Rare Earth Based Anisotropic Nanomaterials: Synthesis, Assembly, and Applications

Abstract

Rare earths (RE) refer to the lanthanide elements La–Lu together with Sc and Y. Conventionally, they have found applications in phosphors, magnets, catalysts, fuel cell electrodes/electrolyte. Here in this chapter, we discuss the synthesis, assembly and applications of rare earth based anisotropic nanomaterials. Regarding synthesis, the anisotropic growth behaviors of these nanocrystals are predominantly governed by their own unique crystal structures. Yet for wet-chemistry synthetic methods where a number of parameters could be finely tuned, the addition of particular coordination agents, templating agents or mineralizers has proven to be an effective way to direct the growth of nanocrystals into some anisotropic structures. Regarding applications, anisotropic nanomaterials, compared to their isotropic counterparts, often exhibit distinct properties. For example, the luminescence of anisotropic nanomaterials can display polarization and site-specific features. As for rare earth nanomaterials as magnetic resonance imaging (MRI) contrast agents, the high surface area of anisotropic nanostructures can give rise to superior performances. And for catalysis applications, anisotropic nanomaterials expose rich, highly active facets, which is of great importance for facet-selective catalytic reactions. In the chapter, we will start with introduction of the crystal structures of rare earth compounds, then briefly summarize the synthesis and assembly of rare earth anisotropic nanomaterials, and discuss their properties and applications in three realms, namely, luminescence, magnetism and catalysis.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

5.1 Crystal Structures of Rare Earth Compounds

5.1.1 Rare Earth Oxides

As rare earth element s show strong affinity to oxygen, they can be easily synthesized in the air, and rare earth oxides enjoy the most investigation among its solid compounds. Rare earth oxides mostly take the form of sesquioxides (RE₂O₃) due to their trivalency, while some elements can also exist as divalent or tetravalent ions in their oxides, thus forming monoxides, dioxides or mixed valence oxides with a complex crystal structure [1]. Europium and ytterbium monoxides are the only two rare earth monoxides that are stable in solid form, and they are readily oxidized in the air. Praseodymium and terbium usually exist as both trivalent and tetravalent ions in their oxides, thus forming mixed valence oxides with non-stoichiometric formulas of Pr₆O₁₁ and Tb₄O₇, respectively. Cerium exists exclusively as tetravalent ions in its most stable oxide, CeO₂ [2].

Rare earth sesquioxide s are the most common category of rare earth oxides. Depending on their cationic radii, rare earth sesquioxides can be most stable in hexagonal phase (A phase), monoclinic phase (B phase) or body-centered cubic phase (C phase) under ambient conditions (as shown in Fig. 5.1). In both hexagonal (A phase) and monoclinic (B phase) sesquioxides, a rare earth cation is coordinated by seven oxygen ions, with six oxygen ions forming an octahedron around the cation. In hexagonal sesquioxides, the last oxygen ion rests on one of the octahedron’s faces, while in monoclinic sesquioxides, the last oxygen ion locates exceptionally far from the cation. Light rare earth sesquioxides such as La₂O₃, Ce₂O₃, Pr₂O₃ and Nd₂O₃ usually exist in hexagonal phase due their large cationic radii, while Sm₂O₃ usually takes monoclinic phase due to its smaller cationic radius. In body-centered cubic (C phase) sesquioxides, a rare earth cation is coordinated by six oxygen ions, resembling a cubic fluoride structure with two anions removed. The cation occupies two different kinds of vacancies formed by differently removed surrounding anions. Y₂O₃ and all lanthanide oxides heavier than Sm₂O₃ (Eu₂O₃ to Lu₂O₃) exist in C phase.

Fig. 5.1

Crystal structures of rare earth oxides. a Hexagonal sesquioxide. Blue spheres denote RE³⁺; red sphere O²⁻. b Monoclinic sesquioxide. Blue, green and orange spheres denote RE³⁺ in three different coordination environments; red spheres O²⁻. c Cubic sesquioxide unit cell projected along the c axis. Blue and red spheres denote RE³⁺ in two different coordination environments; red spheres O²⁻. d Cubic fluorite CeO₂. White spheres denote Ce⁴⁺; red spheres O²⁻

EuO and YbO have a cubic halite structure and the space group Fm3m with the cation coordinated by six oxygen ions. The monoxides can also be mixed with the corresponding sesquioxides in 1:1 ratio to form the mixed valence oxides of Eu₃O₄ and Yb₃O₄, respectively.

As one of the most studied rare earth compound, CeO₂ is a notable exception to the sesquioxide norm. CeO₂ has a cubic fluorite structure and the space group Fm3m, with the cation coordinated by eight oxygen atoms. As the valence of cerium can alternate between Ce(III) and Ce(IV), Ce⁴⁺ in the lattice can be reduced to Ce³⁺ under reductive conditions, accompanied by the generation of oxygen vacancies that serve as highly active catalytic sites and oxygen storage centers. Despite the formation of Ce³⁺, CeO₂ can maintain its cubic fluorite structure under intense reduction. This excellent chemical stability combined with its thermal stability makes it a versatile catalyst and oxygen sensor under rigorous conditions [3]. PrO₂ and TbO₂ also possess a cubic fluorite structure similar to CeO₂. The sesquioxides of praseodymium and terbium adopt a body-centered cubic structure as described above, enabling them to mix with their cubic fluorite dioxides to form mixed valence oxides.

5.1.2 Rare Earth Fluorides

Binary rare earth fluoride REF₃ has two kinds of crystal structures: trigonal fluocerite structure and orthorhombic β-YF₃ structure. Trigonal REF₃ has a hexagonal phase with the space group of P-3c1. The rare earth cations lie on a twofold rotation axis. There are nine fluorine atoms near each rare earth atom. The distance of RE³⁺ and F⁻ ions is between 0.242 and 0.264 nm. There are two F⁻ ions far from the RE³⁺ ions with a distance of 0.301 nm. Rare earth atom is nine-coordinated via 9 fluorine atoms. Each F⁻ anion bonds with three RE³⁺ cations in a three-bridged configuration. β-YF₃ belongs to orthorhombic structure, Pnma space group. In this system, it is a three-dimensional (3D) network molecule with a F⁻ anion as a bridged atom. Each F⁻ anion bonds with three Y³⁺ as three bridges. There are nine F⁻ anions coordinated with Y³⁺, eight of them close to the Y, while the other one is a little far. The nine F⁻ anions form a triangular prism coordination polyhedron (shown in Fig. 5.2).

Fig. 5.2

Crystal structures of trigonal REF₃ and orthorhombic REF₃ (built by CERIUS2 software (http://www.accelrys.com/cerius2)

REF₂ has a cubic fluorite structure, Fm3m space group. REF₂ can produce some complex fluorides. EuF₂ and CsF can form perovskite CsEuF₃. EuF₂ and MgF₂ can form orthogonal SmMgF₄ and EuMgF₄. EuF₂ and AlF₂ can form cubic solid solution (Eu, Al) F_2.00-2.44 and tetragonal system EuAlF₅. EuF₂ can form EuZrF₆, EuSiF₆, EuThF₆ with ZrF₄, SiF₄ and ThF₄ respectively.

5.1.3 Rare Earth Complex Fluorides

According to the coordination states of rare earths, the rare earth complex halide A_mREX_n (A = alkali and/or alkaline earth metal; RE = rare earth; X = halide) compounds can take the following forms: (1) AREX₄, a fourfold coordination of rare earth cations is accompanied by twofold or threefold coordination of the alkali and/or alkaline earth atom. (2) A₂REX₅, rare earth cations are in a fivefold coordination configuration, energetically more stable than a fourfold one. (3) A₃REX₆, the fivefold and sixfold coordination of rare earth cations are energetically competitive. For both A₂REX₅ and A₃REX₆, each coordination state can be realized in various forms that differ in detail but are close in energy.

Common structure types of the AREF₄ include tetragonal anti-scheelite (LiYF₄), hexagonal (NaNdF₄), trigonal (KErF₄), and orthorhombic (KCeF₄) type [4–6]. In addition, a high-temperature modality exists which has A⁺ and RE³⁺ cations at the Ca²⁺ sites of cubic CaF₂ structure with statistical distribution.

As shown in Fig. 5.3, α-NaREF₄ has a cubic phase with the space group of Fm3m. β-NaREF₄ has a hexagonal phase (space group: P-6 or P6₃ /m). In the cubic-phase α-NaRF₄, Na⁺ and RE³⁺ cations are randomly distributed in the cationic sublattice. In the hexagonal-phase β-NaREF₄, the cation sites contain three types: a onefold site occupied by RE³⁺, a onefold site occupied randomly by 1/2 Na⁺ and 1/2 RE³⁺, and a twofold site occupied randomly by Na⁺ and vacancies [7].

Fig. 5.3

Crystal structures of α-NaREF₄ and β-NaREF₄ built by CERIUS2 software (http://www.accelrys.com/cerius2)

5.1.4 Rare Earth Oxyhalides

Rare earth oxyhalide s have two structures: tetragonal phase (P4/nmm space group) and trigonal phase (R3m space group), as shown in Fig. 5.4 [8, 9]. In the tetragonal phase, RE³⁺ cations occupy the position with C _4v symmetry. The tetragonal phase is thermodynamically less stable, and tends to form nonstoichiometric compounds with more complex structures. In the hexagonal phase, RE³⁺ cations occupy the position with C _3v symmetry. At high temperature, the hexagonal phase can transform into cubic structure.

Fig. 5.4

Crystal structures of trigonal REOF, tetragonal REOF and cubic REOF built by CERIUS2 software (http://www.accelrys.com/cerius2)

5.1.5 Rare Earth Phosphates and Vanadates

For rare earth orthophosphate s (REVO₄), there exist four different phases, mostly depending on the cationic radius of rare earth element: monazite, xenotime, rhabdophane, and churchite [10]. Rare earth orthovanadates (REVO₄) have two isomorphic phase structures, namely, monoclinic (m-) monazite type, and tetragonal (t-) zircon type. The phase selectivity for REVO₄ relies on the radius of RE³⁺, only zircon type is observed for all rare earth orthovanadates excluding LaVO₄ and CeVO₄ that can crystallize in both monoclinic and tetragonal phased structure [11, 12].

5.2 Synthesis, Assembly of Rare Earth Based Anisotropic Nanomaterials

5.2.1 One-Dimensional (1D) Nanostructures

One-dimensional (1D) nanostructure s include nanorods, nanowires, and nanotubes. Many solid materials naturally grow into 1D nanostructures, the growth direction of which is usually dominated by the crystallographic symmetry [13–17]. However, rare earth oxides often have isotropic structures; for example, CeO₂ crystals has a face-centered cubic (fcc) structure (space group: Fm3m). Therefore, CeO₂ nanocrystals tend to present high-symmetry morphologies, such as cubes and octahedra. To obtain low-symmetry morphologies, templating or capping agents are generally needed to differentiate the surface energy of each facet and to direct anisotropic growth and the formation of 1D structures. Therefore, various templating methods have been established for preparing CeO₂ nanorods and nanowires. Porous anodic alumina membranes (AAM) are commonly used in such synthesis as hard templates by virtue of their modulated pore diameters, ideally rod-shaped pores and exceptionally narrow size distribution [18–20]. CeO₂ nanowires were successfully synthesized with an improved sol–gel method within the hexagonally ordered nanochannels of the AAM. Ce³⁺ cations and corresponding anions reacted directly inside the nanochannels to form intermediates with 1D nanostructure s. After post-treatment, the intermediates were transformed into the arrays of CeO₂ nanowires within the pores of AAM template [20].

