Hydrothermal conversion of cerium oxalate to CeO₂: a parade of oxalate and water coordination modes

Abstract

Hydrothemal conversion of lanthanide/actinide oxalate salts to nanocrystalline oxide has gained technological importance in the chemistry of f-elements. Herein, we present a study describing the mechanism behind the hydrothermal recrystallization and decomposition of cerium (and Ce-Gd) oxalate to oxide. Structural determination of all intermediate nanocrystalline products was possible thanks to three-dimensional electron diffraction. This work reveals the enormous structural abilities of simple oxalate ligands.

---

Oxalates are low-cost, simple, but essential species in the chemistry of transition metals, lanthanides,¹ and actinides.² Thanks to the low solubility of oxalate salts, oxalic acid is used: as a separation agent of metal ions from aqueous solutions; as a ligand in coordination compounds; in metal–organic frameworks. These salts can be converted readily to nano- and microcrystalline oxides by an appropriate calcination cycle, typically beyond 600 °C.^3,4 Recently, Walter et al.⁵ reported the conversion of uranium(IV) oxalates to oxide under hydrothermal conditions at mild temperatures (160–250 °C). The authors extended their work on the hydrothermal decomposition of various actinide oxalates (Th, U, Np, Pu)⁶ and their solid solutions.^7,8 Typically, an oxalate dispersion in pure water (pH adjusted or in the presence of hydrazine) is loaded into an autoclave and treated at 100–250 °C. This procedure results in agglomerates of AnO₂ nanoparticles of approximate size 3 to 10 nm. Later, comprehensive studies were conducted by Manaud et al. on Th(IV)⁹ and U(IV)¹⁰ while changing the main reaction parameters (pH, temperature, and time). There was a clear link between the treatment temperature, pH of the solution, and the crystallite size and morphology of the final product. Recently, Baumann et al.¹¹ compared the hydrothermal and thermal conversion of Pu(IV) oxalate to nanocrystalline PuO₂ mentioning most of the positive aspects of the hydrothermal route: lower temperature, shorter reaction time, and lower content of residual carbon. All the above-mentioned works described a drastic change in the morphology of the starting powder during the process. However, tracking of the reaction mechanism is complicated due to the hydrothermal conditions, so determination of the structure of intermediates is complicated due to nanocrystallinity. Therefore, the detailed reaction mechanisms are not known. Few studies have described the hydrothermal treatment of cerium oxalate. Flower-like CeO₂ microcrystals were hydrothermally synthesized from an oxalate precursor in an alkaline environment and hydrogen peroxide.¹¹ A similar procedure yielding hierarchical mesoporous ceria was obtained in the presence of some amino acids previously.¹² Recently, mixed U_1−xCe_xO_2+δ·nH₂O (0.1 ≤ x ≤ 0.7) solid solutions were prepared using this method.¹³

Herein, we focused on the hydrothermal conversion of pure cerium oxalate hydrate to oxide. Ceria has been in the spotlight for many decades, mainly due to its applicability in fuel cells,¹⁴ catalysis,¹⁵ and Pu surrogates.¹⁶ The mechanism of the decomposition and recrystallisation of oxalate (which is responsible for such a drastic change in the morphology of the agglomerates) was studied in detail. This was possible thanks to the newly developed technique of three-dimensional electron diffraction (3D ED), which enabled complete determination of the structure without the need for large crystals.^17–22 In addition, we tested the ability of this route to convert mixed Ce_1−xGd_x(C₂O₄)₃·10H₂O solid solutions to Ce_1−xGd_xO_2−x/2 (x = 0.1, 0.2, 0.3).

Pure and mixed oxalates were prepared by heterogeneous precipitation as detailed previously^1,23 (a description is in the ESI†). Briefly, for the hydrothermal conversion, 250 mg (0.35 mmol) of Ce₂(C₂O₄)₃·10H₂O or Ce_1−xGd_x(C₂O₄)₃·10H₂O (x = 0.1, 0.2, 0.3) was placed in a Teflon™-lined autoclave (capacity of 50 mL) together with distilled water (21 mL) or nitric acid solution of the desired pH. Solutions of HClO₄ and HCl, as nonoxidizing acids, were also tested instead of HNO₃. The mixture was treated at 180–220 °C for ≥ 24 h. The final product was purified by repeated centrifugation and dried at 50 °C under air. All prepared materials were analysed by powder X-ray diffraction (PXRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM), and thermogravimetric analysis (TGA). The sample was treated in pure water (pH = 7, 220 °C, 48 h) and analysed using 3D ED. A detailed description of the analytical techniques is given in the ESI.†

Ce₂(C₂O₄)₃·10H₂O can be converted completely to oxide in aqueous nitric acid solution (pH = 1) at >180 °C. Optimal conditions were found to be 24 h and 220 °C to yield a powder of grain size 200–500 nm (Fig. 1). The final size of CeO₂ crystals increased with the applied temperature (see the micrographs in Fig. S1 and infrared spectra in Fig. S2, ESI†).

