Solvothermal Synthesis Routes to Substituted Cerium Dioxide Materials

Abstract

We review the solution-based synthesis routes to cerium oxide materials where one or more elements are included in place of a proportion of the cerium, i.e., substitution of cerium is performed. The focus is on the solvothermal method, where reagents are heated above the boiling point of the solvent to induce crystallisation directly from the solution. This yields unusual compositions with crystal morphology often on the nanoscale. Chemical elements from all parts of the periodic table are considered, from transition metals to main group elements and the rare earths, including isovalent and aliovalent cations, and surveyed using the literature published in the past ten years. We illustrate the versatility of this synthesis method to allow the formation of functional materials with applications in contemporary applications such as heterogeneous catalysis, electrodes for solid oxide fuel cells, photocatalysis, luminescence and biomedicine. We pick out emerging trends towards control of crystal habit by use of non-aqueous solvents and solution additives and identify challenges still remaining, including in detailed structural characterisation, the understanding of crystallisation mechanisms and the scale-up of synthesis.

1. Introduction

Cerium dioxide, ceria, is well known for significant applications in heterogeneous catalysis, where it is frequently used as a redox-active support [1,2]. This chemistry is exploited in numerous practical applications, exemplified by its use as the oxygen storage component in three-way catalytic convertors, as a support for precious metals for destruction of CO, NO_x and hydrocarbon emissions in automotive exhausts [3,4,5]. There are many emerging applications of ceria in catalysis and these include use in steam reforming [6], soot oxidation [7], water–gas shift [8], oxidation of volatile organics [9] and thermochemical water splitting [10]. All of these rely on the ease of reversible release of oxygen from the solid that has the cubic fluorite structure type, which is well known for its oxide ion migration properties, accompanied by the reversible reduction of Ce⁴⁺ to Ce³⁺.

Another important developing application of ceria is as an oxide-conducting electrolyte in solid oxide fuel cells [11]. Ceria-based solid electrolytes show promising conductivity in the temperature range 500–700 °C, and here, oxide ion vacancies are vital to optimise the hopping of ions to carry charge. Beyond heterogeneous catalysis and ionic conductors, biomedical applications of ceria are under development, such as in antioxidant defense [12,13,14]. The material also finds use in UV shielding [15,16,17], as an abrasive agent [18] and in humidity sensors [19].

For most real samples of ceria, the ideal formula CeO₂ is not found: in reality, there are typically some levels of oxide defects and hence the presence of Ce³⁺ to give the formula CeO_2−δ [20]. These intrinsic defects are inherently beneficial for the oxide ion migration that gives rise to many of the materials’ useful properties, since bulk oxygen mobility is enhanced, but in order to further enhance defects, partial substitution of Ce by a second cation, M, to give ternary materials Ce_1−xM_xO_2−δ is a common strategy. (Note that in this article, we refer to this replacement as substitution to indicate that a proportion of cerium is swapped for a second cation, rather than “doping”, which implies the addition of a second species.) If M is aliovalent (a trivalent cation or divalent cation), then a greater proportion of oxide ion vacancies are needed to balance the charge, Figure 1. However, even replacement of Ce by an isovalent cation (such as Zr⁴⁺ or Hf⁴⁺) engenders useful properties: the local distortion of the structure due to the size mismatch of the ions introduces high lattice strain [5]. This makes these solid solutions highly reducible with a significant proportion of the bulk cerium being reduced at moderate temperatures, with redox properties maintained over many cycles of reduction–oxidation. Hence, ceria-zirconia has become one of the most widely studied materials for redox catalysis, especially as a support for precious metals [21].

Coupled with substitutional chemistry is the effect of texture and crystal morphology on the properties of cerium oxide materials. For catalysis, the high surface area offered by small crystallites is desirable, but the shape and crystal habit can also affect properties, and for ceria, this includes the redox associated with oxygen loss and gain. Thus, the activity of ceria in catalysis may depend crucially on crystal habit, particularly when nanoscale crystallites are considered [22,23,24,25,26], and when the material is used as a support for precious metals, when specific crystal faces may offer favourable interactions. Similar relationships have been found for ceria-zirconia solid solutions [27,28].

In general, there is a need to develop reproducible methods for control of the synthesis of nanostructured ceria materials [29], and this article will review the recent literature that deals with the solvothermal synthesis of substituted cerium dioxide materials. We will focus on the literature from the past decade, since 2010, and for the previous literature, we direct the reader to a previous review article that considered the earlier literature on solvothermal crystallisation of various oxides of cerium [30].

2. Solvothermal Synthesis of Oxides

The term hydrothermal refers to the use of water as a reaction medium when heated at or above its boiling point in a closed vessel, where autogeneous pressure is introduced [31,32]. The method has been explored since the 1840s as a means of preparation of analogues of naturally found minerals, is well known for the synthesis of zeolites and as such has been intensely developed since the 1940s [33]. Preparative hydrothermal synthesis was also developed for the growth of large crystals of dense materials, such as quartz [34]. The term solvothermal was later introduced by Demazeau to encompass the use of any solvent heated above its boiling point [35]. In terms of the synthesis of oxides, solvothermal synthesis methods (in fact, commonly hydrothermal) have been applied for a variety of materials from binary phases to more complex multinary compositions. Here, the emphasis is often on the use of rather mild conditions (typically around 200 °C, where 10 s of bar pressure may be developed) for the formation of fine powders of a material [36,37,38,39,40]. Earlier review articles described the changes in water properties that occur when it is heated in sealed vessels, and these may include changes in viscosity and dielectric constant such that otherwise insoluble reagents are solubilised and dissolution–recrystallisation may take place [31,32]. Often, a mineraliser is added to the crystallisation, whose role is to aid the dissolution of reagents, and commonly alkali-metal hydroxides or fluorides are used.

The advantage of solvothermal synthesis over other methods is that crystallisation of the desired phase occurs directly from the solution, offering the potential for control of crystal habit directly at the point of synthesis, without any high temperature annealing necessary to induce crystallisation and that might destroy any intricate crystal form. We will thus focus one such one-step synthesis, without annealing or firing to bring about crystallisation. Commonly, however, calcination up to 500 °C is employed to remove any excess solvent before study, particularly when catalytic properties are to be investigated, but this is a considerably lower temperature method than employed in other methods, such as precipitation firing, combustion or thermal decomposition of precursor salts. Interestingly, early work on ceria synthesis in the 1980s used extreme hydrothermal conditions (up to 700 °C and 100 MPa) with the aim of growth of large crystals (100 μm in dimension) [41], but the focus since then has almost entirely been on the formation of fine powders, particularly with nanostructure. Under these milder conditions, synthesis is conveniently carried out in the laboratory using Teflon™-lined steel reactors.