CeO₂ nanorods can also be synthesized with solvothermal/hydrothermal methods [21–23], which generally feature controlled composition and morphology, low aggregation and high crystallinity, because of the diffusion-controlled growth in solvent media in a closed system. Yan’s group prepared high-aspect-ratio ceria nanorods through a facile hydrothermal treatment without addition of any templates [24]. Using Ce(NO₃)₃ as the precursor, hexagonal Ce(OH)₃ intermediate was formed in the presence of OH⁻ anions (6 mol L⁻¹). After drying, white Ce(OH)₃ nanocrystals were converted into yellow CeO₂·nH₂O in ambient atmosphere, without any change in their shape. Very recently, Yan’s group have obtained lanthanide-doped CeO₂ nanowires through a modified hydrothermal route in the presence of NaOH and NaCl without any additional capping agents (as shown in Fig. 5.5) [25]. Such preparation process can be described with the following equations:

$$ {\text{Ce}}^{ 3+ }_{{({\text{aq}})}} + x{\text{Ln}}^{ 3+ }_{{({\text{aq}})}} + y{\text{OH}}^{ - }_{{({\text{aq}})}} \to {\text{Ce}}\left( {\text{OH}} \right)_{ 3}{:}x\% {\text{Ln}}_{{({\text{s}})}} $$

(5.1)

$$ {\text{Ce}}\left( {\text{OH}} \right)_{ 3}{:}x\% {\text{Ln}}_{{({\text{s}})}} + y{\text{O}}_{{ 2({\text{g}})}} \to {\text{CeO}}_{ 2}{:}x\% {\text{Ln}}_{{({\text{s}})}} + z{\text{H}}_{ 2} {\text{O}} $$

(5.2)

Fig. 5.5

TEM (a, c), HRTEM (b, d), and HAADF-STEM EDS elemental mapping (e, f) images of CeO₂:Nd (a, b, e) and CeO₂:Lu nanocrystals (c, d, f). (Reprinted with permission from [25]. Copyright 2013 American Chemical Society)

In this approach, Ce³⁺ precursor was dissolved in a high-concentration alkaline solution, forming one-dimensional Ce(OH)₃ intermediates with hexagonal structure. Upon calcination at 300 °C for 1 h, the nanowires transformed into CeO₂:Ln nanowires without prominent change in shape.

Zink’s group reported a facile template-free hydrothermal method to obtain hierarchical architectures of CeO₂ nanorods in an acidic medium (as shown in Fig. 5.6) [26]. Na₃PO₄, as a mineralizer, plays a key role in controlling the morphology of CeO₂ nanostructures by adjusting the electrostatic potential and surface energy of CeO₂ nanorods. They investigated the mechanisms of nucleation and crystal growth process by varying the cerium precursors and concentration, concentration of phosphate, reaction temperature, pH of the reaction mixture, seeding and secondary treatment. Aspect ratios of these CeO₂ nanorods and nanowires can be precisely controlled and finely tuned from 4.1 to more than 100. The single-crystalline CeO₂ nanorod s/nanowires grew along the [211] direction by an “oriented attachment” mechanism, followed by Ostwald ripening. Both phosphate and chloride ions were critical to the synthesis of one-dimensional CeO₂ nanorods and nanowires in both primary and secondary hydrothermal process. The strong coordination effect of phosphate with cerium ions was the driving force for the “dissolution” of nanorods under highly acidic condition. Chloride ions acted as a stabilizing agent in maintaining the rod/wire structure.

Fig. 5.6

TEM analyses of CeO₂ nanorods obtained from (a–c) primary and (d–f) secondary synthesis. (Reprinted with permission from [26]. Copyright 2012 American Chemical Society)

Rare earth oxide nanotubes have received extensive interests since the synthesis of carbon nanotubes in 1991 because they might offer unique properties and lead directly to original technological applications [27]. Han et al. first reported a facile method to synthesize CeO_2-x nanotubes by a hydrothermal route with mild reaction conditions [28]. By elevating the precipitation temperature and prolonging the aging time, a large amount of nanotubes were formed. Nevertheless, this method had a few disadvantages, such as the time-consuming procedures, excessive by-products and difficulties for purification.

Yang’s group developed an effective hydrothermal method to synthesize CeO₂ nanotubes with large cavities, thin walls and high quality [29]. A facile oxidation-coordination-assisted dissolution process of the hydroxide precursors was involved. First, Ce(OH)₃ nanowires were used as the starting materials; then the outer wall of Ce(OH)₃ nanostructures was partially oxidized by air. The oxidation-coordination-assisted dissolution process was limited inside the 1D nanomaterials. Finally, the ceria nanotubes with large cavities and thin walls were formed.

Chane-Ching et al. reported a general method for the synthesis of nanostructured materials with large surface area through the self-assembly of functionalized nanoparticles in a liquid-crystal phase [30]. Surface-functionalized ceria nanoparticles with the protonated amino acid interact directly with the ethylene oxide (CH₂CH₂O) groups of the copolymer. Based upon the weak interactions between the surfaces of the nanoparticles and the template, organized hexagonal arrays of CeO₂ nanoparticles are obtained and the symmetry of the arrays was preserved after calcination at 500 °C.

Seal et al. developed a very simple, green chemical route to guide self-assembly and time-dependent evolution of ceria nanoparticles into ultralong polycrystalline ceria nanorods [31]. By freezing and subsequent aging of an aqueous solution, CeO₂ nanoparticles trapped in voids (which form in ice) gradually evolve into polycrystalline nanorods by localized oriented attachment process. MD simulations also predicted that the type of morphology evolution observed for the ceria nanostructures can be driven by the dimensional constraints.

Yan’s group obtained four types of self-organized monolayer patterns (i.e., isolated particles, short chainlike (pseudo-1D aggregated), pearl necklace-like (1D aggregated), and dendritic (pseudo-2D aggregated) alignments during a PVP-assisted alcohothermal synthesis of ceria nanocrystals [32]. Possible self-organization mechanism of the nanosized CeO₂ colloids from isolated particle to 1D and 2D aggregated alignments were associated with the delicate balance of the attractive and repulsive forces caused by the adsorbed hydroxyls, PVP, and alkylammonium cations on the surfaces of CeO₂ nanocrystals during the irreversible evaporation of the solvent.

Using block copolymer Pluronic P123 as the template and ceric nitrate and zirconium oxide chloride as the precursors, highly ordered 2D hexagonal mesoporous Ce_1–xZr_xO₂ solid solutions with a 2D hexagonal (p6mm) structure were synthesized via a novel direct and reproducible method as shown in Fig. 5.7 [33]. The overall synthesis strategy is based on a sol-gel process combined with evaporation-induced self-assembly in ethanol, without any extra reagents to adjust the pH of the sol-gel reaction. A series of mesoporous Ce_1–xZr_xO₂ with different Ce/Zr ratios can be obtained under the optimized conditions, such as appropriate precursors, surfactants, and reaction temperature.

Fig. 5.7

TEM images of the mesoporous Ce_1–xZr_xO₂ (x > 0.5) recorded along the a [001] and b [110] orientations. The inset in (a) is the corresponding FFT (fast Fourier transform) diffraction image, and the one in b is the corresponding SAED pattern. (Reprinted with permission from [33]. Copyright 2007 American Chemical Society)

Nanosized rare earth fluorides and complex fluorides usually adopt plate-like morphologies due to their crystal structures, yet under certain conditions (for example, in the presence of shape-directing surfactants), 1D nanostructures can also be obtained. Chen et al. employed a precipitation method and prepared orthorhombic and hexagonal phase EuF₃ nanocrystals with various shapes, including nanospheres, nanobundles, nanorods, nanowires and nanoplates [34, 35]. Through a hydrothermal route, Li et al. obtained monodisperse LaF₃ and NaLaF₄ nanorods with different aspect ratios [36]. They also developed a solvothermal method to prepare NaYF₄ nanocrystals by controlling the reactant concentration, temperature and duration, rod-like nanocrystals with good crystallinity and different sizes and aspect ratios were obtained [37–40]. Using YCl₃, NaF and EDTA as the starting materials, Qian et al. obtained spindle-like YF₃ with a hydrothermal method [41]. Interestingly, Zhao et al. discovered hexagonally ordered arrays of NaYF₄ nanotubes in a solvothermal reaction [42]. Yan’s group developed a general thermolysis method to prepare rare earth fluorides and complex fluorides in high-boiling solvents (such as octadecene, oleic acid and oleylamine) using trifluoroacetate salts of rare earths and alkali metal; with proper ratio of Na to rare earth, hexagonal-phase NaYF₄ nanorods were obtained [43]. Using a similar approach, they obtained ultrathin EuOF nanowires with diameter below 2 nm [44]. The nanowires are highly flexible and can form ordered superstructures in a parallel configuration on substrates. A proper ratio of Eu precursor and surfactant oleic acid was found to be crucial for obtaining such wire-like structures.

For orthophosphates of light lanthanides (La–Gd), hexagonal rhabdophane and monazite phase rod-like or wire-like products could be obtained (as shown in Fig. 5.8) [45, 46]. With a temperature of l80–240 ºC, the products become the monazite phase with the morphology strongly affected by the acidity [47]. LaPO₄:Ln³⁺(Ln³⁺ = Ce³⁺, Tb³⁺) and LaPO₄:Ce³⁺,Tb³⁺/LaPO₄ core/shell nanowires have been synthesized on a large scale through a direct precipitation in a water-based system under moderate conditions without the assistance of any surfactant, catalyst, or template [48]. The diameters of the obtained nanowires are about 15 nm, and the lengths range from hundreds of nanometers to several micrometers. For heavy lanthanides (Ho–Lu) and Y, a synthesis temperature as low as 70 ºC is required to obtain hydrated monoclinic churchite phase, for example, YPO₄ nanowires [49]. Higher temperature leads to tetragonal zircon phase. Usually, a particle-like morphology is obtained with acidic conditions. When a chelating agent like EDTA is introduced, the hydrated hexagonal nanorods of YPO₄ could be obtained [50]. For intermediate lanthanide s (Gd, Tb, Dy), hexagonal, tetragonal, and monoclinic phases may coexist. Therefore, results are more complex. Typically, rhabdophane type TbPO₄ nanorods form at low temperature and zircon type TbPO₄ nanocubes form at high temperature, both with acidic mother liquors [46]. Hexagonal DyPO₄ nanorod bundles form at low temperature [51]. Usually, the rhabdophane products of light rare earths convert to monazite products, while the churchite phase converts to zircon products with heat treatment [52].