Fig. 1 Transmission (top) and scanning (bottom) electron micrographs of CeO₂ powder obtained by the hydrothermal decomposition of Ce₂(C₂O₄)₃·10H₂O in nitric acid solution (pH = 1) at 220 °C for 24 h.

A short study of the hydrothermal conversion of Ce_1−xGd_x(C₂O₄)₃·10H₂O (x = 0.1, 0.2, 0.3) solid solutions showed that the conditions mentioned above (HNO₃ solution, pH = 1) were valid for mixed oxides giving homogeneous nanocrystalline Ce_1−xGd_xO_2−x/2 (see Fig. S3 and S4 and PXRD in Fig. S5, ESI†). However, the yield of Gd(III) incorporation in CeO₂ was not quantitative and showed some fluctuations (see ICP-MS analyses in Table S1, ESI†). Further studies in this direction are beyond the scope of the current work.

The presence of an oxidizing acid (HNO₃) solution was of paramount importance. It facilitated the oxidation of Ce(III) ions to Ce(IV) ions, leading to the oxide. Hydrothermal conversion in HCl or HClO₄ (non-oxidizing acids) solutions of pH = 1 did not lead to CeO₂ (see the XRD results of samples obtained after 24 h at 220 °C in Fig. S6, ESI†). Different concentrations of nitric acid (pH = 1 to 7) in the autoclave batch led to different degrees of conversion. Fig. 2 shows Ce₂(C₂O₄)₃·10H₂O treated in solutions of pH 1 to 7 (HNO₃). In pure water (pH = 7), recrystallisation and a drastic change in morphology occurred, but decomposition to CeO₂ was not reached (Fig. 2, top and middle row). The original cerium oxalate with a flat needle morphology (Fig. S6, ESI†) changed to large rods (from 10 × 10 × 100 μm and larger) with well-developed crystal facets parallel to the elongation direction. Increasing the concentration of nitric acid led to the decomposition of the recrystallized material to oxide (Fig. 2, bottom row). This process was accompanied by full disintegration of the rods to nanocrystals, which sometimes remained assembled to the initial rod shape. A change in agglomerate morphology has been described for all hydrothermally converted oxlates.^5–10 This is one of the main differences with thermal conversion, during which the initial oxalate shape is retained.^1,3,23,24

Fig. 2 Scanning electron micrographs of cerium oxalate with different degrees of conversion to CeO₂. Assemblies of the recrystallised phase with initial Ce₂(C₂O₄)₃·10H₂O (top), recrystallisation and change of morphology to rods (middle), and disintegration of the rods to nanocrystalline CeO₂ (bottom).

Recrystallisation was also studied from the crystallographic viewpoint. The PXRD patterns of selected samples were plotted (Fig. 3). Unfortunately, structure matching from PXRD was possible only for the initial cerium oxalate and final cerium dioxide. Tracking the intermediate phases and resolving their structures was difficult due to the nanocrystallinity of the material and the coexistence of several phases in the samples. This was overcome thanks to 3D ED enabling structural determination of small, unstable, or problematic crystals unsuitable for single-crystal XRD. The structure of the initial oxalate is well known. It crystallizes in the P2₁/c space group with four Ce(III) ions in the elementary cell, each bonded to three oxalate groups and three water molecules in the primary coordination sphere.²⁵ Two additional water molecules are present in each of the four cavities formed by the oxalate coordination polymer. Therefore, the molecular formula should be written as Ce₂(C₂O₄)₃(H₂O)₆·4H₂O, as for some An(III) oxalates with a similar structure.²⁶ There are two formula units within the elementary cell. The recrystallized sample at 220 °C (pH = 7, 24 and 48 h) was selected for structural analysis by 3D ED. This sample showed full recrystallization of the oxalate but low decomposition to oxide (Fig. 3). In total, 114 datasets were indexed. Except for the known initial decahydrate structure (42 indexed datasets), four other structures were observed (72 indexed datasets in total), as shown in Table 1 (full information is given in Table S2, ESI†). The structures were all solved ab initio by the Superflip program.²⁷ The structure present in most of the transformed oxalate hydrates was the orthorhombic P2₁2₁2₁ of Ce₂(C₂O₄)₃(H₂O)₄. Isostructural Pr and Nd oxalates have been described in the ICSD database.^28,29 The other three structures have not been reported.