3. Survey of Substituted Cerium Oxides

3.1. Ceria-Zirconia

As noted above, the replacement of Ce⁴⁺ by isovalent Zr⁴⁺ is one the most widely studied substitutions used to tune the properties of ceria. In fact, this gives a complete substitutional series, Ce_1−xZr_xO₂, from cubic cerium-rich compositions with x < 0.1, to tetragonal zirconium-rich phases with x > 0.1, and monoclinic for the most zirconia-rich materials, x > 0.88 [2]. Note that these are equilibrium crystal structures for materials prepared at high temperature, and the degree of tetragonality may be small for x < 0.5 and can also be difficult to detect when a diffraction profile is broadened due to the small crystal domain size, giving rise to pseudo-cubic materials. There is also the possibility of various metastable phases that depend on the composition and preparative method, and coupled with the possibility of atomic-scale inhomogeneity and nanoscale phase segregation, this makes determination of the precise structure of ceria-zirconia solid solutions challenging [42]. Even in the pseudo-cubic materials, the smaller ionic radius of zirconium compared to cerium induces local distortion of the structure, which is believed to be responsible for the enhanced oxygen storage capacity and more favourable reducibility, and this is also accompanied by enhanced thermal stability [5]. The precise redox and other catalytic properties depend on composition and which crystal phase is present, and here, the low temperatures of solvothermal synthesis offer the possibility of control over polymorphism, or the formation of mixed-phase samples with desirable properties. For example, Yang et al. used a solvothermal synthesis method from a mixed solvent of ethanol, N,N-dimethylformamide and ethylene glycol in the presence of lauric acid to form a ceria-zirconia with a Ce/Zr molar ratio of ~7/3; even in this ceria-rich composition, a mixed cubic (41%) and tetragonal (59%) sample was formed with distinct cubic–tetragonal interfaces on the nanoscale, Figure 2 [43]. Typically, to engineer such interfaces, high-temperature treatment above 1000 °C is needed to induce phase separation, and the advantage of the solvothermal approach is the high surface areas of the materials, which Yang et al. showed offered good activity for soot oxidation, ascribed to the presence of cubic–tetragonal interfaces, which significantly enhance the formation of oxygen vacancies from release of surface lattice oxygen.

In general, solvothermal synthesis allows surface properties of ceria materials to be modified, compared to materials prepared by conventional methods. Luciani et al. prepared Ce_0.75Zr_0.25O₂ by two hydrothermal routes, one with 1-propanol as a co-solvent and the other with urea as an additive, and X-ray photoelectron and electron paramagnetic resonance spectroscopies showed an increase in both surface and bulk Ce³⁺ content, corresponding to a higher oxygen mobility [44]. Wang et al. also used urea-assisted hydrothermal synthesis to prepare defect-rich ceria-zirconia materials, with the highest levels of defects seen for the zirconia-rich phases, and used these materials as catalysts for the oxidative dehydrogenation of ethylbenzene with CO₂ [45].

The benefits of solvothermal synthesis in controlling crystal morphology have been explored for ceria-zirconia materials, and Table 1 summarises some of the key examples. These illustrate how the crystal shape on the nanoscale can be adjusted from isotropic to anisotropic plates and rods by choice of reagents and solvent. Figure 3 shows an example of how morphology can be selected from facetted polyhedral to rod-shaped and plate-like with the crystal dimension on the nanoscale by choice of reagents, solution additives and temperature (and/or time of reaction) [46].

An orientated attachment mechanism was proposed by Liu et al., who formed hollow, truncated octahedral agglomerates of primary particles of cerium-rich compositions [47]. However, in general, the crystallisation mechanism and the origin of the shape control that is possible by synthesis have not been determined. Clearly, this is an area where further work is needed. Some observations have been made on the relationship between crystal morphology and the resulting properties. For example, CeO₂ nanorods were shown to have higher activity towards CO oxidation than CeO₂ nanoparticles, primarily because of the preferential exposure of the reactive {110} planes, but ZrO₂ doping promoted the reducibility of the nanoparticles more significantly than that of the nanorods [48].

Table 1. Summary of solvothermal conditions used to prepare various crystal morphologies of Ce_xZr_1−xO₂.

Crystal Morphology	Composition CexZr1−xO2	Solvent Temperature Time a	Reagents	Reference
Facetted polyhedral 3–5 nm diameter	0 ≤ x ≤ 1	6 M NaOH(aq) 150 °C 48 h	Ce(NO3)3·6H2O ZrO(NO3)2·xH2O	[49]
Hollow, truncated octahedral ~80 nm edge	x = 0.01, 0.1, 0.2	water 100 °C 18 h	CeCl3·7H2O ZrOCl2·8H2O Poly(vinylpyrrolidone) H2O2 formic acid tert-butylamine	[47]
Rods diameter ~8 nm length ~40 nm	0 ≤ x ≤ 0.2	10 M NaOH(aq) 100 °C 20 h	Ce(NO3)3·6H2O Zr(NO3)4·5H2O	[48]
Facetted <10 nm diameter	0 ≤ x ≤ 0.2	water 95 °C 8 h	(NH4)2Ce(NO3)6 Zr(NO3)4·5H2O urea	[48]
Polyhedral 10–15 nm diameter	x = 0.2	water 180 °C 12 h	Ce(NO3)3·6H2O ZrO(NO3)2·6H2O poly(vinylpyrrolidone) hydrazine	[46]
Rods diameter ~10 nm length ~20–70 nm	x = 0.2	6 M NaOH(aq) 100 °C 24 h	Ce(NO3)3·6H2O Zr(NO3)4·5H2O	[46]
Plates 6–12 nm diameter	x = 0.2	NH3(aq) 100 °C 24 h	Ce(NO3)3·6H2O ZrO(NO3)2·6H2O CTAB b	[46]

^a: A post-synthesis calcination ~600 °C was commonly applied to yield the studied product. ^b: CTAB = cetyltrimethylammonium bromide.

Table 1. Summary of solvothermal conditions used to prepare various crystal morphologies of Ce_xZr_1−xO₂.

Crystal Morphology	Composition CexZr1−xO2	Solvent Temperature Time a	Reagents	Reference
Facetted polyhedral 3–5 nm diameter	0 ≤ x ≤ 1	6 M NaOH(aq) 150 °C 48 h	Ce(NO3)3·6H2O ZrO(NO3)2·xH2O	[49]
Hollow, truncated octahedral ~80 nm edge	x = 0.01, 0.1, 0.2	water 100 °C 18 h	CeCl3·7H2O ZrOCl2·8H2O Poly(vinylpyrrolidone) H2O2 formic acid tert-butylamine	[47]
Rods diameter ~8 nm length ~40 nm	0 ≤ x ≤ 0.2	10 M NaOH(aq) 100 °C 20 h	Ce(NO3)3·6H2O Zr(NO3)4·5H2O	[48]
Facetted <10 nm diameter	0 ≤ x ≤ 0.2	water 95 °C 8 h	(NH4)2Ce(NO3)6 Zr(NO3)4·5H2O urea	[48]
Polyhedral 10–15 nm diameter	x = 0.2	water 180 °C 12 h	Ce(NO3)3·6H2O ZrO(NO3)2·6H2O poly(vinylpyrrolidone) hydrazine	[46]
Rods diameter ~10 nm length ~20–70 nm	x = 0.2	6 M NaOH(aq) 100 °C 24 h	Ce(NO3)3·6H2O Zr(NO3)4·5H2O	[46]
Plates 6–12 nm diameter	x = 0.2	NH3(aq) 100 °C 24 h	Ce(NO3)3·6H2O ZrO(NO3)2·6H2O CTAB b	[46]

^a: A post-synthesis calcination ~600 °C was commonly applied to yield the studied product. ^b: CTAB = cetyltrimethylammonium bromide.

The high surface areas offered by small particles of ceria-zirconia prepared by solvothermal routes are suited to act as supports for precious metal nanoparticles in heterogeneous catalysis, and, indeed, some of the examples reported in Table 1 were used for this purpose, with the precious metals added in a subsequent step [46]. Taking a different approach, Das et al. used a one-step approach to prepare such catalysts, by inclusion of ruthenium chloride in the hydrothermal reaction of cerium and zirconium chlorides, with hydrazine present as a reducing agent and CTAB as a structure-influencing agent [50]. After a mild calcination, the material consisted of Ce_0.5Zr_0.5O₂ nanorods with a fine dispersal of Ru nanoparticles; these proved effective catalysts for the partial oxidation of methane. This illustrates how solution chemistry offers the possibility to form multicomponent systems with some control of their form.