Fig. 5.8

a TEM image of LaPO₄ nanowires; b HRTEM image of a single 23 nm LaPO₄ nanowire; c TEM image of SmPO₄ nanowires; d TEM image of a uniform 50 nm SmPO₄ nanowire; e HRTEM image of a single SmPO₄ nanowire taken from the highlighted section with inset showing the electron diffraction pattern; f TEM image of EuPO₄ nanowires; g TEM image of a uniform 80 nm EuPO₄ nanowire; h HRTEM image of a single EuPO₄ nanowire taken from the highlighted section with inset showing the electron diffraction pattern; i TEM image of CePO₄ nanowires; j TEM image of PrPO₄ nanowires; k TEM image of NdPO₄ nanowires; and l TEM image of GdPO₄ nanowires/nanorods. (Reprinted with permission from [46]. Copyright 2003 Wiley-VCH.)

Surfactants, chelating agents, as well as block copolymers are used to control the size and shape of rare earth phosphate nanocrystals. Bu et al. reported a P123-assisted hydrothermal synthesis of CePO₄:Tb single-crystalline thin nanorods of 10–12 nm in width. The surfactant Pluronic P123 was found to play a crucial role both to improve luminescence properties and nanorod homogeneity [53]. When the pH value of reaction system is adjusted to below 1.0, uniform spindle like nanowire bundles of LaPO₄ could be obtained [54]. Xing et al. reported the synthesis of uniform CePO₄ nanorods by reaction of aqueous [(CTA)₃PO₄] micelles with [Ce(AOT)₃] reverse micelles prepared in isooctane [55]. Ghosh et al. reported the synthesis of LaPO₄:Er,Yb and LaPO₄:Er@YbPO₄ nanorods using a reverse micelles system [56]. Li et al. reported an OA-assisted solvothermal routes in mixed solution of water and ethanol for the synthesis of uniform hexagonal REPO₄·nH₂O nanocrystals at 140 ºC (as shown in Fig. 5.9) [57]. The synthesis in high-boiling solvents gives out a number of advantageous features for the synthesis of doped REPO₄ nanocrytals. Yan et al. reported a series of high-quality dispersible REPO₄ nanocrytals with shapes of nanopolyhedra, quasinanorods, nanorods, and nanowires which are synthesized at 180–260 ºC in oleic acid and oleylamine solvents via a limited anion-exchange mechanism [58]. The assembly behavior during the synthesis of rare earth orthophosphates has been observed. Li et al. reported a one-pot synthesis of CePO₄ nanowires attached to CeO₂ octahedral micrometer crystals [59]. Ce(NO₃)₃·6H₂O and (NH₄)₂HPO₄ are used in a molar ratio of 2:1, pH value is tuned to be about 1, and the treatment temperature is 180 ºC. The photoluminescence properties of CePO₄ nanowires attached to CeO₂ octahedral micrometer crystals are enhanced strongly in comparison with pure CePO₄ nanowires.

Fig. 5.9

TEM image of REPO₄·nH₂O nanocrystals: a Dy, b Er, c Ho, d Tm, e Yb, and f Lu. (Reprinted with permission from [57]. Copyright 2007 Wiley-VCH)

For anisotropic orthovanadates, the first example reported is LaVO₄. Different from YVO₄ that only crystallizes in tetragonal phased structure, LaVO₄ exhibits both monoclinic and tetragonal phased structure. Therefore, phase modulation is the first task for the synthesis of LaVO₄-based materials. Wet chemistry routes always show superiority on the selective synthesis of nanomaterials with desired phase structure and morphology [60]. Sun et al. contributed several reports on this subject. In their synthesis, they used NaVO₃ and La(NO₃)₃ as the starting agents. The phase structure and morphology of LaVO₄ are dominated by aqueous pH value. Below pH 3.5, LaVO₄ crystallizes into irregular shaped m-phase nanoparticles. The t-phase nanorods only form in the pH range of 4.5–6.0. It is noticed that at early stage, NaVO₃ and La(NO₃)₃ solution were mixed to form a yellow suspension of m-LaVO₄. Then, the crude precipitates were transformed into t-LaVO₄ nanorods during a 48 h hydrothermal treatment at 180 °C with pH value of 4.5–6.0. The obtained nanorods had average diameters of 20 nm and lengths close to 100 nm along [001] direction. Besides the direct precipitation combining hydrothermal treatment, introducing chelating agent is also a powerful tool to restrict crystal growth, regulate the crystallization and get anisotropic materials. Yan et al. firstly used EDTA-assisted hydrothermal method to prepare t-LaVO₄ nanorods. The transformation from monoclinic phase to the metastable tetragonal structure dramatically enhanced the luminescence intensity of LaVO₄:Eu, validating the correlation between structure and properties of materials [61]. They subsequently found that weak coordinating ligands, like sodium acetate or sodium citrate, only promote the crystallization of m-LaVO₄, also benefits the homogeneous doping of other rare earth ions in LaVO₄ lattice [61], but have no effect on the polymorph selection for the tetragonal phase. Also, by varying the ratio of [EDTA]/[La³⁺], the aspect ratio of the LaVO₄ nanorods/nanowires also could be tuned (as shown in Fig. 5.10) [62, 63]. Employing a reverse microemulsion system with SDS (sodium dodecylsulfate) as the surfactant, Fan et al. obtained t-LaVO₄ nanowires and nanotubes after 170 °C treatment [64]. By altering the SDS concentration, the morphology of LaVO₄ was variable between nanowires and nanorods, and the aspect ratio along the [001] direction could reach 100.

Fig. 5.10

TEM and HRTEM images of as-obtained t-LaVO₄ nanocrystals (a and b) and t-LaVO₄ nanorods (c and d). (Reprinted with permission from [63]. Copyright 2005 American Association)

5.2.2 Two-Dimensional (2D) Nanostructures

Recently, 2D nanomaterial s, such as nanoplates and nanosheets have attracted broad attention because of their special structures and properties coming from the quantum confinement of electrons [65–68]. However, the facile synthesis of 2D rare earth oxide nanomaterials remains a challenge [69, 70]. Yan’s group synthesized a series of dispersible rare earth oxide nanocrystals using rare earth benzoylacetonates [71], acetylacetonates, or acetates [72] as precursors by thermal deposition method (as shown in Fig. 5.11). Interesting, these nanocrystals exhibit a striking ability to self-assemble into large-area nanoarrays. By tuning the polarity of the dispersant, the RE₂O₃ nanocrystals could be aligned to form “side-to-side” or “face-to-face” self-assembly nanoarrays on carbon-coated copper grids [71, 72]. Rare earth oxides with different morphologies, including nanopolyhedra, nanoplates, and nanodisks, were synthesized in oleic acid/oleylamine mixed solvents. The nature of metal cations as well as the selective adsorption of the capping ligands plays a critical role in the shape-controlled growth process.

Fig. 5.11

a Formation of rare earth oxide nanopolyhedra, nanoplates, and nanodisks. b–d TEM images of the as-obtained Eu₂O₃: b OA/OM = 1:7, 310 °C, 1 h (inset HRTEM image of an Eu₂O₃ nanoparticle; scale bar 10 nm); c OA/OM = 3:5, 310 °C, 20 min; d OA/OM = 3:5, 330 °C, 1 h. (Reprinted with permission from [71]. Copyright 2005 John Wiley and Sons)

Murray et al. synthesized ultrathin 2D ceria nanoplates via a thermal decomposition method in the presence of mineralizers (sodium diphosphate) [73]. Many reaction parameters, such as precursor ratio, concentration, and reaction time, were varied to tune the morphology of nanoplates. Sodium diphosphate proved to be critical for accelerating the crystallization process and controlling the morphology of ceria nanocrystals. The obtained CeO₂ nanoplates exhibit much higher oxygen storage capacity than that of 3D CeO₂ nanomaterials prepared with other methods due to their higher theoretical surface-area-to-volume ratio and desirable (100) surfaces.

Generally, metal oxides with a cubic crystal structure, including ceria, inherently do not show any preference for 2D anisotropic growth behaviors. However, Xia et al. prepared ultrathin, single-crystalline ceria nanosheets with a thickness of approximately 2.2 nm and lateral dimension up to 4 μm by a simple aqueous route (as shown in Fig. 5.12) [74]. A balance of the anisotropic hydrophobic attraction and the electrostatic interactions may govern the spontaneous 2D self-organization of initially formed small ceria nanocrystals. Subsequently, the nanosheets were formed through a recrystallization process. Due to the quantum size effect associated with the extremely small thickness, these ceria nanosheets hold great potentials for fundamental studies.

Fig. 5.12

Characterization of ceria nanosheets. a, b TEM images of nanosheets with different sizes. c TEM image of a self-folded nanosheet. d Tapping-mode AFM image and the height along the line shown in the AFM image. e Powder XRD pattern. f HRTEM image and the corresponding FT pattern (inset). (Reprinted with permission from [74]. Copyright 2010 John Wiley and Sons)

In order to study the role of active sites in catalysis, Xie’s group successfully synthesized three-atom-layer thin CeO₂ sheets with about 20 % pits occupancy using an “ultrafast open space transformation” strategy [75]. An intermediate of ultrathin CeCO₃OH sheets with thin 2D structures were obtained using sodium oleate, CeCl₃ and NH₃·H₂O as the starting materials. Subsequently, freestanding ultrathin CeO₂ sheets with numerous surface pits were synthesized after direct heating of the CeCO₃OH intermediate at 400 °C for 2 min in air. By contrast, they prepared clean three-atom-thick CeO₂ sheets without surface pits by elevating the heating temperature to 550 °C and shortening the duration to 10 s.

Rare earth fluorides and complex fluorides often take 2D shapes. Li et al. synthesized uniform LaF₃ nanoplates with Yb³⁺/Er³⁺ and Yb³⁺/Ho³⁺ dopants in a hydrothermal autoclave [76]. Yan’s group employed the thermolysis approach with the single precursor La(CFCOO)₃ and obtained highly monodisperse LaF₃ triangular nanoplates [77]. By altering the solvent, the nanoplates can self-assemble into highly ordered superlattices on substrates in a side-by-side or face-to-face manner (as shown in Fig. 5.13). They subsequently obtained monodisperse ultrathin LaOCl nanoplates (ca. 4.0 nm in thickness) using a similar method [78], and found that when different long chain amines (oleylamine, hexadecylamine and octadecylamine) were used as the surfactant, the LaOCl nanoplates could self-organize into face-to face and side-by-side superstructures, nanowire-like and nanorod-like superstructures. The different interaction strengths of amine capping ligands were presumed to be responsible for the different behaviors in self-assembly configurations. Apart from self-assembly, the same group also employed Langmuir-Blodgett technique to fabricate dense monolayers of a variety of differently shaped hydrophobic nanostructures [79], including sphere-like NaYF₄ and LiYF₄ polyhedra, 1D NaYF₄ nanorods, 2D LaF₃ triangular nanoplates, CaF₂ square nanoplates and EuF₃ hexagonal nanoplates. The effects of size, shape symmetry on the assembly kinetics were discussed.