Fig. 3 Powder X-ray diffraction of the initial Ce₂(C₂O₄)₃·10H₂O (bottom), hydrothermally treated (middle), and final CeO₂ (top).

Table 1 Crystallographic details of hydrothermally recrystallized cerium oxalate phases

Stoichiometric formula	Ce2(C2O4)3(H2O)4	Ce2(C2O4)3(H2O)4·H2O	Ce2(C2O4)3(H2O)4·H2O	Ce2(C2O4)3(H2O)4·H2O
Occurrence (% of cases)	40	39	14	7
Crystal system	Orthorhombic	Triclinic	Orthorhombic	Monoclinic
Space group	P212121	P1̄	Pbcn	P21/n
a, b, c (Å)	8.64(1), 9.59(1), 16.87(1)	10.50(1), 11.77(1), 12.97(1)	22.06(8), 12.96(1), 10.61(1)	21.67(1), 10.51(1), 12.97(1)
α, β, γ (°)	90, 90, 90	92.91(4), 90.58(4), 112.95(4)	90, 90, 90	90, 93.58(4), 90
Z	4	4	8	8

---

The second most abundant phase was a structure with triclinic P1̄ symmetry. If all the water molecules were fully occupied, then the structure was pentahydrate Ce₂(C₂O₄)₃(H₂O)₄·H₂O (see the detailed discussion of the level of hydration in the ESI†). Four water molecules in the formula unit were coordinated to cerium ions and the fifth one resided in a pore. Two other phases, orthorhombic Pbcn and monoclinic P2₁/n, were significantly less abundant and much more disordered judging from the observed diffuse scattering. Both phases were pentahydrate, with four water molecules coordinated to two independent cerium atoms. The fifth water molecule was in a cavity. Hydrolysis or formation of hydroxo-complexes were not observed until the final oxalate decomposition. Probably a higher pH would be required to observe such behavior, as has been observed with Nd.³⁰ The results of 3D ED were supported by PXRD Rietveld analysis of the sample treated at 220 °C for 48 h (Fig. S7, ESI†). The best fit was obtained using the P2₁2₁2₁ and P1̄ phases having the highest abundance in the sample.

Evidently, the four structures had common structural motives worth mentioning. In all structures, oxalate ions adopted bridging coordination modes (Fig. S8, ESI†) to connect cerium ions with coordination modes 1 : 2, 1 : 2 + 1 and 1 : 2 + 2 (Fig. S8, top from left, ESI†), extending the list of oxalate coordination modes for rare-earth ions.³¹ The oxalates in the P2₁2₁2₁ structure formed infinite chains along two nearly perpendicular directions, which created a “wine rack” structural motif (Fig. 4, left). Higher coordination numbers in the Pbcn, P1̄ and P2₁/n structures were probably caused by a specific organization of the infinite chains of oxalates with the 1 : 2 coordination mode solely, all aligned in the same direction. Layers of oxalates with higher denticity were then perpendicular to these polymer chains (Fig. 4, right), here 1 : 2 and 1 : 2 + 2 modes were present.

Fig. 4 Oxalates in the P2₁2₁2₁ structure form infinite chains running along two nearly perpendicular directions (black arrows; the other oxalate molecule and coordinated water molecules have been omitted for clarity). Pbcn structure with infinite oxalate chains running only along one direction marked by black arrows (coordinated water molecules and water in pores have been omitted for clarity).