The use of ceria as a redox catalyst in solar reactors for the upgrading of carbon dioxide and/or water to carbon monoxide and/or hydrogen in a thermochemical splitting cycle has received some attention, and the design of active materials based on ceria-zirconia has proved possible using hydrothermal synthesis. Luciani et al. showed how Ce_0.75Zr_0.25O₂, when prepared in 1-propanol as a solvent at 120 °C with urea as an additive, yields materials with an increase in both surface and bulk Ce³⁺, leading to a higher oxygen mobility from and towards the catalyst surface, and, notably, even after several cycles of high-temperature treatment, these defects were maintained, leading to high activity for solar thermochemical splitting of water and carbon dioxide [44]. It should be noted that in such high-temperature applications, one possible disadvantage of materials prepared by a low-temperature solvothermal route is that any metastable composition or structure may irreversibly phase separate upon heating; this is an area in which future research is needed.

Solvothermal synthesis methods lend themselves to scaling for industrial manufacture, especially when continuous flow reactors can be designed to make use of superheated solvents. Rapid-mixing devices may also be used to induce crystallisation in confined geometries, allowing control over the crystal form. Ceria-zirconia solid solutions have been prepared in this way, offering the prospect of production of catalyst materials for real application. Kim et al. used supercritical water (250 bar, 550 °C) as a reaction medium and produced fine powders of Ce_1−xZr_xO₂ (x = 0.35, 0.5, 0.8) that were much less agglomerated than those formed by co-precipitation [51] and could be used to support Rh for catalytic reduction of NO by CO [52]. Auxéméry et al. produced nanostructured Ce_xZr_1−xO₂ (0 ≤ x ≤ 0.75) in flow reactors using an ethanol/water mixture (25 MPa, 400 °C). This yielded nanocrystallites that proved to be effective supports for Ni in dry methane reforming catalysis [53].

3.2. Transition Metal-Substituted Cerias

Through the addition of divalent (D²⁺) and trivalent (T³⁺) transition metal ions, compositions of the form Ce_1−xD_xO_2−x or Ce_1−xT_xO_2−x/2, respectively, can be expected. The oxygen site vacancies that this produces can modify lattice oxide mobility, making them attractive materials for use as electrolytes in solid oxide fuel cells, and may also be of benefit in oxygen storage materials. Some transition metals are of particular interest if they can form redox pairs along with Ce⁴⁺/Ce³⁺. This is because these materials are more capable of undergoing reduction and oxidation reactions, due to the possibility of cooperative electron exchange between the two metals. This is of relevance for applications in heterogeneous catalysis. An example of this effect can be found when doping CeO₂ with CuO. The Cu²⁺/Cu⁺ redox cycle has been shown to couple with the Ce⁴⁺/Ce³⁺ pair, helping to promote the redox features [54]. These materials have also been shown to have an increased number of oxygen vacancies associated with the Cu²⁺ ions [55].

Wang et al. demonstrated that Cu-doped ceria can be synthesised hydrothermally using Ce and Cu nitrates in NaOH, heated to 180 °C for 24 h [56]. The resulting material was shown to be single phase and indexed to the Fm$\stackrel{-}{3}$m cubic space group associated with ceria. TEM showed that the synthesised particles had sizes of between 4 and 5 nm. Other groups have shown that morphology control is possible: samples with 5 and 10 mol% Cu were made using chlorides and H₂O₂ heated at 180 °C, and 5 h of heating produced spheres with diameters of 60–80 nm, Figure 4, [57], while addition of PVP, ammonia and urea to the mixture with heating for 6 h gave spheres of ~100 nm in diameter [58].

Rood et al. were able to synthesise Cu-containing ceria nanotubes that were single phase up to 7 mol% Cu [59]. Their method involved using Ce(NO₃)₃ and CuSO₄ in NaOH solution heated to 100 °C for 10 h. In the same work, Cr-containing ceria nanotubes, up to a concentration of 5 mol%, were reported where a similar method was used, with Cr(NO₃)₃ as a precursor. These materials were tested as catalysts for both CO oxidation and NO reduction, with the Cu samples outperforming pure CeO₂ in terms of CO oxidation, but no overall improvement for NO reduction, with the Cr samples showing the opposite effect, an improvement in NO reduction, with no noticeable improvement in CO oxidation. XPS data suggested that Cu inclusion led to an increase in surface Ce³⁺, while Cr increased the percentage of Ce⁴⁺ on the surface. A Cu-Cr-co-doped sample showed an improvement in conversion of CO and NO over pure ceria, and only a dopant level of 1% Cu and 1% Cr was required to reach a similar CO oxidation to a sample of 7% Cu doping, suggesting a synergistic effect in the co-doped samples. Separately, Zhang et al. showed that the catalytic performance of hydrothermally prepared Cu-CeO₂ remains constant over six cycles of CO oxidation, suggesting that the material has the potential to be a stable, recyclable oxidation catalyst [60].

Kurajica et al. produced ceria samples doped with ~10 mol% Mn, Cu or Zn from sulfate precursors and 8 M aqueous NaOH heated at 120 °C for 16 h. Elemental analysis using electron energy loss spectroscopy (EELS) and energy-dispersive X-ray spectroscopy (EDS) analysis showed that Mn substitution had occurred to the expected concentration, but the Cu and Zn concentration was lower than expected [61]. These materials appeared single phase by powder X-ray diffraction (XRD) and had average crystallite sizes between 4.2 and 6.3 nm, with Mn doped > pure ceria > Cu doped > Zn doped, as confirmed using scanning transmission electron microscopy (STEM). Prior to testing the materials’ catalytic activity, the samples were fired to 500 °C to test for the presence of excess metal oxides on the surface of the materials. XRD of the materials after firing showed no extra peaks, suggesting that the as-made samples were unlikely to have amorphous metal oxides present. A trend that was noted, with regard to crystallite size, was that the samples with larger as-made crystallites showed only minor growth relative to pure ceria, whereas the smaller as-made crystallites showed greater growth. This resulted in the Cu and Zn samples having a lower specific surface area (SSA) than pure CeO₂, while Mn had a greater SSA. To test the catalytic activity of these samples, they were used to oxidise toluene. All samples reached 50% conversion at lower temperatures than pure CeO₂.

Syed Khadar et al. demonstrated a one-step hydrothermal synthesis of 2–8 mol% Co-doped ceria [62]. Their method used Ce(NO₃)₃ and CoCl₂ in water in the presence of Na₃PO₄, heated to 180 °C for 15 h. The role of the phosphate additive was not proposed, but these nanocrystalline samples consisted of cuboid particles, which were used in their as-prepared form for their antibacterial properties. Zhang et al. hydrothermally synthesised Mn-doped CeO₂ using Ce and Mn nitrates, in a mixture of ethanol, formic acid, ammonia and peroxide heated at 150 °C for 16 h [63]. The resulting nanospheres showed single-phase XRD patterns and had an average diameter of 130 nm, as seen from electron microscopy. EDS analysis shows that there is an even distribution of all elements across the particles, Figure 5. The catalytic performance of these materials was tested to measure CO oxidation. The sample with 7 mol% Mn was cycled six times, and after an initial increase in activity at low temperatures from first to second cycle, the activity remained stable at low temperatures. The initial change in activity was associated with the removal of impurities from the surface of the nanospheres.