Fig. 5.13

TEM images of the a side-by-side and b face-to-face superlattices of LaF₃ nanoplates. Insets are the SAED patterns. (Reprinted with permission from [77]. Copyright 2005 American Chemical Society)

Examples of 2D rare earth phosphates and vanadates are relatively rare, since they usually take rod-like or spindle-like shapes due to their crystal structures. Li et al. reported an oleic-acid-assisted solvothermal routes in mixed solution of water and ethanol for the synthesis of uniform REPO₄·nH₂O nanocrystals at 140 ºC [57]. Hexagonal nanoplates with sub-100 nm diagonal diameter could be obtained, which self-assembled on substrate into ordered hexagonal monolayer s. Qian et al. synthesized a series of rare earth vanadates and compared the effect of different chelating ligands including trisodium citrate, sodium tartrate and sodium malate [80]. They found when the amount of citrate or tartrate was higher than twofold molar ratio of Y³⁺, a preferential growth would arise, yielding YVO₄ nanoplates with a donut-like morphology (as shown in Fig. 5.14). This method can be extended to vanadates of heavy lanthanide elements, such as DyVO₄ and ErVO₄. Li et al. prepared LaVO₄:Eu square nanoplates using a hydrothermal method with the assistance of oleic acid as surfactant [81]. The monodisperse 40 nm × 40 nm × 10 nm nanoplates tend to self-organize into ordered 2D arrays. They also synthesized a series of t-REVO₄ colloidal nanocrystals by using the same method [82]. They found the preferential growth direction of nanocrystals was decided by the steric repulsions of active points on crystal facets. LaVO₄, CeVO₄, and PrVO₄ nanocrystals revealed a square sheet-like morphology since their [100] and [010] directions had smaller rejection. With reducing ionic radius, YVO₄, NdVO₄, SmVO₄, and EuVO₄ showed similar square morphology, but the corners and edges of the nanoplates were broken or damaged. From EuVO₄ to LuVO₄, the breakage on nanocrystals edges became more pronounced and the proportion of irregular nanocrystals gradually increased. For TmVO₄ and LuVO₄, the selectivity of crystalline growth direction almost disappeared.

Fig. 5.14

SEM images of monodisperse YVO₄ assemblies obtained from 24 h hydrothermal reaction at 140 °C. Different chelating agents were used. a, b 2:1 molar ratio of citrate/Y³⁺; c, d, 3:1 molar ratio of citrate/Y³⁺; e, f, 4:1 molar ratio of citrate/Y³⁺; g, h, 2:1 molar ratio of tarate/Y³⁺. (Reprinted with permission from [80]. Copyright 2009 Wiley-VCH)

5.3 Applications of Rare Earth Based Anisotropic Nanomaterials

5.3.1 Luminescence Properties and Optical Applications

5.3.1.1 General Introduction of the Luminescence Properties of Rare Earths

Featuring abundant energy levels of 4f electron configurations, trivalent rare earth ions activated nanocrystals exhibit unique and fascinating luminescence properties [83]. Compared with semiconductor quantum dots and organic dyes, rare earth activated nanocrystals possess excellent photostability, large Stokes/anti-Stokes shifts, sharp-band emissions, and long luminescence lifetimes [84]. Benefiting from these advantages, RE-activated nanocrystals are promising candidates for applications in lighting and displays [85], optical fibers and amplifiers [86], bioimaging [87], photoactivation reactions [88], and photovoltaic devices [89].

With differently arranged energy levels, various rare earth activators are endowed with distinctive transition pathways. Ce³⁺, widely studied and used for phosphor activators, is featured by its broadband 4f → 5d transition [90]. For other RE³⁺ activators, intra-configurational 4f → 4f transitions are the main intrinsic mechanism for their luminescence processes. Three different pathways for intra-configurational transition may occur in rare earth ions, namely, down-shifting, quantum-cutting (also known as downconversion), and upconversion (as shown in Fig. 5.15). In a down-shifting process, a high-energy excitation photon results in a low-energy one emitted, with the quantum efficiency below 100 %. Tb³⁺, Eu³⁺, Sm³⁺ and Dy³⁺ are typical activators for the down-shifting studies [91]. Due to their sufficiently high efficiencies and long luminescence lifetimes, down-shifting-based nanocrystals have been employed as candidates for lighting and displays and time-resolved biodetection applications [92, 93]. In a quantum-cutting process, one high-energy excitation photon splits into two (or more) low-energy ones by cascading down the ladder-like energy levels, with the quantum efficiency above 100 %. Nanocrystals with Pr³⁺/Gd³⁺–RE³⁺ (RE = Eu, Tb, Er) pairs and Yb³⁺–RE³⁺ (RE = Tb, Tm, Pr) pairs dominate the quantum-cutting studies [94]. Considering the high quantum efficiency and specific visible or near-infrared (NIR) emissions, quantum-cutting-based nanocrystals are promising in lighting, displays, and particularly, improvement for solar cells [95, 96]. The photon upconversion process refers to a non-linear optical process where two (or more) low-energy excitation photons generate a high-energy one assisted by real intermediate excited states, with the quantum efficiency below 100 % [97]. In this section, we will mainly discuss the photon upconversion processes and extended optical applications of rare earth activated anisotropic nanocrystals.

Fig. 5.15

Simplified illustration of intra-configurational transition pathways for rare earth ions. a Down-shifting. b Quantum-cutting (or downconversion). c Upconversion. The upward, downward and dashed arrows represent excitation, emission and relaxation processes, respectively

5.3.1.2 Photon Upconversion in Rare Earth Anisotropic Nanostructures

Due to the high conversion efficiency, energy transfer upconversion (ETU) emissions have been widely studied. Two different types of luminescent centers, namely, a sensitizer and an activator, are involved in the ETU process [98–100]. Yb³⁺, with only one excited state (²F_5/2) and a large absorption cross-section at 980 nm, is an outstanding sensitizer for the ETU process. Moreover, the energy of ²F_5/2 state matches well with that of most rare earth ions, which also facilitates Yb³⁺ as a commendable sensitizer. With ladder-like arranged energy levels, long-lived intermediate excited states and excellent resonance with the energy gap from ²F_7/2 to ²F_5/2 of Yb³⁺, Er³⁺, Tm³⁺, and Ho³⁺ (especially Er³⁺ and Tm³⁺) ions are ideal pairing activators for the ETU process. In addition, the host matrices also have significant influences on the upconversion behaviors of doped luminescent centers. Generally, rare earth based fluorides and complex fluoride salts are considered as optimal host matrices due to their low phonon energy and high transparency for NIR photons [101–103].

In Yb³⁺–Er³⁺ activated nanocrystals, three-photon 415 nm emission (²H_9/2 → ⁴I_15/2), two-photon 525, 545 and 655 nm emissions (²H_11/2 → ⁴I_15/2, ⁴S_3/2 → ⁴I_15/2, ⁴F_9/2 → ⁴I_15/2) can be observed after the successive energy transfer from Yb³⁺ to Er³⁺ (Fig. 5.16a) [104]. In Yb³⁺–Ho³⁺ activated nanocrystals, three-photon 485 nm emission (⁵F₃ → ⁵I₈), two-photon 545 nm emission (⁵F₄, ⁵S₂ → ⁵I₈) and 650 nm emission (⁵F₅ → ⁵I₈) are the typical upconversion spectral fingerprints (Fig. 5.16b) [105]. The densely arranged energy levels of Er³⁺ and Ho³⁺ determine that multiple photon upconversion processes are scarcely observed in Yb³⁺–Er³⁺ and Yb³⁺–Ho³⁺ activated nanocrystals. However, multiple photon upconversion emissions, coming from three-, four-, and five-photon transitions, can be frequently detected in Yb³⁺–Tm³⁺ activated nanocrystals. This should be attributed to the discretely arranged energy levels of Tm³⁺, which, to a large extent, reduce the non-radiative relaxation processes. Therefore, two-photon 800 and 695 nm emissions (³H₄ → ³H₆, ³F₃ → ³H₆), three-photon 645 and 475 nm emissions (¹G₄ → ³F₄, ¹G₄ → ³H₆), four-photon 450 and 365 nm emissions (¹D₂ → ³F₄, ¹D₂ → ³H₆), and even five-photon 345 and 290 nm emissions (¹I₆ → ³F₄, ¹I₆ → ³H₆) can be simultaneously released from Yb³⁺–Tm³⁺ pairs (Fig. 5.16c) [106].

Fig. 5.16

Proposed upconversion energy transfer pathways in a Yb³⁺–Er³⁺, b Yb³⁺–Ho³⁺ and c Yb³⁺–Tm³⁺ pairs. Excitation and emission processes are represented by upward and downward solid arrows, respectively. The dotted and dashed arrows denote energy transfer and non-radiative relaxation processes, respectively

When rare earth activated nanocrystals adopt anisotropic morphologies, their upconversion properties differ greatly from those with isotropic ones, especially the polarized upconversion emissions. Qiu et al. for the first time investigated the polarized upconversion emissions from Yb³⁺–Tm³⁺ activated hexagonal phased NaYF₄ single nanorod [107]. Upon excitation with a 980 nm linearly polarized laser, sharp energy level splitting of the ¹D₂ and ¹G₄ states of Tm³⁺ was observed at room temperature, generating singlet-to-triplet emissions at 735 nm (¹D₂ → ³F₃) and 768 nm (¹G₄ → ³H₅), which previously were assumed to be partially forbidden (Fig. 5.17a). In the meanwhile, the multiple discrete emission intensity periodic variation with polarized direction was also observed (Fig. 5.17b). Moreover, the aspect ratio of nanorod was studied for the effect on the polarization property (Fig. 5.17c). Hexagonal-phased NaYF₄:Yb,Tm single nanorod and single nanodisk were found to possess quite similar fitting contours in each transition of Tm³⁺, suggesting a negligible effect of aspect ratio on the polarization anisotropy. However, when Gd³⁺ ions were doped into the nanorods, the polarization anisotropy became distinctly different. Two different kinds of intensity variation states were observed for each transition. The authors reasoned that the different local symmetry of Tm³⁺ in NaYF₄:Gd,Yb,Tm nanorods should be responsible. Based on the comprehensive studies, the intrinsic transition properties and crystal local symmetry should dominate the polarization anisotropy.

