Liquid‐Phase Synthesis of Highly Reactive Rare‐Earth Metal Nanoparticles

Abstract

The first liquid‐phase synthesis of high‐quality, small‐sized rare‐earth metal nanoparticles (1–3 nm)—ranging from lanthanum as one of the largest (187 pm) to scandium as the smallest (161 pm) rare‐earth metal—is shown. Size, oxidation state, and reactivity of the nanoparticles are examined (e.g., electron microscopy, electron spectroscopy, X‐ray absorption spectroscopy, selected reactions). Whereas the nanoparticles are highly reactive (e.g. in contact to air and water), they are chemically stable as THF suspensions and powders under inert conditions. The reactivity can be controlled to obtain inorganic and metal–organic compounds at room temperature.

Reactive but easy to handle: For the first time, rare‐earth metals (La, Ce, Sm, Eu, Tm, Y, Sc) are obtained as high‐quality nanoparticles (1–3 nm in size) in the liquid phase. They can be highly interesting as reactive starting materials to realize new rare‐earth metal compounds and/or phase‐pure rare‐earth‐metal‐based materials.

Rare‐earth metals are highly reactive, in principle.^[1] Performing reactions with rare‐earth metals in a controlled manner is nevertheless difficult. On the one hand, the surface area of the bulk metals is small and crushing them into pieces is difficult due to their hardness and sensitivity. On the other hand, the surface of rare‐earth metals is usually passivated by hydroxides, oxides, and/or carbonates, which kinetically inhibit reactions at low temperature. In chemical syntheses, these passivation layers naturally also cause impurities. However, a purification of rare‐earth metals is elaborate and requires high‐temperature distillation (>800 °C), special equipment (e.g. steel cylinders), and profound experience.^[2] In contrast to alkali metals and alkali earth metals, rare‐earth metals also show only low solubility in liquid ammonia.^[3] As a result, reactions with zerovalent rare‐earth metals are often limited to high‐temperature solid‐state reactions (>500 °C).[¹, ⁴]

The availability of rare‐earth metals in the form of small‐sized nanoparticles in suspension or as powder sample could significantly change the aforementioned situation and allow an easy handling with reactions of rare‐earth metals near room temperature (≤100 °C) and/or directly in the liquid phase. A liquid‐phase synthesis of rare‐earth metal nanoparticles, however, is largely unknown. This can be ascribed to the reactivity of such nanoparticles, which needs to be expected as considerably higher than for the respective bulk rare‐earth metals. So far, few reports address physical methods to obtain rare‐earth metal particles (e.g., γ‐radiation, laser ablation, plasma treatment, electrodeposition),^[5] which, however, result in heavily agglomerated particles with broad size distribution and significant oxygen contamination. Some reports claim aqueous syntheses or even biosyntheses,^[6] which seems hardly traceable when considering the position of the rare‐earth metals in the voltage series with electrochemical potentials of about −2.4 V.^[7] Our experience is the observation of violent reactions with explosion if the rare‐earth metal nanoparticles come into contact with water.

The dilemma of the high reactivity can be exemplarily illustrated for Gd⁰ nanoparticles. Here, the only reliable liquid‐phase synthesis relates to the reduction of GdCl₃ with crown‐ether‐stabilized alkalides (e.g. [K(18‐crown‐6]⁺[K]⁻) as powerful reducing agents.^[8] Due to the high reactivity of the Gd⁰ nanoparticles, the authors could only identify Gd₂O₃ nanoparticles via electron microscopy. Recently, they could validate the formation of Gd⁰ for Gd@Au core–shell nanoparticles with a dense gold shell to protect the Gd⁰ core from oxidation.^[8a] With La⁰, Ce⁰, Sm⁰, Eu⁰, Tm⁰, Y⁰, and Sc⁰, we now present a reliable liquid‐phase synthesis of high‐quality, small‐sized rare‐earth metal nanoparticles ranging from lanthanum as a large (187 pm) to scandium as the smallest (161 pm) rare‐earth metal for the first time (Figure 1 a).^[9]

Figure 1.