Zhang et al. produced ceria nanocubes doped with 0.5–7 mol% Mn, using nitrates and NaOH and heating to 180 °C for 24 h [64]. They observed that the size of the cubes decreased with increasing Mn concentration, but no additional peaks were observed in the XRD patterns. Temperature-programmed reduction was used to demonstrate that these materials were more reducible than pure ceria nanocubes, and the materials were used for the selective oxidation of styrene tert-butyl hydroperoxide.

The work of Jampaiah et al. on manganese-, cobalt- and nickel-containing ceria nanorods (10 mol% transition metal added) demonstrated that Co doping increased the concentration of surface Ce³⁺ and that these materials showed greater oxidation of CO than CeO₂ produced using the same method [65]. The reason for the formation of extremely anisotropic crystal forms, Figure 6, is not apparent, since no solution additives were employed: the synthesis method simply used nitrate salts in sodium hydroxide solution heated at 100 °C. Song et al. used a similar synthesis approach to Fe- and Co-containing ceria samples and also found rod-like crystal morphologies; in this case, beneficial activity for the catalysis for the total oxidation of propane was found [66]. Xing et al. also reported rod-shape particles upon inclusion of Mn into CeO₂ by a direct hydrothermal crystallisation [67].

Liu et al. demonstrated the synthesis of 2 mol% Co-doped nanospheres from (NH₄)₂Ce(NO₃)₆ and CoCl₂ in acetic acid and ethylene glycol heated at 180 °C for 8 h [68]. When tested as a catalyst for CO oxidation, the Co-doped sample outperformed pure CeO₂. The same method was used to prepare Ni-containing samples, and these outperformed the Co samples at CO oxidation at lower temperatures, but as the temperature increased, the Co samples reached maximum conversion faster. The work of Du et al. focussed on Ni- and Co-doped ceria nanosheets, synthesised hydrothermally using nitrates, acrylamide, glucose and ammonia, heated to 180 °C for 96 h and then fired at 400 °C for 4 h, testing the materials’ ability to reduce NO [69]. It was found that the Ni-doped samples performed better, likely due to them having more oxygen vacancies than other comparable samples. Venkataswamy et al. produced CeO₂ doped at 5, 10 and 15 mol% with Cr [70]. This was achieved by mixing Ce and Cr nitrates in NaOH and heating to 100 °C for 24 h. They then fired the sample to 350 °C for 2 h. They found that the addition of Cr caused an increase in the catalytic oxidation of CO at lower temperatures up to 10 mol% The 15 mol% Cr-doped CeO₂ was shown to be less efficient than the 5 mol% sample. This is likely due to Cr being less effective as an oxidation catalyst. At lower amounts, the lattice distortion created by the dopant enhances the catalytic effect of Ce, but as the concentration increases, the loss of Ce from the sample eventually outweighs the benefit from the distortion.

Fe-substituted CeO₂ was synthesised by Phokha et al. by mixing nitrates in PVP and heating at 200 °C for 12 h [71]. They produced samples up to 7 mol% Fe that appeared single phase by XRD and Raman spectroscopy, but higher concentrations of Fe gave Fe₂O₃ as an impurity. Wang et al. were successful in producing Fe-doped ceria, up to a concentration of 20 mol% [72]. Their method involved using nitrates in ethylene glycol, heated to 180 °C for 28 h. They followed this up by heating the sample to 450 °C for 4 h. The resulting materials all matched the XRD pattern associated with CeO₂, rather than Fe₂O₃. Zheng et al. also produced Fe-doped ceria, forming the materials in a “nanoflower” morphology, Figure 7 [73]. This required mixing nitrates in urea, citric acid and ammonia, heating to 155 °C for 26 h and then firing to 500 °C for a further 3 h. The materials synthesised had an iron content of between 4 and 12 mol% and were applied for the selective oxidation of H₂S.

The use of additives in solution has been shown to result in unusual crystal forms. In the case of Ni-containing ceria, Liu et al. used a mixed water–ethylene glycol solvent with the addition of acetic acid to form particles with spherical-like nanocluster morphology that possessed hollow centres [74]. The porosity of the particles was dependent on the volume of Ni(NO₃)₂ solution used in synthesis, and materials with ~6.5 mol% Ni showed high catalytic performance in carbon monoxide oxidation. Mesoporous cerium–iron binary oxides prepared by a hydrothermal technique using CTAB as a template were tested as catalysts in methanol decomposition and total oxidation of ethyl acetate: low levels of Fe were dispersed in the fluorite structure, but higher levels gave clustering of Fe and ultimately phase separation above 50% Fe [75].

As well as properties in catalysis, transition metal-substituted cerias have been prepared by hydrothermal synthesis for other applications. Barbosa et al. synthesised Ce_1−xM_xO₂ (M = Mn, Cr, Co and Fe) from aqueous sodium hydroxide solution and studied the magnetic behaviour [76]. Radović et al. synthesised ceria materials with up to 5 mol% iron by crystallisation from aqueous ammonia solution, and optical properties were studied by spectroscopic ellipsometry [77]. Fe-substituted CeO₂ prepared from nitrate salts in aqueous ammonia was found to have a high level of oxide vacancies useful for photocatalytic ozonation [78]. Recently, Fe-substituted ceria has been examined for use in rechargeable lithium batteries: Ce_0.9Fe_0.1O₂ was prepared by a simple hydrothermal reaction of nitrates salts in aqueous sodium hydroxide, and its behaviour towards lithium insertion and removal was tested [79]. It was proposed that reversible reduction to metallic iron occurred on cycling, and an anode was fabricated to produce a high-powered lithium-ion battery with cycling stability and high power density; Figure 8 shows a proposal for the structural chemistry at play in this material, where the off-centring of Fe in the ideal CeO₂ structure provides space for Li insertion.

Dual substitution of more than one transition metal into ceria is also of interest to tune properties. Zhang et al. prepared Ce_1−x(Fe_0.5Zn_0.5)_xO_2−δ with x up to 0.018 at 200 °C and found a contraction of the lattice parameter and reduction in the crystallite size with increasing transition metal content, and XPS proved the presence of Fe³⁺ [80]. Dosa et al. prepared Ce_0.95Mn_0.025Cu_0.025O_2−δ alongside samples that contained only Cu or Mn using similar methods and found that the Mn-containing sample was more effective for soot oxidation and the Cu-containing sample showed the highest activity for CO oxidation: in this case, the co-substitution offered no benefit over CeO₂ itself [81].

A surfactant-assisted hydrothermal method was used by Zhu et al. to produce materials Mn_xCu_yCe_1−x−yO_2−δ [82]. Cetyltrimethylammonium bromide (CTAB) was added to aqueous solutions of metal salts and the pH was adjusted to 11 with NaOH before heating to 200 °C. Calcination at 550 °C yielded samples with high surface areas, >74 m²g¹, and with evidence of mesoporosity, as seen by gas adsorption measurements. The materials showed enhanced catalytic activity for CO oxidation.

It is also worth drawing attention to the possibility of forming composite materials by hydrothermal chemistry. For example, Ji et al. produced some striking hierarchically structured particles of CuO-CeO₂ composites, Figure 9 [83]. While powder XRD showed signals from both crystalline phases, implying limited inclusion of Cu into the CeO₂ structure, electron microscopy showed an intimate association of the two phases to give composite particles of agglomerated CuO needles with surface CeO₂. Although strictly not homogeneous single phases, this work illustrates the possibilities in the preparation of novel materials with unique morphologies by choice of reagents and conditions in hydrothermal reactions.