Fig. 5.17

a Scanning electron microscopy image of NaYF₄:Yb,Er nanorod and its upconversion emission spectrum upon excitation with a 980 nm linearly polarized laser. b Polarization angle (φ_em) dependent upconversion emission spectra. c Polar plots of the upconversion peak intensity as a function of the emission polarization angle, which is corresponding to the transitions of Tm³⁺ in NaYF₄:Yb,Er nanorod, nanodisk, and NaYF₄:Gd,Yb,Er nanorod. (Reprinted with permission from [107]. Copyright 2013 American Chemical Society)

Apart from the polarization anisotropy, the pleochroism of upconversion emissions is also embodied in anisotropic nanocrystals. Yan et al. prepared a series of hexagonal phased NaYF₄:Yb,Er nanocrystals with different sizes ranging from ~20 to ~300 nm. As the size increased, the morphology of the nanocrystals changed from spherical to disk-like. And the green-to-red emission ratio of Er³⁺ was found to increase with the nanocrystals from isotropic to anisotropic (Fig. 5.18) [104, 108]. Furthermore, the authors investigated the upconversion behavior of one-dimensional NaMgF₃:Yb,Er nanorods. Different from the upconversion profile of hexagonal-phased NaYF₄:Yb,Er nanodisks, for which the green emission dominated the visible regime, the green emission from NaMgF₃:Yb,Er nanorods was significantly weaker than the red emission [109]. Recently, they developed a novel structure of NaYF₄:Yb,Er nanocages, deriving from the removal of NaCl in NaCl/NaYF₄:Yb,Er core/cage nanocomposites. The upconversion behavior of NaYF₄:Yb,Er nanocages and nanospheres were examined. It was found that the red-to-green emission ratio of Er³⁺ decreased to 0.44 in nanocages from 1.89 in nanospheres [110]. Zhu et al. compared the upconversion profiles of orthorhombic KSc₂F₇:Yb,Er nanorods with hexagonal NaYF₄:Yb,Er nanorods [111]. They observed greatly enhanced red and violet emissions of Er³⁺ in KSc₂F₇:Yb,Er nanorods. Liu et al. prepared 1D orthorhombic KYb₂F₇:Yb,Er nanorods, where Yb³⁺ ions were localized as tetrad clusters. The spectral results exhibited that multiphoton violet emission of Er³⁺ was significantly enhanced compared with that in hexagonal NaYF₄ nanodisks [112]. Moreover, the violet emission of Er³⁺ was assigned as a four-photon emission.

Fig. 5.18

a Upconversion emission spectra of different sized hexagonal NaYF₄:Yb,Er nanocrystals in cyclohexane. b Diagram of green to red emission ratio versus the size of NaYF₄:Yb,Er nanocrystals, with insets representing the morphology transition from isotropic to anisotropic. (Reprinted with permission from [104]. Copyright 2007 American Chemical Society)

When multiple activators are embedded in anisotropic nanocrystals, the upconversion emission s become more intriguing. Li et al. studied the upconversion behavior of hexagonal phased NaYF₄:Yb,Ho nanorods after Ce³⁺ doping [113]. They observed that the red-to-green emission ratio of Ho³⁺ increased after the introduction of Ce³⁺, with the emission output from pure green to greenish-yellow. Moreover, unusual ⁵G₅ → ⁵I₇ and ⁵F₂, ³K₈ → ⁵I₇ transitions of Ho³⁺ and 5d → 4f transitions of Ce³⁺ can be simultaneously observed. Qin et al. modulated the red-to-green emission ratio of Er³⁺ in hexagonal-phased NaYF₄ one-dimensional microtubes by precisely defining the doping ratio of Tm³⁺ dopant [114]. An increased red-to-green emission ratio of Er³⁺ was detected with elevated content of Tm³⁺ from 0 to 2 mol%, with the emission output altering from green to red. Multiphoton cross-relaxations within Tm³⁺–Er³⁺ pairs were proposed to account for the spectral variation. Lee et al. fabricated a series of quasi-two-dimensional anisotropic core/shell nanostructures, where the shell layers preferentially grew on the side faces (100) of the hexagonal phased NaYF₄ nanocrystals. Er³⁺ and Tm³⁺ ions were separately embedded in the core and shell regions, respectively, giving out multicolor upconversion emissions upon NIR excitation [115]. Recently, Liu et al. employed an end-on growth of upconversion nanocrystals compromising Tm³⁺ and Er³⁺ ions onto the hexagonal one-dimensional NaYF₄ nanorods. The emission output was further modulated by tuning the molar ratio of sensitizer and activator. A series of dual-color upconversion nanorods displaying different combinations of the three primary colors were obtained (Fig. 5.19). With the dual-color emitting nanorods, they achieved some prototype models for barcoding for anti-counterfeiting application [116].

Fig. 5.19

a Scanning electron microscopy image of NaYF₄ based nanorod. b–f Optical micrographs showing five sets of dual-color-banded upconversion nanorods. Note that the appearance of a tinge color at the tip junction is due to the chromatic aberration and limited resolution of the microscope. (Reprinted with permission from [116]. Copyright 2014 American Chemical Society)

Recent progresses show that upconversion emissions from anisotropic nanocrystals can be tailored by the other co-assembled nanomaterials. Yan et al. investigated the upconversion emissions from a two-dimensional dense film of NaYF₄:Yb,Er nanocrystals by co-assembling with Ag nanowires [117]. They noticed 2.3 fold and 3.7 fold enhancement for the green and red emissions, respectively. And the large scattering efficiency of Ag nanowires was proposed to account for the enhancement. Kagan et al. studied the upconversion emission in an assembly comprising single NaYF₄:Yb,Er nanodisk and single Au nanorod (as shown in Fig. 5.20) [118]. The longitudinal surface plasmon resonance of the Au rod was tuned to match the 977 nm excitation wavelength of the NaYF₄:Yb,Er nanodisk by tailoring the dimension of the rod. And the spatial distance between the two composites was within the range of the intense near-fields surrounding the Au rod tips. As a result, two and threefold enhancements were achieved for the green and red emissions of Er³⁺, respectively. Moreover, the enhancements exhibited a strong dependence on the polarization of the excitation light relative to the Au rod axis. The longitudinal polarization exhibited more spectral changes than the transverse one. Zink et al. encapsulated single NaYF₄:Yb,Er nanorod and single smaller superparamagnetic Fe₃O₄ nanocrystal into one mesoporous silica nanoparticle, and further exposed the nanocomposites to oscillating magnetic field [119]. The upconversion behavior of NaYF₄:Yb,Er nanorod was modulated by heat magnetically induced by Fe₃O₄ inside the nanocomposites. Spectral results showed that the emission ratio of 525 to 545 nm of Er³⁺ enhanced with elevated temperature, which resulted from prolonged exposure time to the magnetic field.

Fig. 5.20

Simulated field intensity enhancement maps for a transverse and b longitudinal excitation polarization. c Representative upconversion emission spectra from the center of the NaYF₄:Yb,Er nanodisk under longitudinal (blue) and transverse (red) incident polarization for the Au–NaYF₄:Yb,Er assembly. (Reprinted with permission from [118]. Copyright 2014 American Chemical Society)

The excitation pathways can be modulated in anisotropic nanocrystals as well. Yan et al. extended the NIR excitation band of Yb³⁺–RE³⁺ (RE = Er, Tm) pairs by introducing Nd³⁺ into a separated layer with an array of Yb³⁺ ions as the energy bridgers (Fig. 5.21) [120]. As a result, almost similarly efficient upconversion emissions can be obtained upon 980 nm (for Yb³⁺) and 808 nm (for Nd³⁺) excitations. Moreover, the potential overheating issue induced by longtime 980 nm irradiation was largely minimized by shifting the excitation wavelength to 808 nm. Similarly, Wang et al. also realized energy transfer management in multishell structured anisotropic nanorods upon 808 nm excitation [121]. Besides spatial separation of Nd³⁺ and activators, Han [122] and Liu [123] also achieved cascade sensitization of upconversion emissions by co-doping Nd³⁺ and Er³⁺/Tm³⁺ in the same layer.

Fig. 5.21

a Proposed energy transfer pathways for Nd³⁺ activated upconversion emissions. b Typical upconversion emission spectra of Yb³⁺–Er³⁺ activated nanocrystals upon 980 and 808 nm excitations. (Reprinted with permission from [120]. Copyright 2013 American Chemical Society)

5.3.1.3 Optical Applications

The past few years witnessed the rapid development of rare earth upconversion nanocrystal s in various aspects, including bioimaging studies [124], theranostics [125], sensing and detection [126], photovoltaic devices [127], and photoactivation reactions [128]. In this part, we will present a brief summary of rare earth upconversion nanocrystals for applications in biology and photoresponsive materials.

Due to the specific NIR excitation, resolvable UV to NIR emission and deep tissue penetration, rare earth upconversion nanocrystals have been widely used as biomarkers for biological studies. Various models, such as cells [129], C. elegans [130], rats [131] and rabbits [132], have been employed to assess the bioimaging performance of rare earth upconversion nanocrystals. Yan et al. examined the bioimaging capability and assessed the toxicity of NaYF₄:Yb,Tm nanocrystals in vitro and in vivo [133]. After incubation with the upconversion nanocrystals, bright upconversion emission signals can be collected from HeLa cells and C. elegans (Fig. 5.22). Moreover, no obvious toxicity effect was found for both cells and C. elegans. Li et al. synthesized sub-10 nm hexagonal NaLuF₄:Gd,Yb,Er nanocrystals and applied them for sensitive in vivo imaging. Subcutaneous imaging of 50 KB cells and intramouse imaging of 1000 KB cells after intravenous injection were obtained [134].

Fig. 5.22

In vitro and in vivo bioimaging of NaYF₄:Yb,Tm nanocrystals. a, b False-colored upconversion emission signals in visible (400–700 nm) and NIR (720–800 nm). c Merged images of HeLa cells and upconversion emissions. d, e False-colored NIR upconversion emission signal and the merged image of C. elegans. (Reprinted with permission from [133]. Copyright 2011 Elsevier B.V.)

Besides single-modal optical imaging, rare earth upconversion nanocrystals have also been assembled with magnetic nanocrystals, noble metal nanocrystals and photosensitive molecules to support multi-modal imaging, theranostic and biodetection applications. Liu et al. deposited Au and Fe₃O₄ nanocrystals onto hexagonal NaYF₄:Yb,Er nanodisks to conduct photothermal therapy , magnetic resonance imaging and upconversion imaging studies [135]. Yan et al. fabricated multifunctional NaGdF₄:Yb,Er@CaF₂@SiO₂ nanocomposites, which were further grafted with photosensitizers to generate singlet oxygen, for upconversion imaging, magnetic resonance imaging and photodynamic therapy [136]. Li et al. decorated NIR cyanine dye hCy7 molecules on hexagonal NaYF₄:Yb,Er,Tm nanocrystals. When MeHg⁺ ions were added to the nanocomposites, the red emission of Er³⁺ became significantly enhanced while the NIR emission of Tm³⁺ sharply decreased. They realized both in vitro and in vivo detection of MeHg⁺ ions with the detection limit of 0.18 ppb [137].