Six cerium coordination polyhedra were found in the four structures. Type 1 comprised the cerium ion coordinated by nine ligands. Two water molecules were in the cis conformation, with one monodentate oxalate and three bidentate oxalates (Fig. S9, ESI†). Types 2, 3, and 4 are shown in Fig. S10 (ESI†). Coordination was achieved by 10 ligands, two water molecules in trans positions, and two oxalates in the perpendicular direction forming infinite chains. There were two bidentate oxalates in the other perpendicular direction. Type 5 revealed sole coordination by five oxalates (nine oxalate oxygens (Fig. S11, ESI†)), four bidentates and one monodentate. Type 6 involved non-participation in the infinite chains and, except for three bidentate oxalates, it was coordinated to four water molecules (Fig. S12, ESI†). For more details, see the ESI.†

As mentioned above, in the Pbcn, P2₁/n, and P1̄ structures, there were oxalate layers perpendicular to the direction of the infinite oxalate chains. The structural motifs in these layers were almost identical for Pbcn and P2₁/n phases (Fig. 5, left and middle). The difference was that four water molecules were in the layer in the Pbcn structure, but an oxalate replaced it in the P2₁/n phase. The four coordinated molecules in Pbcn pointed outwards from the layer in the direction of the infinite oxalate chains and were coordinated to one cerium atom (type 4). In the P1̄ structure, the layer motif was slightly simpler (Fig. 5, right).

Fig. 5 Structural motifs of oxalate layers perpendicular to the infinite coordination polymer chains. Coordinated water molecules (except those forming the oxalate layer and water in pores) have been omitted for clarity.

The global water content in the sample treated at 220 °C for 24 h was measured by thermogravimetry, and Ce₂(C₂O₄)₃·3.5H₂O is shown in Fig. S13 (ESI†). From the crystallographic analysis considering two main phases, the composition was estimated to be Ce₂(C₂O₄)₃·4.5H₂O (i.e., one water molecule more than determined by thermogravimetry). This was probably caused by lower occupancy of water molecules coordinated to cerium (type 4) and those within the pores (deeper discussion is provided in the ESI†).

Similar to that in,^32,33 hydrothermal crystallization was typically not a single-step process. Here, we reported the mechanism of hydrothermal conversion of Ce₂(C₂O₄)₃·10H₂O to oxide. The whole mechanism is expressed graphically in Fig. 6.

Fig. 6 Graphical illustration of the mechanism of hydrothermal conversion of Ce₂(C₂O₄)₃·10H₂O to CeO₂.

Hydrothermal dehydration and recrystallisation of oxalate take place prior to the conversion to oxide, which is responsible for the enormous change in morphology during this process. 3D ED was crucial for structural determination of nanocrystalline samples. The loss of water was linked with an increase in the denticity and coordination mode of the oxalate ligand, and elicited several structurally analogous phases. The change in the oxidation state of cerium, together with oxalate decomposition, seemed to be the final one-step process visible as a disintegration of the recrystallised phases to oxide. In addition to structural investigations, we tested the possibility of hydrothermal conversion of Ce-Gd oxalates to mixed oxides.

Conflicts of interest

There are no conflicts of interest to declare.

Acknowledgements

This study was funded by the Czech Science Foundation (GAČR) (project 20-20936Y). P.B. and J. R. acknowledge GAČR (project 21-05926X), the CzechNanoLab Research Infrastructure (MEYS CR, LM2018110) and the Operational Programme Research, Development and Education by the EU ESIF and MEYS CR (SOLID21 CZ.02.1.01/0.0/0.0/16_019/0000760).