Rare‐earth metals: a) Row of metals with atomic radii (red‐marked metals prepared here, *aperiodic radii due to half‐filled f ⁷ and filled f ¹⁴ electron shell). b) Scheme of the exemplary liquid‐phase synthesis of Sm⁰ nanoparticles.

Several prerequisites need to be addressed to obtain high‐quality nanoparticles of such highly reactive metals. This includes a powerful reducing agent as well as a sample handling under strict inert conditions, starting from synthesis to all analytical characterization. Based on our experience with base metal nanoparticles (e.g., Ti⁰, Mo⁰, W⁰, Re⁰, Fe⁰, Zn⁰) and different synthesis strategies (e.g., microemulsions with liquid ammonia, sodium‐driven reduction in liquid ammonia, lithium/sodium naphthalenide‐driven reduction in tetrahydrofuran, lithium/sodium pyridinyl‐driven reduction in pyridine),^[10] we have selected lithium naphthalenide ([LiNaph]) as the reducing agent of choice to obtain the even more reactive rare‐earth metals. Simple metal halides were used as low‐cost, well‐soluble starting materials (see SI). All syntheses were performed in Schlenk‐type vessels; sample handling and centrifugation were carried out in gloveboxes; suitable transfer systems were used for all analytical characterization (SI: Figure S1).

The synthesis strategy is exemplarily illustrated for Sm⁰ nanoparticles (Figure 1 b). Accordingly, SmI₂(THF)₂ was dissolved in tetrahydrofuran (THF). Moreover, lithium and naphthalene were dissolved in THF. Thereafter, the resulting [LiNaph] solution was injected with intense stirring into the SmI₂(THF)₂ solution. The reduction and nucleation of Sm⁰ nanoparticles are indicated by the immediate formation of a deep black suspension after injection of [LiNaph]. Electron paramagnetic resonance (EPR) spectra confirm the completeness of the reaction (SI: Table S1, Figure S2). Thus, the ESR signal of the paramagnetic [LiNaph] vanishes completely upon reduction of SmI₂(THF)₂. Subsequent to the reaction, the Sm⁰ nanoparticles were separated by high‐power centrifugation (25.000 rpm, 55.200×g), purified by resuspension/centrifugation in/from THF or toluene. The as‐prepared nanoparticles can be easily redispersed in THF or toluene, or they can be dried in vacuum at room temperature to obtain powder samples.

Size, shape, and size distribution of the as‐prepared Sm⁰ nanoparticles were evaluated by transmission electron microscopy (TEM) (Figure 2). Overview TEM images show uniform, non‐agglomerated nanoparticles with narrow size distribution (Figure 2 a,b). A statistical evaluation of 250 nanoparticles on TEM images results in a mean diameter of 1.7±0.2 nm (Figure 2 c). High‐resolution (HR)TEM images confirm a size of 1–3 nm (Figure 2 d). Moreover, HRTEM indicates the crystallinity of the as‐prepared Sm⁰ nanoparticles and shows highly parallel lattice fringes (Figure 2 d). The observed lattice fringe distance of 1.9 Å is well compatibel with hexagonal β‐Sm⁰ (d₁₀₅ with 1.9 Å).^[11] The presence of zerovalent Sm is further confirmed by Fourier transformation (FT) analysis of HRTEM images of single particles (Figure 2 e), which agrees with the calculated diffraction pattern of hexagonal bulk‐Sm⁰ (space group: P6₃/mmc, a=3.682, c=11.043 Å) in the [‐551] zone axis.^[11] We also observed few particles with disordered structure (Figure 2 b), which could be caused by electron‐beam‐induced knock‐on damage. The formation of very small Sm⁰ nanoparticles might be surprising at first sight, but is well in agreement with the LaMer‐Dinegar model of particle nucleation and growth.^[12] Thus, the injection of [LiNaph] initiates a fast reduction and formation of metal nanoparticles, which are highly insoluble in THF. The resulting high oversaturation promotes a fast nucleation and, therefore, very small particles.