Upconversion emissions from rare earth activated nanocrystals are also demonstrated to trigger photoactivation reactions. Diarylethene derivatives, o-nitrobenzyl containing molecules, and azobenzenes are typical photoresponsive molecules. Branda et al. used alternative irradiation of upconversion emissions from NaYF₄:Yb,Tm and NaYF₄:Yb,Er nanocrystals to remotely control the photoswitching process of diarylethene derivatives [138]. By precisely defining the composition of upconversion nanocrystals and tuning the NIR excitation power, reversible photoswitching process of diarylethene derivatives was achieved [139]. Xing et al. reported a novel type of nanocomposites for controllable release of D-luciferin molecules, which were decorated on the surface of NaYF₄:Yb,Tm nanocrystals via the o-nitrobenzyl group. Upon NIR excitation, UV emissions from Tm³⁺ can trigger the disassociation of D-luciferin molecules [140]. Recently, Yan and Li et al. fabricated a NIR-light-responsive self-organized helical superstructure by doping chiral azobenzenes and two-dimensional (2D) Yb³⁺–Tm³⁺ activated nanodisks into a liquid crystal host. Upon high power excitation, UV emissions (Fig. 5.23a) from Tm³⁺ can direct the trans to cis isomerization of azobenzenes, resulting a red-shifted reflection wavelength of the liquid crystal (Fig. 5.23b). At low power excitation, the blue emissions (Fig. 5.23c) from Tm³⁺ are observed to trigger the reversible process of the blue-shifted absorption of liquid crystal (Fig. 5.23d) [141].

Fig. 5.23

Upconversion emission spectra from Yb³⁺–Tm³⁺ activated nanodisks and reflection spectra of liquid crystal upon high power excitation (a, b) and low power excitation (c, d). (Reprinted with permission from [141]. Copyright 2014 American Chemical Society)

5.3.2 Magnetic Properties and MRI Applications of Anisotropic Rare Earth Nanostructures

The rare earth elements are well known for their superior magnetic properties. Their trivalent ions are characterized by fⁿ configurations, which renders them large paramagnetic susceptibility (expect Sc³⁺, Y³⁺, La³⁺ and Lu³⁺, due to their lack of 4f unpaired electrons) [142]. Different from the 3d electrons of some transition metal ions (Fe³⁺, Mn²⁺ and Co²⁺, etc.), the 4f electrons of rare earth ions are less influenced by the microenvironment, as they are shielded by outer-shell (5s²5p⁶) electrons. Besides, some of their compounds possess distinctive magnetic properties, such as large saturation magnetization, large magnetocrystalline anisotropy constant, large magnetostriction constant and remarkable magneto-optic effect [143]. As a result, rare earth materials have been widely used in traditional magnet technology over the past decades, for example, SmCo₅ and Nd₂Fe₁₄B are well-known permanent magnets for their large static magnetic field; yttrium barium copper oxide are famous for displaying high-temperature superconductivity. With the development of nanotechnology, rare earth magnetic materials also have proved their great potential for applications in nanoscale, particularly in the field of magnetic resonance imaging (MRI) (Table 5.1).

Table 5.1 Ground state, g values, calculated and experimental room temperature χT value for Ln(III) ions

As a very powerful noninvasive detection protocol, MRI has superb advantages, such as fast scan time, deep penetration, high spatial resolution, and no radiochemical damage [144, 145]. Nevertheless, the low imaging sensitivity hampered its further application. MRI contrast agents (CAs), which are able to alter the relaxation time of surrounding protons, have been utilized to improve the sensitivity both in clinical diagnosis and biological research [146]. Longitudinal relaxation enhancement is mainly related to the inner-sphere regime that chemically exchanges directly with the paramagnetic centers, while transverse relaxation enhancement is mainly related to proton’s effective diffusion and interaction with magnetic dipolar moment in outer-sphere regime (Fig. 5.24) [147]. The ability of CAs to shorten longitudinal relaxation time (T ₁) and transverse relaxation time (T ₂) is evaluated by longitudinal relaxivity (r ₁) and transverse relaxivity (r ₂), respectively. Generally, CAs with low r ₂ to r ₁ ratio (r ₂/r ₁) are used for T ₁-weighted images, while CAs with high r ₂/r ₁ are used for T ₂-weighted images. For clinic, Gd-based chelates (especially Gd-DTPA) and superparamagnetic iron oxide nanoparticles (SPIO) are most widely used T ₁ and T ₂ CAs, respectively. To lower the risks of free Gd³⁺, [148] most clinical lanthanide CAs are used in form of chelates, so as to stabilize Gd³⁺ by coordination [149, 150]. However, the leakage of Gd³⁺ is inevitable due to coordination equilibrium. Moreover, the Gd³⁺-based chelates usually have small molecular weight, resulting in rapid clearance from human body. Aiming to solve the present problems of chelate CAs, the concept of nanoparticulate CAs has been raised [151]. The advantages of nanoparticle-based CAs include: (1) rigid inorganic structure could minimize the leakage of Gd³⁺; (2) high payload of Gd³⁺ per nanoparticle could ensure large local contrast enhancement; (3) larger size than chelates could prolong not only the circulation time in vivo, but also the rotational correlation time (τ _R); (4) multi-functional bio-applications can be realized through doping and further surface engineering.

Fig. 5.24

Schematic illustrations of proton phenomena in magnetic systems relating to T ₁ and T ₂ relaxations. a Paramagnetic metal complex system and the selected key parameters to T ₁ relaxation of protons: molecular tumbling time (τ _R), proton residence lifetime (τ _M), and the coordinating water molecular number (q), while the magnetic gradient field around paramagnetic center is neglected. b Phenomena of proton interaction with a spherical magnetic nanoparticle system: water molecular diffusion and chemical exchange with surface magnetic metals, related to their T ₂ and T ₁ contrast enhancements, respectively. (Reprinted with permission from [147]. Copyright 2014 American Chemical Society)

5.3.2.1 Gd-Based Anisotropic Nanostructures as MRI CAs

With seven unpaired electrons and isotropic orbitals, Gd³⁺ has both large magnetic moment and long electric relaxation time (due to the negligible spin–orbit interaction), and is widely accepted as the best choice for longitudinal proton relaxation enhancement [150]. Nano-sized gadolinium oxide nanoparticles were first evaluated for their physicochemical and NMR properties by McDonald et al. in 2003 [152], which opened the research field of rare earth-based nanoparticulate MRI CAs. Up to now, the studies of Gd-based T ₁ CAs have been extended from gadolinium oxides (Gd₂O₃) to gadolinium hydroxide (Gd(OH)₃), gadolinium fluoride (GdF₃), sodium gadolinium fluoride (NaGdF₄) and gadolinium oxysalts (such as GdPO₄), etc. Besides, doping Gd³⁺ into other inorganic matrices has become an effective strategy to combine MRI with other imaging modalities [153].

Relaxivity is one of the most important indexes for CAs, which could reflect the contrast enhancement efficacy. High relaxivity is always pursued, since it will bring benefits including improved image quality, reduced injection dose and lower risks. However, the r ₁ value of Gd-DTPA is quite low (about 4 s⁻¹ mM⁻¹ at 1.5 T) [154]. Although Gd-based nanoparticles were supposed to have higher relaxivities than Gd-DTPA in the beginning, the r ₁ values of those nanoparticles in early stage were still not sufficiently high. This probably could be explained by the immaturity of synthetic methods, since those nanoparticles usually had irregular shapes and wide size distribution. In recent years, with the development of synthetic method, uniform and monodispersed nanoparticles could be readily obtained, and size has been found to significantly affect the relaxivity. Lee et al. synthesized ultrasmall paramagnetic Gd₂O₃ nanoparticles with the size around 1 nm and a large r ₁ was estimated (9.9 s⁻¹ mM⁻¹) [155]. They discussed the size dependence of r ₁ in two terms, surface to volume ratio (S/V) and cooperative induction effect. On one hand, it is indicated that only surface Gd³⁺ significantly contribute to the longitudinal relaxation. Therefore, decreasing the size will lead to larger S/V and more surface Gd³⁺, and positively influence the r ₁. On the other hand, several surface Gd³⁺ could possibly accelerate the longitudinal relaxation of water protons by cooperative induction. When the size is too small to produce this effect, the r ₁ value would decrease. The optimal range of particle size that they suggested is 1–2.5 nm (Fig. 5.25). van Veggel et al. reported on the size-tunable synthesis of NaGdF₄ nanoparticles below 10 nm [156]. Through the analysis of the relationship between particle size and r ₁, they revealed that surface Gd³⁺ are the major contributors to the relaxivity enhancement, and surface Gd³⁺ on a larger nanoparticle affect the relaxivity more strongly than those on a smaller nanoparticle.

Fig. 5.25

Schematic diagram showing that a four surface Gd(III) ions as an example cooperatively induce the longitudinal relaxation of the water proton, whereas b such an effect does not exist in individual Gd(III)−chelates. The interacting and non-interacting Gd(III) ions with a water proton are denoted as dotted and solid arrows for their spins, respectively. The ligands are drawn arbitrarily. (Reprinted with permission from [155]. Copyright 2009 American Chemical Society)

Besides S/V, τ _R is another important parameter affecting relaxivity, which is mainly determined by hydrated radii of nanoparticles [157]. For 0D nanomaterials, the τ _R value shortens with decreasing size, counteracting the effect of S/V towards relaxivity. Whereas for rare earth based nanomaterials with anisotropic morphology, the effect of S/V and τ _R could be consistent, as a result, high relaxivity could be realized through rational design of anisotropic nanoparticles.

For 1D nanostructures, most of the Gd-based nanomaterials have rod-like morphology. Suzuki et al. synthesized dextran coated rod-like GdPO₄ nanoparticles, with 20–30 nm in the major axis and 6–15 nm in the minor axis [157]. The nanoparticles showed high r ₁ and r ₂ values of 13.9 and 15.0 s⁻¹ mM⁻¹, respectively, and 1.1 as the r ₂/r ₁ value at 0.47 T. The tumors in rabbit could be effectively visualized with only 1/10 of the applied dose compared to the clinically used Gd-DTPA [158].

Rod-like Gd(OH)₃ nanoparticles with the length of 150 nm were synthesized by Gillis et al., showing r ₁ and r ₂ of 4.03 and 8.0 s⁻¹ mM⁻¹, respectively, at 1.5 T [159]. As the magnetic field increased to 7 T, the r ₁ slightly decreased and r ₂ increased linearly with the field. Zhao et al. synthesized GdPO₄ nanorods (up to ~100 nm in length and ~10 nm in diameter) in the presence of PVP molecules [160]. The r ₁ of GdPO₄ was 2.08 s⁻¹ mM⁻¹ at 4.7 T. Size-dependent MRI relaxivity of Eu-doped GdPO₄ rod-like nanoparticles was investigated by Talham et al. [161]. They found although ionic molar relaxivity decreases for larger particles, the relaxivity per nanoparticle can be significantly greater. By doping light-emitting lanthanide ions into GdPO₄ host matrix, bimodal nanoprobe (luminescent and magnetic) could be obtained.