As described thus far, much of the work on transition metal substitution has focussed on the first-row series, and few examples of inclusion of the second-row and third-row elements in ceria by solvothermal routes have also been proven. The case of precious metals is important to consider since ceria is often used as a support for precious metal particles. Hiley et al. showed the inclusion of Pd²⁺ into the cubic fluorite structure of CeO₂ was possible for up to 20 mol% substitution from CeCl₃·7H₂O and PdCl₂ in aqueous NaOH/H₂O₂ [84]. As well as inducing charge-balancing oxide vacancies, the square planar preference of Pd²⁺ induces a structural distortion, as evidenced from EXAFS spectroscopy, Figure 10a. This material is unstable on heating in air above 700 °C, and the Pd is extruded to give metallic particles on CeO₂, which is, in itself, of interest for heterogeneous catalysis. The case of niobium substitution is an interesting situation as its preferred oxidation of +5 requires a different charge-balancing mechanism than substituents with lower charge. Hydrothermal synthesis from NbCl₅ and CeCl₃·7H₂O in aqueous NaOH solutions produced substituted ceria materials in which charge compensation was achieved by co-inclusion of Na⁺ to give materials of composition (Ce_1−xNb_x)_1−yNa_yO_2−δ (where x ≤ 0.30 and y ≥ ~x/3) [85]. Here, neutron pair distribution function analysis showed how the lower coordination number preference of Nb⁵⁺ leads to a local distortion of the structure, Figure 10b, which can be used to explain the enhanced oxygen storage properties of the material.

3.3. Main Group Substituents

The main group elements include those from the s and p blocks of the periodic table and cover a wide range of possible substituents for cerium dioxide, ranging from the hard cations of Groups 1 and 2 to the more covalently bound atoms from the p block. Hydrothermal synthesis provides a convenient way to explore this substitutional chemistry since phases that might otherwise phase separate during high-temperature synthesis may crystallise to yield unique compositions with unusual structural features and, potentially, properties.

From the alkali earths in Group 2, Ma et al. reported magnesium-containing materials Ce_1−xMg_xO₂ (x = 0.00, 0.02, 0.04, 0.06, 0.08) in the form of nanowires, prepared in a water–ethylenediamine mixture with NaOH as a mineraliser and CeCl₃·7H₂O and MgCl₂·2H₂O heated at 180 °C for 14 h [86]. Calcium- and strontium-containing CeO₂ with nominal compositions Ce_0.90Ca_0.10O_1.90, Ce_0.90Ca_0.05Sr_0.05O_1.90 and Ce_0.90Sr_0.10O_1.90 was prepared from Ce(III) nitrate and salts of Ca²⁺ and Sr²⁺ in aqueous NaOH at pH = 14, heated at 220 °C for up to 8 h [87]. Only a low level of incorporation of Sr into the ceria structure was found, explained by the larger size of Sr²⁺ compared to Ce⁴⁺. The small crystal size (12–16 nm) of the powders allowed effective sintering into dense ceramic pellets at 1350 °C for 5 h, and for the Sr-containing material, a secondary phase of SrCeO₃ was then seen, consistent with the fact that Sr is not effectively incorporated into ceria. The Ca-containing sample showed favourable ionic conductivity after sintering.

From Group 12, zinc-containing ceria with up to 20% of the cerium replaced by the substituent was prepared by a continuous hydrothermal synthesis approach using supercritical water as a reaction medium (450 °C and 24.1 MPa) and KOH as a mineraliser. The samples were formed as nanocrystalline powders with crystallite sizes less than 5 nm [88]. A 10% Zn-containing CeO₂ was found to be an effective catalyst for the CO₂ hydrogenation reaction, ascribed to the highest level of Ce³⁺ defects [89]. Although not an elemental substitution, the formation of ZnO-CeO₂ heterojunctions by hydrothermal synthesis has also been studied, where the purposeful growth of the two phases in intimate contact leads to superior visible-light photocatalytic performance [90].

Of the Group 13 elements, only indium has been introduced in ceria via hydrothermal conditions. Younis et al. published a synthesis of indium-containing CeO₂ that appears single phase up to 15 mol% In [91]. This involved mixing aqueous solutions of Ce and In nitrates and adding toluene, oleic acid and tert-butylamine. The reaction mixture was heated to 200 °C for 36 h and the solid product calcined at 180 °C for 1 h. The lattice parameters of the synthesised samples decreased with increasing In³⁺ concentration, attributed to the smaller ionic radius of In. The 5 and 10 mol% In samples outperformed pure CeO₂ in the photocatalytic degradation of organic dyes.

From Group 14, the inclusion of Sn and Pb into CeO₂ has been examined by solvothermal routes. Abbas et al. produced phase-pure samples on Sn-containing ceria up to 7% Sn, designed for use as a cytotoxic nanomedicine using Ce(NO₃)₃ and SnCl₂ in polyethylene glycol and NaOH heated to 95 °C for 12 h and then fired to 300 °C for 2 h [92]. The addition of Sn changed the morphology of the material from spherical nanoparticles to a heterogeneous nanowires–nanosheets structure. A contraction of the unit cell was reported, although the oxidation state of Sn was not measured. Shajahan et al. successfully synthesised single-phase Pb-containing CeO₂ using nitrates in NaOH, heated to 140 °C for 14 h followed by heat treatment at 350 °C for 3 h [93]. The samples appeared single phase under XRD in samples up to 12% Pb, and it was found that up to 6% Pb, the band gap of the material was lower, making it a better photocatalyst than pure ceria, while higher Pb content increased the materials’ bandgaps. The oxidation state of Pb was proved to be +2 using X-ray photoelectron spectroscopy.

From the elements in Group 15, there are several examples of how Bi³⁺ has been studied as a way of modifying CeO₂, and hydrothermal synthesis allows a convenient route to the introduction of high levels of substitution with up to 60% of the cerium replaced. The hydrothermal oxidation of Ce³⁺ in basic aqueous solution by NaBiO₃ at 240 °C provides a redox synthetic method, where the product phase contains Ce⁴⁺ and Bi³⁺, to form nanocrystalline powders (Figure 11) that show high reducibility, although the extrusion of the Bi metal will limit the practical application of the materials [94]. Bi-substituted CeO₂ has frequently been investigated for use in photocatalysis, although it has also been looked at for uses such as antibacterial activity: Frerichs et al. produced single, cubic-phase samples up to 20 mol% Bi by reacting Bi and Ce nitrates in NaOH and heating to 80 °C for 25 h [95]. Samples made using a higher level of Bi were a two-phase mix of the cubic CeO₂ structure and the monoclinic Bi₂O₃ material. Houlberg et al. synthesised materials using Bi(NO₃)₃ and (NH₄)₂Ce(NO₃)₆ in NaOH, heated to 200 °C for 20 h, and produced samples that were single cubic phase up to 60% Bi [96]. This study employed in situ X-ray diffraction to track the formation of the product and proved that mixed metal oxide phases were formed at the earliest stages of reaction even before heating under hydrothermal conditions. Other methods that have been used to create phase-pure samples up to ~10% Bi have been reported. Shanavas et al. used nitrates in NaOH, heated to 140 °C for 14 h and then firing at 350 °C for 3 h to remove surface impurities [97]. Veedu et al. used nitrates in ammonia and HCl, heated to 100 °C and then firing to 500 °C [98].

Unlike bismuth, which is found in the +3 oxidation state when included in ceria, the earlier member of Group 15, antimony, has been shown to be present in the +5 oxidation state: this is the case whether SbCl₃ or SbCl₅ is used under hydrothermal conditions in the presence of NaOH and H₂O₂ [99]. Here, like the case of niobium mentioned above, the co-substitution of sodium is found to balance the charge of the pentavalent substituent, resulting in chemical formulae (Ce_1−xSb_x)_1−yNa_yO_2−δ (where x < 0.4 and y ≥ x/3). As with bismuth, the antimony-containing samples are unstable in highly reducing conditions and phase separate to yield elemental antimony.

