and 9 (Table 1) probably because of the high degree of Fe dissolution. U₃O₈ components were identified in 5 and 8, and the U₃O₈ component in 8 is a U₃O₈ component undetected by XRD (Table 1). The (Fe,Cr)UO₄ component with a high Fe/(Fe+Cr) ratio was identified only in 5, and the spectral patterns as shown in Figure 3 were not observed in 6–10, consistent with the XRD result (Table 1).

Spinel oxides were identified in 5, 6, 7, 9, and 10, and U is probably dissolved similarly to the U-SUS simulated debris because high wavenumber positions of the ~700-cm⁻¹ band were observed in 9: 720 cm⁻¹ in Figure 6 and 725 cm⁻¹ in Figure S5. Additionally, Zr is also expected to dissolve in the spinel oxides because the Raman spectrum of spinel oxides observed in a heated SUS-Zr sample (Figure 7a) also shows a band at ~745 cm⁻¹, which is higher than that of the heated SUS (~700 cm⁻¹ in Figure 5a). This higher wavenumber is likely due to the dissolution of Zr in the spinel oxides. The difference in the XRD signals of the spinel oxides between the heated SUS and the SUS-Zr samples (Figure 7b) supports the possibility of Zr dissolution in the structure of spinel oxides.

Figure 7. (a) Averaged Raman spectrum of a heated SUS-Zr sample. (b) XRD patterns of heated SUS and SUS-Zr samples.

3.2.2. ZrO₂

Two crystal forms of zirconia, monoclinic (m-ZrO₂) and tetragonal (t-ZrO₂), were identified in the U-SUS-Zr simulated debris. The m-ZrO₂ components observed in 5 and 8 (the two top spectra in Figure 8) show spectral patterns similar to those reported in the literature (sharp bands at ~90, ~180, ~330, ~380, ~475, and ~620 cm⁻¹) [63–66], whereas 6 shows various spectral patterns that are somewhat different from those of m-ZrO₂, suggesting that the chemical states of ZrO₂ in 6 are close to those of m-ZrO₂ but do not have a pure crystal structure and have some structural heterogeneity. Furthermore, 9 showed a Raman spectrum of the ZrO₂ component close to that of m-ZrO₂. The bottom spectrum of 6 in Figure 8 is assignable to t-ZrO₂ because a band characteristic of t-ZrO₂ was observed at ~260 cm⁻¹ [63,64,67], and the ~180-cm⁻¹ band characteristic of m-ZrO₂ (Figure 8) was absent. The t-ZrO₂ signals at ~260 cm⁻¹ were dominantly observed in the averaged spectra of 7, 9, and 10 in Figure 6, indicating that t-ZrO₂ is a dominant species in these simulated debris. Importantly, t-ZrO₂ components were not observed in the heated SUS-Zr, which only consists of spinel oxides and m-ZrO₂, indicating that U is dissolved in the structure of t-ZrO₂ and is necessary for the formation of t-ZrO₂ in the present synthesis conditions. The spectral patterns of t-ZrO₂ observed in the simulated debris were not exactly identical to those in the literature [63,64,67], supporting some structural distortion of t-ZrO₂ by U dissolution.

Figure 8. Representative some examples of nonaveraged (single-point) spectra assignable to ZrO₂ components observed in the 5, 6, 8, and 9 U-SUS-Zr debris.

3.2.3. Zr₃O and the fluorescent signal

The Zr₃O component was the dominant spectral signature observed in 8 (Figure 6), and this assignment was confirmed by comparing with the Raman spectrum of Zr₃O (Figure 9a). The Zr₃O sample was prepared by heating Zr at 1600°C in 2% O₂ for 1 h and was confirmed to be mainly Zr₃O by XRD (Figure 9b). A strong fluorescent signal was observed in 8 in the wavenumber region greater than ~900 cm⁻¹ (Figures 9a and 6). This signal was also observed in alumina dishes used in the synthesis of the simulated debris. Hence, the fluorescent signal is due to the contamination of the alumina to the simulated debris.

Figure 9. (a) Averaged Raman spectrum of the 8 U-SUS-Zr debris (the same as that in Figure 2) and Raman spectrum of a Zr sample heated at 1600°C in 2% O₂ for 1 h. (b) XRD pattern of the heated Zr (1 h) and the standard XRD data of Zr₃O in the ICSD [56].

4. Summary

Various chemical components included in several types of simulated nuclear fuel debris synthesized from uranium, stainless steel, and zirconium were identified and characterized by Raman spectroscopy. UO₂ and UO₂(s.s.) were identified, but UO₂(s.s.) dissolving a large amount of Fe was not. U₃O₈ was sensitively observed, and a U₃O₈ component undetected by XRD was identified in some simulated debris. (Fe,Cr)UO₄ was identified, and the Fe/(Fe+Cr) ratio in (Fe,Cr)UO₄ can be estimated from the Raman spectral patterns. Spinel oxides were sensitively detected for most of the simulated debris, and many types of composite spinel oxides that are slightly different from each other were revealed to be formed in the simulated debris. The Raman signals of the spinel oxides suggest that U and Zr elements are also dissolved in the spinel component. Raman signals somewhat different from pure m-ZrO₂ were observed, implying that a structural change occurred in the m-ZrO₂ components. The observation of the Raman signal of t-ZrO₂ suggested that the formation of t-ZrO₂ necessitated the dissolution of U in the t-ZrO₂ component in the present experimental condition. The Zr₃O signal was sensitively detected by Raman. Overall, Raman spectroscopy can identify chemical components that XRD can do with different sensitivity, and can discuss chemical states, especially concerning elemental dissolution in spinel oxides. The results obtained in this study will be beneficial when actual nuclear debris, such as that in the Fukushima Daiichi nuclear power plants, is analyzed using Raman spectroscopy.

Acknowledgments

The authors thank T. Yomogida for the SEM measurements. This work was supported by the JAEA Nuclear Energy S&T and Human Resource Development Project through Concentrating Wisdom, Grant Number [JPJA18P18071886].

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/00223131.2022.2128460

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.