Tan et al. demonstrated a simple synthetic strategy for the fabrication of Tb and Yb/Er doped Gd₂O₃ nanorods [162]. As refluxing progressed, monodispersed quasi-spherical Gd₂O₃ nanocrystals were produced at first, and then fused into nanorods (18.8 ± 5.7 nm in length). The Yb/Er co-doped Gd₂O₃ nanorods exhibit good T ₁-weighted MRI contrast enhancing and upconversion fluorescence properties. They also prepared NaDyF₄:Yb³⁺/NaGdF₄:Yb³⁺,Er³⁺ nanorods, with the average diameter and length of 21 and 45 nm, respectively [163]. The r ₁ (0.321 s⁻¹ mM⁻¹) is quite small, and the presence of Dy³⁺ is inferred to affect the T ₁ induced by Gd³⁺. However the r ₂ (437.97 s⁻¹ mM⁻¹) is much higher than other reported Dy-based materials, probably due to the additional synergistic contribution of T ₂ shortening by the Gd³⁺ sitting adjacent to Dy³⁺ in the nanorods.

Jin et al. reported a facile solvothermal approach to synthesize monodispersed ultrathin GdF₃ nanowires (lengths of 20–40 nm and diameters of ca. 2 nm) [164]. The nanowires were formed through assembly of GdF₃ nanocrystals. The surface capping ligands anchored to some specific GdF₃ surfaces selectively, and thus facilitated the oriented attachment along the no (or less) capped crystal planes. After modificated with the surfactant Pluronic F127, the GdF₃ nanowires showed high r ₁ of 15.0 s⁻¹ mM⁻¹ at 7 T. The large length in 1D direction and low symmetry of nonspherical structure prolonged their τ _R compared to 0D nanoparticles, which was believed to be responsible for the high r ₁.

2D nanostructures have also been well studied. Murray et al. reported shape-controlled synthesis of Gd₂O₃ tripodal and triangular nanoplates [165]. With an increase of magnetic field strength from 9.4 to 14.1 T, the r ₁ of PEI-capped Gd₂O₃ tripodal nanoplates decreased from 1.41 to 0.89 s⁻¹ mM⁻¹, and r ₂ increased from 140 to 193 s⁻¹ mM⁻¹. Reduced spectral density at higher magnetic field strengths results in less efficient dipole-dipole relaxation and a corresponding decrease in r ₁. The magnetic moment of paramagnetic nanoparticles is proportional to local magnetic field strength, resulting in more effective dephasing of proton and thus increased r ₂ at higher fields.

Colvin et al. synthesized Gd₂O₃ nanoplates via a thermal decomposition method [166]. The nanoplates have a core diameter varying from 2 to 22 nm and a thickness of 1–2 nm. When coated with an octylamine modified poly(acrylic acid) polymer layer, the r ₁ of 2 nm Gd₂O₃ nanoplates reached 47.2 s⁻¹ mM⁻¹ at 1.41 T, about 10 times higher than that of Gd-DTPA.

Li et al. reported the synthesis of lanthanide ion doped GdVO₄ tetragonal nanosheets (with a thickness of ~5 nm and a width of ~150 nm) using a facile solvothermal reaction [167, 168]. After the ligand exchange reaction with poly(acrylic acid), the hydrophilic nanosheets showed excellent paramagnetic properties (r ₁ reached 37.8 s⁻¹ mM⁻¹ at 0.5 T). This high r ₁ was attributed to the special 2D nanostructure, since most of the Gd³⁺ were exposed on the surface of nanosheets and water-soluble polymer coating may offer a hydrophilic interface between the surface of the nanosheets and water molecules.

Alexandrou et al. synthesized Eu-doped GdVO₄ nanoparticles which combine magnetic resonance contrast enhancing properties with luminescence properties and hydrogen peroxide sensing features [169]. The 2D projection of the nanoparticles is an ellipse with the two major axes having lengths of 13.1 ± 1.1 and 26.6 ± 4.8 nm, respectively. The paramagnetic nanoparticles showed high r ₁ (8.18 s⁻¹ mM⁻¹) and r ₂ (9.38 s⁻¹ mM⁻¹).

5.3.2.2 Ln³⁺(Other Than Gd³⁺)-Based Anisotropic Nanostructures as MRI CAs

Currently, the development of MRI is towards high field to further increase the signal-to-noise ratio, and to further improve the temporal and spatial resolution [170]. Unfortunately, the efficiencies of current clinically used CAs are reduced when field strength increased to higher than 3 T [145, 171, 172]. Therefore, CAs with high performance under high magnetic field are in urgent need. Paramagnetic trivalent RE³⁺ ions other than Gd³⁺ are less efficient T ₁ relaxation agents, mainly due to their short electronic relaxation time. However, those paramagnetic RE³⁺ with large magnetic moments (such as Tb³⁺, Dy³⁺, Ho³⁺ and Er³⁺) can induce field inhomogeneities, resulting in transverse relaxation enhancement of protons [173]. The r ₂ of lanthanide nanoparticles appeared to be proportional to the magnetic field strength and to the square of the effective moment of the lanthanide [174]. Thus, lanthanide-based nanomaterials hold great potential as T ₂ CAs for high field MRI [171, 175].

Same with T ₁ CAs, relaxivity properties of T ₂ CAs are also believed to be influenced significantly by size and shape. van Veggel et al. synthesized NaDyF₄ nanoparticles with different sizes, and studied size dependence of r ₂ [176]. They found larger nanoparticles show higher r ₂ than smaller ones, and attributed this to spin-canting effects. With the decrease in the nanoparticle size, the S/V ratio increases and curvature of nanoparticles becomes more pronounced. Spins located near the surface tend to be slightly tilted, resulting in low magnetization value for small nanoparticles. Zeng et al. reported the synthesis of NaErF₄ with different shapes and discussed the shape induced magnetic properties [177]. They speculated that the different shape anisotropy leads to the difference of magnetization, and consequently the rod-like morphology had highest magnetization. After doping with Yb³⁺, multi-functional NaErF₄ nanorods were obtained, combining upconversion emission, X-ray attenuation and negative relaxation enhancement properties [178].

Shen et al. reported Co²⁺-doped NaYF₄:Yb³⁺,Tm³⁺ nanorods (mean width of 19.8 nm and mean length of 33.2 nm) as dual-modality imaging nanoprobes [179]. The r ₂ of nanorods was up to 1544 s⁻¹ mM⁻¹, due to the doping of Co²⁺. Except the aforementioned sodium lanthanide fluoride, hydroxides with rod-like morphology have also been reported. Lee et al. reported water-soluble and biocompatible D-glucuronic acid coated Dy(OH)₃ nanorods (20 × 300 nm) and Eu(OH)₃ nanorods (9.0 × 118.3 nm) [180, 181]. The r ₁ and r ₂ of Dy(OH)₃ was 0.97 and 181.57 s⁻¹ mM⁻¹ at 1.5 T, respectively, while both of r ₁ and r ₂ were very small for Eu(OH)₃, probably due to the insufficient surface coating. To compensate for the deficiencies of individual imaging modalities, NaDyF₄ nanorods coated with α-cyclodextrin and modified with Gd-DTPA were prepared by Liao et al. [182] Dy³⁺ in the host matrix induces a high X-ray attenuation ability for CT and negative enhancement for T ₂-weighted MRI (r ₂ = 7.68 s⁻¹ mM⁻¹), while Gd-DTPA induces positive enhancement for T ₁-weighted MRI (r ₁ = 4.65 s⁻¹ mM⁻¹).

Reports of 2D lanthanide-based nanomaterials as MRI CAs are quite rare, probably due to the relatively large S/V ratio, which is unfavorable for T ₂ CAs as we discussed above. Lanthanide (Gd, Dy and Yb) oxides nanodisks were produced with high-temperature colloidal syntheses by Helms et al. [183]. These nanodisks were 10–14 nm in diameter with a thickness of a single unit cell. For Dy₂O₃ nanodisks, their r ₂ values showed ca. 1.7-fold enhancement over Gd-DTPA when passivated with poly(acrylic acid)-methoxy-terminated polyethylene oxides alongside Gd-DTPA at 1.41 T.

5.3.3 Catalytic Application of Rare Earth Anisotropic Nanostructures

The catalytic applications of rare earth nanomaterials draw large attention due to the unique electronic structures as rare earth elements possess 4f orbitals without full electron occupancy and lanthanide contraction. In particular, La and Ce are extensively used as catalyst or catalytic components. Among the compounds of La and Ce, ceria is the most representative catalyst and has been widely used in a number of catalytic reactions. Ceria can uptake and release oxygen owing to the transformation between Ce³⁺ and Ce⁴⁺ [184]. When the size decreases to nanometer scale, ceria exhibits more excellent performances due to high ratio of atoms exposed and thus high specific area. A large number of researches related to the controllable synthesis and applications of ceria nanomaterials have been conducted in the past decade. Therefore, here in this section, we will mainly focus on nanoceria to demonstrate the properties and applications of rare earth nanomaterials.

5.3.3.1 Basic Properties of Ceria Nanocrystals

As mentioned before, CeO₂ crystals are in fluorite structure, consisting of a face-centered cubic (fcc) unit cell of cations with anions occupying the octahedral interstitial sites [185]. In this structure, each cerium cation is coordinated by eight nearest-neighboring oxygen anions, and each oxygen anion is coordinated by four nearest-neighboring cerium cations. Of note, there exist both +3 and +4 oxidation states for cerium, and the transformation between Ce³⁺ and Ce⁴⁺ is reversible. Ceria can form a non-stoichiometric oxide (CeO_2–x) through releasing or uptaking oxygen at the ambient partial pressure of oxygen, while maintaining the crystal structure [186]. However, owing to the transformation between Ce³⁺ and Ce⁴⁺, the discrepancy in the number of atomic neighbors and interatomic distances can cause surface defects in ceria, which is often responsible for the catalytic activity of ceria [17]. The main defect in CeO₂ is the formation of oxygen vacancies. The fluorite structure allows oxygen atoms for rapid diffusion as a function of the number of oxygen vacancies. This character endows ceria nanomaterials the capacity to store and release oxygen. In order to maintain the charge balance, the oxygen vacancies form and disappear during the transformation between Ce³⁺ and Ce⁴⁺ [187]. Oxygen vacancies can be produced via the following defect reaction:

$$ {\text{Oo}}\text{ + } 2 {\text{Ce}}_{\text{Ce}} \text{ = }1/2{\text{O}}_{ 2} \text{ + }{\text{V}}{\ddot{\text{o}}} + 2 {\text{Ce}}_{\text{Ce}}^{{\prime}} $$

(1)

in which Ce_Ce, Oo, Vö and Ce′_Ce represent Ce⁴⁺, O²⁻, oxygen vacancies and Ce³⁺, respectively. Oxygen vacancies play an important role in the activity of ceria nanomaterials for the various catalytic reactions like CO oxidation [188]. Besides the defect properties, ceria also possesses versatile acid–base properties [189]. The defect and acid–base properties are the bases of catalytic applications of ceria nanomaterials.