Finally, we consider a more unusual case, where nitrogen can partially replace oxygen in CeO₂ to give an oxynitride analogue of ceria. Xu and Li produced samples from Ce(NO₃)₃ and (NH₄)₂S₂O₈ in water heated to 200 °C for 24 h [100]. XRD recorded before and after a calcination at 500 °C for 2 h shows that the sample adopts the cubic ceria structure, with no additional peaks visible. Large excesses of nitrogen, from the ammonium persulfate reagent, were needed to achieve substitution, but the highest Ce/N ratio achieved in the product was 1:0.052, which is close to the solid solubility limit. The lattice parameters of all doped samples were larger than the expected value for pure Ce, due to the lattice distortion caused by the N³⁻ being larger than O²⁻. This was found to enhance the oxygen storage capacity of these materials.

3.4. Rare Earth-Substituted Ceria

The inclusion of rare earths (the other lanthanides and yttrium) into ceria has attracted much attention by various synthetic approaches, and these include numerous solvothermal routes. All of the available rare earths have been studied in this respect and a complete survey of materials prepared solvothermally in the past decade are included as Supporting Information (Table S1). Typically, up to 50% of the Ce can be replaced by a trivalent substituent, before impurity phases, such as Ln(OH)₃ or LnO(OH), are seen, particularly for the rare earths with smaller ionic radii [101,102,103]. Most synthesis is performed under hydrothermal conditions from aqueous solutions of metal salts: typically nitrates of the trivalent cations, but occasionally chlorides or acetates, and in the case of cerium, Ce(IV) salt (NH₄)₂Ce(NO₃)₆ has also been used. In the simplest cases, the pH is increased by addition of NaOH, KOH or ammonia, and temperatures spanning 80–250 °C have been reported to produce a solid product. In many cases, the solid formed is crystalline, but often a subsequent heat treatment up to 500 °C is applied, simply to remove any surface-bound species. Under these conditions, the typical morphology of CeO₂ itself is of cube-shaped crystallites, Figure 12 [104], which may appear spherical when the size is 10 nm or less. The size of the crystals may depend on pH, temperature and time of reaction.

Upon introduction of small amounts of a second lanthanide, the cube-shaped crystal is usually maintained, but it has been reported that at higher substituent levels, and under otherwise identical reaction conditions, the crystal morphology becomes increasingly rod-shaped, with a distribution of both crystal forms at intermediate compositions. This is exemplified by the case of praseodymium, where the same phenomenon has been reported independently by several different research groups: [105,106,107,108] at 25% Pr substitution and above, the rod-like crystal morphology is detected, and an example is shown in Figure 13. Praseodymium may also be found in the +4 oxidation state, and this has been inferred from the redox properties of some of these samples [106,108] and detected by X-ray spectroscopy in others [105,107]. This, however, does not explain the formation of the rod-like crystallites. Interestingly, Mendiuk et al. found for Ln = Pr, Sm, Gd, Tb, at higher Ln substitutions, the formation of Ln(OH)₃ rod-like crystals [109]. It may well be possible that the formation of these anisotropic crystallites has a seeding effect in the subsequent growth of the oxide, or that when a second step of firing is applied, even at moderate temperature, the dehydration of the hydroxide occurs with retention of crystal morphology. Consistent with this idea, Hong et al. separately showed the anisotropic crystallites of phase-pure (Ce,Gd)(OH)₃ could be dehydrated to the mixed oxide on heating at just 400 °C with crystal morphology maintained [110]. Yoshida and Fujihara found that for Sm substitution, hydrothermal treatment at 100 °C prior to heat treatment gave rod-shaped crystals, even for very low Sm content, which may imply the initial formation of a hydroxide phase at mild conditions [111].

Regarding crystallisation mechanisms, Mendiuk et al. proposed that for the cube-shaped crystallites, their assembly occurred by an oriented attachment mechanism, owing to the observation of a bimodal crystallite size distribution [102,109]. A careful study of Sm-CeO₂ nanocubes using spatially resolved electron energy loss spectroscopy on the TEM showed the that presence of the Ln³⁺ substituent did not affect the Ce³⁺ concentration at the surface of the crystallites [112].

The addition of solution additives has been explored as a way to modify the crystal morphology of lanthanide-substituted materials. In one of the simplest cases, Ke et al. added NaCl to the crystallisation of a set of lanthanide-substituted cerias (Ln = La-Lu with 10 mol% substituent) from chloride precursors in 12 M NaOH and found nanowires ~5 nm in diameter with an aspect ratio of ~500, Figure 14 [113]. In the processing of these samples, heating to only 300 °C was performed, which does imply that the ionic strength of the solution may play a role in the crystallisation process.

An additive that has been well established to affect crystal morphology in hydrothermal formation of cerium oxides is Na₃PO₄·12H₂O. Xu et al. found for yttrium substitution that octahedral crystals were produced but with surface nanorods with increasing addition of sodium phosphate [114]. The same octahedral crystal morphology was found by Małecka et al., who studied the formation of ytterbium-substituted CeO₂ [103,115], Roh et al., who examined europium substitution [116], Bo et al. for Er-CeO₂ [117] and Yang et al., who prepared erbium-substituted materials, some of which were co-substituted with lithium, Figure 15, where morphology is maintained even after a high-temperature annealing of 800 °C [118]. Małecka also observed that rod-shaped crystals were found at high concentrations of Na₃PO₄·12H₂O [115]. The octahedral morphology, bounded by {111} faces, is known to be a stable form for large crystals of CeO₂ itself [119], and while earlier work showed the effectiveness of Na₃PO₄ as a mineraliser for growth of octahedral nanocrystals of CeO₂, the mechanism of this precise morphological control has not been established, beyond discussion of how it might control pH and affect the dissolution–recrystallisation processes leading to crystallisation [120].

Urea has been investigated as an additive for when crystallising ceria with substituents Pr, Gd [121] and Tb [122,123], and this may be observed to have the general effect of forming small, ~10 nm, crystallites that are often agglomerated. Muñoz et al. used microwave heating (see below) and produced Gd-CeO₂ materials that consisted of spherical assemblies of the primary particles, Figure 16. The combination of urea and sodium citrate was studied by Xu et al. for La, Nd, Sm and Eu substitution, who found “broom-like” agglomerates of needle-like crystals [124]. Urea is known to decompose under hydrothermal conditions into ammonium and cyanate, and this has been speculated to control the nucleation by pH control, in the case of ceria itself leading to smaller crystallites at higher urea concentrations [125].

Other additives that have been investigated for control of crystal morphology of substituted cerias include hexamethylenetetramine [126,127,128] and tetramethyl ammonium hydroxide [129,130,131]. The role of these additives has not been determined, and often simply the outcome of their influence on crystal morphology is described. Sato et al. speculated that tetramethylammonium ions cap the growing crystal faces of Gd-CeO₂ to avoid agglomeration and ripening, thus forming cuboidal crystals of only 4 nm in dimension [131].

Mixtures of additives have also been investigated, such as polyvinylpyrrolidone, HCl, KClO₃ and N,N-dimethylformamide [132], octadecylamine, ethanol and ethylenediamine [133] or 6-amino hexanoic acid and tert-butylamine [134], and non-aqueous solvents have been studied, such as ethylene glycol [135,136,137,138], ethanol [139] and triethylene glycol [140]. In some cases, the choice of combination of additive and solvent has led to some striking crystal morphologies; for example, the combination of epigallocatechin-3-gallate in ethanol as a solvent allowed the formation of hierarchical nanostructures of Eu-containing CeO₂ with flower-like shapes of intergrown plates, microns in thickness, Figure 17 [141].