Notes and references

1. A. Alemayehu, D. Zákutná, S. Kohúteková and V. Tyrpekl, J. Am. Ceram. Soc., 2022, 105, 4621.
2. F. Abraham, B. Arab-Chapelet, M. Rivenet, C. Tamain and S. Grandjean, Coord. Chem. Rev., 2014, 266–267, 28.
3. V. Tyrpekl, J.-F. Vigier, D. Manara, T. Wiss, O. Dieste Blanco and J. Somers, Low temperature decomposition of U(IV) and Th(IV) oxalates to nanograined oxide powders, J. Nucl. Mater, 2015, 460, 200.
4. L. De Almeida, S. Grandjean, N. Vigier and F. Patisson, Eur. J. Inorg. Chem., 2012, 4986.
5. O. Walter, K. Popa and O. Dieste Blanco, Open Chem., 2016, 14, 170.
6. K. Popa, O. Walter, O. Dieste Blanco, A. Guiot, D. Bouëxière, J. Y. Colle, L. Martel, M. Naji and D. Manara, CrystEngComm, 2018, 20, 4614.
7. L. Balice, D. Bouëxière, M. Cologna, A. Cambriani, J. F. Vigier, E. De Bona, G. D. Sorarù, C. Kübel, O. Walter and K. Popa, J. Nucl. Mater., 2018, 498, 307.
8. V. Baumann, K. Popa, O. Walter, M. Rivenet, G. Senentz, B. Morel and R. J. M. Konings, Nanomaterials, 2023, 13, 340.
9. J. Manaud, J. Maynadié, A. Mesbah, M. O. J. Y. Hunault, P. M. Martin, M. Zunino, N. Dacheux and N. Clavier, Inorg. Chem., 2020, 59, 14954.
10. J. Manaud, J. Maynadié, A. Mesbah, M. O. J. Y. Hunault, P. M. Martin, M. Zunino, D. Meyer, N. Dacheux and N. Clavier, Inorg. Chem., 2020, 59, 3260.
11. W. Liu, L. Feng, C. Zhang, H. Yang, J. Guo, X. Liu, X. Zhang and Y. Yang, J. Mater. Chem. A, 2013, 1, 6942.
12. G. Zhang, Z. Shen, M. Liu, C. Guo, P. Sun., Z. Yuan, B. Li, D. Ding and T. Chen, J. Phys. Chem. B, 2006, 110, 25782.
13. S. Benarib, N. Dacheux, X. F. Le Goff, J. Lautru, L. Di Mascio and N. Clavier, Dalton Trans., 2023, 52, 10951–10968.
14. H. Inaba and H. Tagawa, Solid State Ionics, 1996, 83, 1.
15. T. Montini, M. Melchionna, M. Monai and P. Fornasiero, Chem. Rev., 2016, 116, 5987.
16. K. Choi, W. Tong, R. D. Maiani, D. E. Burkes and Z. A. Munir, J. Nucl. Mater., 2010, 404, 210.
17. L. Palatinus, P. Brázda, M. Jelínek, J. Hrdá, G. Steciuk and M. Klementová, Acta Crystallogr., Sect. B: Struct. Sci., Cryst. Eng. Mater., 2019, 75, 512.
18. P. Brázda, L. Palatinus and M. Babor, Science, 2019, 364, 667.
19. P. Brázda, M. Klementová, Y. Krysiak and L. Palatinus, IUCrJ, 2022, 9, 735.
20. P. B. Klar, Y. Krysiak, H. Xu, G. Steciuk, J. Cho, X. Zou and L. Palatinus, Nat. Chem., 2023, 15, 848–855.
21. L. Palatinus, V. Petříček and C. A. Correa, Acta Crystallogr., Sect. B: Struct. Sci., Cryst. Eng. Mater., 2015, A71, 235.
22. V. Petříček, L. Palatinus, J. Plášil and M. Dušek, Z. Kristallogr., 2023, 238, 271.
23. V. Tyrpekl, P. Markova, M. Dopita, P. Brázda and M. A. Vacca, Inorg. Chem., 2019, 58, 10111.
24. R. Podor, N. Clavier, J. Ravaux, L. Claperde and N. Dacheux, J. Amer. Ceram. Soc., 2012, 95, 3683.
25. Dicerium trioxalate decahydrate, ICSD: 109637, Inorganic Crystal Structure Database, https://www2.fiz-karlsruhe.de/icsd_home.html, January 2018.
26. C. Tamain, B. Arab-Chapelet, M. Rivenet., X. F. Legoff, G. Loubert, S. Grandjean and F. Abraham, Inorg. Chem., 2016, 55, 51.
27. L. Palatinus and G. Chapuis, J. Appl. Crystallogr., 2007, 40, 786.
28. J. C. Trombe and J. Jaud, J. Chem. Crystallogr., 2003, 33, 19.
29. C. J. Hao and H. Xie, Acta Cryst., 2012, E68, m444.
30. A. Mer, M. Rivenet, L. De Almeida, S. Granjean and F. Abraham, Inorg. Chem. Commun., 2013, 31, 90.
31. M. Hernández-Molina, P. A. Lorenzo-Luis and C. Ruiz-Pérez, CrystEngComm, 2001, 16, 1.
32. L. Wang, R. Zhao, C. Wang, L. Yuan, Z. Gu., C. Xiao, S. Wang, X. Wang, Y. Zhao, Z. Chai and W. Shi, Chem. – Eur. J., 2014, 20, 12655.
33. L. Wang, R. Zhao, X. Wang, L. Mei, L. Yuan, S. Wang, Z. Chai and W. Shi, CrystEngComm, 2014, 16, 10469.

---

Footnote

† Electronic supplementary information (ESI) available. CCDC 2296196–2296199. For ESI and crystallographic data in CIF or other electronic format see DOI: https://doi.org/10.1039/d3nj04635d

---