There are three low-index lattice planes on the surface of CeO₂ nanocrystals: (100), (110) and (111) [185]. According to density functional theory calculations, their stability follows this sequence: (111) > (110) > (100), while their activity follows the opposite order [190, 191]. And the oxygen vacancies are different on three planes. The energy required to generate oxygen vacancies on (111) plane surface is greater than those on (110) and (100) plane surfaces. Therefore, (110) and (100) planes in ceria are favorable in catalytic reactions owing to more vacancies. Generally, the shape of CeO₂ will determine the exposed lattice planes. For example, ceria nanoparticles consisting of octahedra or truncated octahedral shapes mainly expose the most stable (111) facets, while nanorods can be enclosed by (110) and (100) planes, and nanocubes can expose (100) planes. So there are more oxygen vacancies on the surface of nanorods and nanocubes. In view of the intrinsic activity of different planes, the catalytic performance and applications are considered to be closely linked to morphologies of ceria nanomaterials. Recently, ceria nanomaterials have been applied in a number of catalytic reactions, such as carbon monoxide (CO) oxidation, photocatalysis, organic chemistry and bio-anti-oxidation.

5.3.3.2 Catalysis for CO Oxidation

Ceria nanomaterials have been extensively applied in heterogeneous catalysis based on its ability to release and uptake oxygen [192]. CO oxidation is a simple but most extensively studied reaction in heterogeneous catalytic reaction. The reports on catalytic application of ceria nanomaterials for CO oxidation are abundant. Recently, the morphology-dependent catalytic properties of ceria nanomaterials have been investigated including 0D nanostructures (nanocube [193, 194], nanopolyhedron [194]), 1D nanostructures (nanowire [193, 195], nanorod [193, 196], nanotube [193, 196]) and 2D nanostructures (nanoplate [196]).

The morphology-dependent activities toward CO oxidation of ceria nanomaterials are essential for developing advanced catalysts. The activities of ceria toward CO oxidation are regulated through controlling morphologies by the modulation of specific area, exposed facets and oxygen vacancies. Ceria nanomaterials with high specific area which have more active sites show superior catalytic performance [197]. For example, Li et al. [197] prepared CeO₂ hollow spheres via a layer-by-layer method and observed the activity of the CeO₂ hollow spheres was higher than that of the non-hollow counterpart. Furthermore, CeO₂ nanotube also exhibited better catalytic performance in CO oxidation due to the exposed inner and outer surface [198]. Therefore, multifold ceria nanomaterials with high specific area were investigated so as to design favorable catalyst, such as spindle-like CeO₂ [198], platelet-like CeO₂ [199] and mesoporous CeO₂ [200].

However, specific area is not the sole factor influencing the catalytic activity toward CO oxidation. For example, Zhou et al. [22] reported that single CeO₂ nanorods are more active for CO oxidation than irregular nanoparticles. The higher activity was attributed to more exposed reactive crystal planes (100) and (110). Recently, morphology-controlled synthesis of ceria nanomaterials, such as nanocubes, 1D nanomaterials (wire, rods and tubes) and 2D nanomaterials (plates), provide an opportunity for designing excellent catalytic materials with desirable exposed planes. For example, Yan et al. [24] focused on the shape-dependent oxygen storage capacity of ceria nanocrystals with various shapes (nanopolyhedra, exposing (111) and (100); nanorods, (110) and (100), and nanocubes, (100)). The results showed that nanorods and nanocubes were more active than nanopolyhedra. Therefore, the design of highly active catalyst with well-defined crystal planes has been attractive and challenging in CO oxidation. Wu et al. [201, 202] investigated CO oxidation over ceria nanocrystals with defined surface planes including rods, cubes, and octahedra. They revealed that the reducibility of these ceria nanoparticles is in accordance with their CO oxidation activity, following the sequence: rods > cubes > octahedral. In addition, the intrinsic factor for this order was surface oxygen vacancies on different planes. The redox activity of ceria is proposed to be controlled by the nature of the oxygen vacancies which is critical for ceria catalysis. Wu et al. [201] probed defect sites on CeO₂ nanocrystals with well-defined surface planes by Raman spectroscopy and O₂ adsorption. They observed nanorods have the most abundant intrinsic defect sites, followed by nanocubes and nanooctahedra when ceria was oxidized. In addition, when ceria was reduced, more defect sites induced are clustered on nanorods than on nanocubes. However, very few defect sites can be generated on the nanooctahedra due to its lowest reducibility. Moreover, increasing the proportion of active planes is an efficient approach for enhancing the catalytic activity. Both CeO₂ nanorods and CeO₂ nanowires prepared by Tana et al. [203] predominantly exposed the reactive {110} and {100} planes. However, the CeO₂ nanowires exhibited higher activity for CO oxidation than nanorods because CeO₂ nanowires expose a larger proportion of active planes on the surface. Other 1D and 2D ceria nanomaterials were also assessed in CO oxidation, such as CeO₂ nanotubes, spindle-like CeO₂, platelet-like CeO₂, CeO₂ nanoplates. The CO oxidation properties of nanorods, nanoplates, and nanotubes were investigated by Pan et al. [14]. They observed an enhanced catalytic activity for CeO₂ nanoplates compared with CeO₂ nanotubes and nanorods, and the crystal surfaces (100) of CeO₂ nanoplates were considered to play an important role in determining their catalytic oxidation properties.

In addition, even if the morphology of ceria is the same, the activity can be modulated by synthetic conditions. For example, Liu et al. [188] prepared two nanorod samples by changing the precursor. Sample A with exposed (100) and (111) facets showed lower activity than Sample B with exposed (100) and (110) facets. The difference could be attributed to the difference of oxygen defects. Therefore, in summary, the activity of nanoceria for CO oxidation is dependent on the morphology, due to the differences in specific area, exposed planes and oxygen vacancies.

5.3.3.3 Photocatalysis

The researches regarding to photocatalytic applications of ceria nanomaterials are mainly focused on the degradation of organic pollutants, including methylene blue, methyl orange, Congo red and so on [204–207]. For example, Lu et al. [206] reported the application of ceria nanowire in removing the Congo red. In addition, ceria nanomaterials also exhibit activity in photolysis of water [208]. Lu et al. [208] observed that CeO₂ nanorods with (110) planes showed considerable photocatalytic activity for hydrogen evolution due to their redox capacity.

5.3.3.4 Catalysis of Organic Chemistry Reactions

The redox and acid–base properties of ceria are important factors that allow for activating complex organic molecules and selectively orienting their transformation [189]. Sato’s group conducted significant research on catalytic application of ceria nanomaterials in organic synthesis, including alcohol dehydration reaction, alkylation reaction, condensation reaction and so on. For example, they reported CeO₂ (111) facets have active sites for the reactions of butanediols and of propanoic acid [209]. In addition, ceria nanomaterials also exhibit excellent catalytic properties in redox organic reaction. Lv et al. [210] reported that CeO₂ nanocube with exposed (100) facets showed exclusive selectivity for the oxidation of toluene to benzaldehydes in liquid phase by O₂. CeO₂ nanoplates prepared by Zhang et al. can selectively catalyze the reduction of nitro-aromatic compounds [211, 212].

5.3.3.5 Catalysis for Anti-oxidation

Recently, ceria nanomaterials attract large attention as an excellent anti-oxidant due to the reversible conversion between Ce³⁺ and Ce⁴⁺ [213]. It was reported that ceria nanomaterials can promote the decomposition of superoxide anion (O₂•⁻), hydrogen peroxide (H₂O₂) and hydroxyl radical [213–215]. Biologically, owing to the oxygen defects on the surface acting as sites for free radical scavenging, ceria nanoparticles exhibit excellent anti-oxidation performance in protecting cells from radiation, inflammation, ischemic strokes and so on [26, 216–218]. Kim et al. reported that ceria nanoparticles can protect against ischemic stroke [216]. Chen et al. showed that nanoceria particles can prevent increases in the intracellular concentrations of ROIs in primary cell cultures of rat retina and, in vivo, prevent loss of vision due to light-induced degeneration of photoreceptor cells [217]. The results reported by Nel et al. showed that CeO₂ nanoparticles can suppress the generation of reactive oxygen species (ROS) production, induce cellular resistance to an exogenous source of oxidative stress and protect cells from oxidant injury [218].

Most published studies show that nanoceria is well tolerated in living organisms. However, the bio-response was closely linked to the morphology of ceria. Therefore, the bio-safety of ceria nanomaterials with particular morphology should be concerned, such as nanorods. Nel group conducted the systematic study about the effect of aspect ratio (R) of ceria nanorods on lysosomal damage, cytotoxicity, and 1L-1β production by the human myeloid cell line (THP-1) [26]. Their results showed that short CeO₂ nanorods with R = 1–16 were nontoxic. The nanorods with intermediary aspect ratios (R = 22, 31) did not cause any obvious cell death but did induce IL-1β production. The two nanowire samples with the highest aspect ratio led to significant IL-1β release as well as cell death. They also performed comparative studies in mouse and gastrointestinal tract (GIT) of zebrafish larvae in order to verify this toxicological paradigm for large-aspect-ratio CeO₂ nanoparticles [219].

5.3.3.6 Other Applications

Ceria as catalyst is also applied in other catalytic reactions, such as steam reforming [220], reduction of nitrogen oxide [221], gas sensitivity [222] and so on. Han et al. studied the application of CeO₂ nanotubes in steam reforming reaction [220], and observed that CeO₂ nanotubes showed higher activity than bulk CeO₂. In addition, rare earth composite nanomaterials are also widely used in industrial catalysis acting as catalyst. Besides acting as the active component in catalytic field, CeO₂ possess extensive applications acting as support. For example, the composite catalyst, consisting of CeO₂ and noble metal, was applied in three way catalyst, CO oxidation, reforming reaction, water–gas shift reaction, NO reduction and so on [223–227].

5.4 Perspective

The research field of rare earth based anisotropic nanomaterials has witnessed a rapid development during the past few decades, due to their rich interests in numerous applications. Though tremendous success has been achieved, a few scientific challenges are yet to be addressed:

• (1) Synthesis mechanism. The synthesis mechanism for plenty of systems remain unclear or ambiguous, or simply still in the phase of presumption. This issue is largely owing to the lack of powerful, in situ and real-time characterization tools to monitor and document the reaction process. Therefore, developing some general and highly sensitive on-line monitoring methodologies is urgently necessary for investigating the synthesis mechanism. And undoubtedly, a more in-depth understanding of the synthesis mechanism would help scientists to design synthetic routes for nanostructures with desired morphologies, even those thermodynamically unfavorable ones.

• (2) Synthetic methods. The synthetic methods developed so far generally require stringent conditions, such as high temperature, high pressure and inert gas atmosphere. In addition, quite a few methods have to be conducted in non-aqueous solutions, affording nanoparticles with strong hydrophobicity and poor biocompatibility. And up to date, most synthetic procedures are performed only in the lab in small batches, and can hardly be scaled up to an industrially relevant level. Therefore, the research field still calls for advanced synthetic methods that are green, facile, high-throughput, and readily scalable.

• (3) Defect management. Defects in/on nanocrystals are inevitable, and can have both positive and negative effects. For optical applications, it is generally expected that the phosphors should possess good crystallinity and minimal defects, and display high luminescence efficiency; whereas for catalysis applications, defects in/on nanocrystals can usually serve as active catalytic sites owing to their inherent high-energy states. Therefore, it is still desired to develop novel methods for effectively annihilating or generating defects in/on nanocrystals, whether during or after the synthesis.