As well as conventional heating, microwaves have been used to facilitate solvothermal crystallisations of rare earth-substituted cerias. The advantage here lies in the short reaction times, from a few hours down to just a few minutes [102,103,115,121,122,123,128,142,143,144,145,146].

Continuous flow synthesis has also been used for some rare earth-substituted ceria materials. Slostowski et al. produced La- and Pr-substituted ceria in a custom-built, continuous flow reactor in supercritical water, up to 400 °C, and formed crystallites less than 10 nm in size, observing that higher temperature allowed a higher level of element substitution [147]. Xu et al. used supercritical water at 396 °C and rapid mixing, ~29 s, with solutions of metal salts and KOH to yield Gd-CeO₂ [148]. Their apparatus, Figure 18, had a further two stages of heating points before allowing the product to be tapped as a slurry. Depending on pH, the crystallites produced were between 6 and 40 nm in size with a polyhedral or octahedral shape.

Applications of rare earth-substituted ceria materials span the types of fields expected for these materials, from oxide ion conductors in solid oxide fuel cells, where the enhanced sinterability of fine powders into ceramics is beneficial [136,137,143,149,150], to catalysts for CO and soot oxidation, where the high surface and exposure of specific crystal faces may offer some benefits [103,105,107,108,113,115,132,133,135,150,151,152,153]. Photocatalytic properties are of increasing focus for degradation of pollutants or organic transformations, and some of the rare earth-substituted materials have been screened for these applications [101,124,141,154].

Photoluminescence is a property that has commonly been used to characterise rare earth-substituted ceria materials [111,114,117,144,155,156,157,158,159,160], while Eu-substituted materials are of particular interest for their luminescent properties. Eu-CeO₂ materials prepared by solvothermal routes have provided some useful materials: this includes in applications such as solar cells, Figure 19 [116], and detection of fingerprints [141]. The use of solvothermal conditions to prepare nanocrystals of ceria materials has allowed specimens to be prepared for study of their toxicity: Dunnick et al. used gadolinium-substituted samples and examined their properties in vivo in rats [129,130].

In terms of practical applications, the formation of composites is of relevance, and the solvothermal routes have allowed some useful possibility of accessing novel materials with practical applications. Pr-doped CeO₂ nanorods were introduced into a Sr-doped LaMnO₃–Y₂O₃-stabilized ZrO₂ composite by an in situ hydrothermal growth process: this gave an enhanced performance of the material for use as an electrode in solid oxide fuel cell applications [161]. Jiang et al. produced gadolinium-substituted ceria crystallites on a graphene support, by including the support material in the hydrothermal reaction [162]. This was tested for application in non-aqueous lithium–oxygen batteries, and a favourable catalytic effect on Li₂O₂ formation and decomposition was found. In a similar manner, Er-CeO₂ was loaded onto graphene oxide directly via in situ hydrothermal crystallisation, and then in a second hydrothermal step, CuO was co-loaded; these composite materials showed enhanced photocatalytic reduction of CO₂ to methanol compared to the individual components [163]. Xu et al. used polystyrene microspheres as templates to produce hollow microspheres constructed from agglomerated crystallites of Y-substituted CeO₂: these showed enhanced activity for the photocatalytic decomposition of acetaldehyde [158].

3.5. Multi-Element Substitutions

Given the wide range of chemical elements that can easily be added to ceria using the very mild conditions offered by solvothermal approaches, it is not surprising to find that multi-element inclusion has been increasingly explored as a way of fine-tuning properties such as oxygen storage or photocatalysis and perhaps also introducing the possibility of multifunctionality in the solid materials. This includes combining elements of similar chemistry that might provide some cooperative effect on resulting properties, such as more than one of the rare earths [164,165,166,167,168]. Mixing elements from different parts of the periodic table is also increasingly being explored, and one important example is the inclusion of other rare earths in ceria-zirconia solid solutions that may further tune redox properties for heterogeneous catalysis. For example, by addition of both Zr and Pr to ceria, via a hydrothermal route, it was shown that praseodymium contributes positively towards the soot oxidation reaction, but it has an adverse effect on CO oxidation over the same catalysts, as compared to pure ceria [169]. The inclusion of both Nd and La to ceria-zirconia, to give a phase with 40% cerium, 50% zirconium, 5% lanthanum and 5% neodymium, via a mild hydrothermal crystallisation, gave materials with enhanced low-temperature reducibility and thermal stability [170]. Materials in the CeO₂-Y₂O₃-ZrO₂ system have been prepared under solvothermal conditions. Wang et al. found that the activation energy of crystal growth is significantly dependent on the CeO₂ concentration [171], and continuous flow synthesis (410 °C, 27 MPa) has been used to form ultrafine stabilised tetragonal zirconia particles [172], while a wide range of nanoscale crystal forms, with intricate morphologies, were possible by inclusion of surfactants, Figure 20 [173]. In the even more complex system ceria-zirconia-yttria-lanthana, high-surface area nanostructured materials have been formed with the aid of lauric acid and dodecylamine [174].

Europium has been added to ceria-zirconia in low concentrations (<10%) using aqueous ammonia as a reaction medium and potassium oleate as an additive to produce nanocrystalline materials with photoemission and photoluminescence properties [175,176]. Hydrothermal synthesis using poly block copolymer surfactants provides high-surface La-containing ceria zirconia [177].

Mixed rare earth ceria materials with other substituents have been prepared via hydrothermal methods, such as Sr,Gd-ceria [178] and Bi,Gd-ceria [179,180], and sintered into ceramics for solid oxide fuel cell applications.

4. Summary, Challenges and Future Directions

Our survey of the past decade of research on solvothermal synthesis methods for producing cerium dioxide materials has shown how elements from every part of the periodic table can be introduced into the cubic fluorite lattice of CeO₂. Our earlier review 10 years ago found relatively few examples of substitutional chemistry, with the main emphasis being on ceria-zirconia solid solutions and rare-earth substituents, in all cases using water as the majority solvent [30]. This work has now been broadened to encompass an extensive set of element substitutions, including mixtures of elements, and using a variety of solvents, solvent mixtures and solution additives. As with other oxide materials, the advantage of the solvothermal method lies in the accessing of novel substitutional chemistry and in the control of crystal form, both of which are difficult when higher temperatures are employed, such as in classical solid-state chemistry, or combustion methods, or even in simple precipitation routes that require a firing step to induce crystallisation [181]. The solvothermal synthesis method typically yields materials in a crystalline state, even if a mild heat treatment (up to 500 °C) may be used to remove any surface-bound hydroxide, solvent or additive, particularly when the materials are to be used for applications in heterogeneous catalysis. Much of the research reported is towards practical applications and as well as classical fields of research, such as in redox catalysis, photocatalysis and oxide ion conductivity, novel directions are being explored and, indeed, enabled by the availability of nanocrystalline forms of ceria. This includes applications as components in solar devices, as electrodes in rechargeable batteries and as catalysts for thermochemical water splitting and has allowed properties such as luminescent sensing, magnetism and toxicity to be studied.

The nanoscale crystal morphologies are important where high surface areas are needed but also provide highly sinterable powders for forming ceramics, such as electrodes for solid oxide fuel cells. In this respect, it is important to consider the level of control over crystal morphology under solvothermal reaction conditions and whether this has truly been demonstrated with any degree of predictability. As this review has shown, a wide variety of solution additives and solvents have now been employed, and, certainly, in some cases, enough examples are provided, independently by different research groups, to identify certain conditions that yield specific crystal morphologies. For example, Na₃PO₄·12H₂O has been used to form octahedral-shaped crystals for CeO₂ itself, as well as for various substituted versions. However, the mechanism of its action has not been uncovered, beyond speculation of control of the solution pH. The formation of nanorod morphologies is often related to the formation of lanthanide trihydroxides, implying either a seeding effect of Ln(OH)₃ which forms early in the synthesis, or thermal decomposition yielding oxides with morphological control. In many cases, the role of the solution additives (along with the choice of solvent) is unknown and many authors reported the outcome of the use of the additive without any consideration of its role.

From our review, three trends are emerging in tuning the properties of materials for applications which make use of the full advantages of solvothermal chemistry, and these are ripe for future development.

(1) Multi-element substitution is now increasingly being used to adjust properties, and this has been extended from mixtures of rare-earth substituents to combinations of, for example, rare earths in ceria-zirconia to mixtures of transition metal ions. This includes compositions that might be expected to phase separate in alternative synthesis approaches, particularly those that use an annealing to induce crystallisation. Here, computer simulation, which has long been used to understand the effect of M³⁺ substituents in ceria [182], may provide important guiding principles in narrowing down likely candidate combinations of substituting elements.

(2) The formation of composite materials is being tuned to introduce dual functionality or enhanced properties beyond the individual components: this includes the purposeful growth of mixed-phase materials where one oxide is grown in intimate proximity to a second. This may give interesting properties from the interfaces present between the two phases but may also yield morphologies not possible by other routes, and the example of CeO₂-CuO illustrates this.

(3) The use of in situ growth on templates or substrates allows fabrication of materials in a form suited for application. For example, impregnation of a porous support material with a reagent solution, followed by hydrothermal crystallisation, provides a convenient method for the formation of a device, and an example of ceria nanorods grown within a ceramic matrix for an electrode in solid oxide fuel cells illustrates this concept. The same idea could be used to produce coating or films of an immobilised material. This is of importance in emerging applications of ceria materials, such as in electrocatalysis, where the role of a support is crucial to determine properties [183].

In consideration of the directed synthesis of active materials for applications, there are some outstanding issues that are apparent from the literature we have reviewed:

(1) A detailed atomic-scale characterisation of materials is needed to understand fully structure–property relationships to then justify and predict choices of elemental substitution. The vast majority of work reviewed above has relied on the use of powder X-ray diffraction patterns and electron microscopy to prove sample identity and purity, in some cases supplemented by Raman spectroscopy as an additional fingerprint. Very few works have performed an in-depth analysis of the local atomic structure in the substituted materials, which is likely to be highly distorted from the ideal fluorite structure, when the coordination preferences of cations are taken into consideration: for example, the Jahn–Teller distortion expected for Cu²⁺, or the asymmetric environment expected for Bi³⁺ in an oxide. In multi-element substitutions, the issue of local atomic clustering, i.e., the homogeneity of element mixing, must be considered. Even for the case where an isovalent substitution is made, such as in ceria-zirconia solid solutions, there are complex structure–composition phase diagrams, and nanoscale phase segregation is possible. Here, methods such as EXAFS spectroscopy and pair distribution function analysis must be employed to determine these structural subtleties that are hidden in conventional diffraction measurements, especially when powder diffraction profiles are broadened by a small crystallite domain size.

(2) Crystallisation mechanisms, including the mode of operation of crystal habit modifiers mentioned above, must be understood in order to plan synthesis effectively and to develop tailored synthesis to target materials with a desired structure and properties. The use of high-energy X-rays has proved highly beneficial in this respect to track the evolution of a crystalline material from within a realistic reaction vessel as a function of temperature and time [96]. This yields quantitative kinetic information and also the possibility of observation of any intermediate phases, and an example of such a study of Bi-CeO₂ is provided above. If pair distribution functions can be measured in situ, then this provides local structure information, without the requirement for a crystalline structure, and such an approach has already been applied for CeO₂ itself where nucleation was observed from previously unknown cerium dimer complexes [184]. The more widespread use of such methods will lead to the mechanistic detail needed for predictive synthesis, and this will include a proper understanding of the role of additives in modifying crystal habit. Computer simulation may also play a part in such efforts.

(3) Scale-up of synthesis for applications is crucial if materials’ properties are to be put into real use. As was highlighted above several times, the use of continuous flow reactors has been the focus of development for a number of years, where a flow of reagents is rapidly mixed with a heated solvent, often under supercritical conditions, to give rapid nucleation and an output stream of a slurry of a product. This has already been put to use on a commercial scale, and cerium oxide products are available commercially [185]. Much more development needs to be conducted on developing scaled synthesis when the use of additives and crystal habit modifiers is used, and with the much more complex chemical compositions that are now being investigated; here, mixing and inclusion of additives must be carefully considered.

(4) The choice of precursors in the preparation of any functional material is of crucial importance since contaminant ions can modify the properties of the solid. In the case of ceria, alkali metal cations can act as a poison when used in selective catalytic oxidation of NO_x or NH₃ [186,187]. On the other hand, the presence of sodium can have a beneficial effect in other applications; for example, it can act as a promoter when ceria is used for ethane oxidation [188], and its inclusion in ceria-zirconia can lead to enhanced low-temperature reproducibility [189]. Similarly, chloride ions can influence the surface properties of ceria, in some cases enhancing interaction with supported species [190], while in others having a detrimental effect by the formation of CeOCl that results in loss of surface area [191]. It is also the case that the choice of precipitant can affect the sintering behaviour of ceria powders [192]. In an industrial setting, extensive washing of powders must be avoided, and therefore to achieve reproducibility in the properties of ceria materials prepared from solution, future work must consider the role of surface contaminants that arise from the choice of chemical reagents; this is true whether the contaminants have a detrimental or a beneficial effect.

(5) The toxicity of nanocrystals to humans and other organisms is a growing concern, and ceria has attracted some attention in this respect, especially given its biological activity arising from its antioxidant properties. With a growing range of specimens available from solvothermal chemistry, each with their own distinctive chemistry from the substituent cations, clearly consideration of biotoxicity must be the focus of greater attention. As well as studies of the action of nanoparticles in biological systems, their transport into the human body is another important field of research, whether by inhalation or through the skin. This has recently been examined for CeO₂ itself [193], but there is much further work to be conducted in this respect.

Finally, we note that in the case of CeO₂ itself, a recent article by Xu et al. extensively reviewed the various solution methods for synthesis, including solvothermal methods, and attempted to classify the various influences on crystal growth with the aim of having some predictability in the control of the crystal form, Figure 21 [194]. While the addition of a substituting cation, and indeed habit-modifying additives, introduces further complexity, it is highly likely that the same principles put forward will apply to the wider family of cerium oxide materials. This accumulation of experimental evidence leads to the possibility of controlled synthesis of materials with highly tuned properties for applications.

5. Conclusions

Solvothermal synthesis routes to substituted cerium oxides provide versatile conditions for the preparation and, ultimately, manufacture of functional materials for use in established and emerging applications. The accumulated literature points towards a move to control of the chemical composition and crystal form, with particular emphasis on the nanostructure. We have emphasised developing trends in targeting complex compositions with multi-element replacements, and in the preparation of composite phases, including in situ growth onto substrates to fabricate devices for application. There are also a number of challenges in developing future targeted synthesis of novel materials, which include more in-depth structural characterisation of existing materials to provide proper structure–property relationships, and in situ studies of crystallisation to provide the formation mechanism with a firmer basis. Finally, scale-up of synthesis requires further consideration, bearing in mind the complex mixtures of reagents being used to tailor the crystal form.