Figure 2.

Electron microscopy of the as‐prepared Sm⁰ with a,b) TEM images at different levels of magnification (red dots indicate single nanoparticles), c) size distribution (based on statistical evaluation of 250 nanoparticles on TEM images), d) HRTEM image with lattice fringes, e) FT‐HRTEM image in (c) with reflection positions of calculated diffraction pattern in [‐551] zone axis indicated by red circles.

In addition to HRTEM and FFT analysis, the oxidation state of the as‐prepared Sm⁰ nanoparticles was validated by X‐ray absorption near edge structure (XANES) spectroscopy, electron energy loss spectroscopy (EELS), and energy‐dispersive X‐ray spectroscopy (EDXS). Sm‐XANES L₃ ‐edge spectra of the as‐prepared nanoparticles were compared to bulk samarium, SmI₂(THF)₂, and literature data of Sm⁰ clusters isolated in argon matrices (Figure 3, SI: Figure S3,S4). Particularly striking is the high intensity pre‐edge peak of SmI₂(THF)₂ in the L₃ ‐edge spectra, which can be ascribed to a 2p ⁶…4f ⁶ 6s⁰ → 2p ⁵…4f ⁷ 6s⁰ transition on Sm^+II (Figure 3 a). This pre‐edge peak is totally missing for the nanoparticles. In comparison to bulk Sm⁰, however, the Sm⁰ nanoparticles exhibit a double‐peak structure (6708, 6716 eV). This double‐peak structure was similarly reported for Ar‐matrix‐isolated samarium clusters and ascribed to a different valence electron configuration of volume and surface Sm atoms (Figure 3 b).^[13] Thus, Sm atoms in the cluster volume exhibit a 4f ⁵(5d 6s)³ configuration, whereas Sm atoms located at the cluster surface have a 4f ⁶ 6s² configuration with a shift of 8 eV between the two peaks (Figure 3 a,b).^[13] The intensity of both peaks depends on the size of the cluster. For the as‐prepared Sm⁰ nanoparticles with an average size of 1.7±0.2 nm about 40 % of the Sm atoms are located at the surface, so that a significant contribution of surface atoms in the XANES spectra is not a surprise. Taken together, XANES spectra prove the presence of small‐sized, zerovalent Sm nanoparticles, which was also confirmed by EELS and EDXS (SI: Figure S5,S6).

Figure 3.

XANES spectra of the as‐prepared Sm⁰ nanoparticles: a) Sm‐L₃ spectra with bulk metal and SmI₂(THF)₂ as references. b) Comparison with Ar‐matrix‐isolated Sm clusters.^[13a] c) Scheme of Sm⁰ nanoparticle with different electron configuration of volume atoms (indicated by *) and surface atoms (indicated by **) according to literature data.^[13]

Similar to Sm⁰, size, size distribution, and crystallinity of the as‐prepared La⁰, Ce⁰, Eu⁰, Tm⁰, Y⁰, and Sc⁰ nanoparticles were also examined by TEM (Figure 4, SI: Figure S7–S12). Overview TEM images again show uniform, non‐agglomerated nanoparticles with predominantly spherical shape (Figure 4 a–f) and mean diameters of 2–3 nm at very narrow size distribution (±0.3 nm) (Figure 5). HRTEM images prove the crystallinity of the as‐prepared nanoparticles and show highly parallel lattice fringes (Figure 4 a–f). The observed lattice fringe distances are well compatible with bulk metals (La⁰: d _01‐1 with 3.265 Å, Ce⁰: d _1‐1‐1 with 3.071 Å, Eu⁰: d₀₀₂ with 2.695 Å, Tm⁰: d _11‐1 with 2.693 Å, Y⁰: d _1‐1‐1 with 2.785 Å, Sc⁰: d _1‐1‐1 with 2.51 Å).^[14] Crystallinity and crystal structure of the as‐prepared rare‐earth metal nanoparticles were also confirmed by FT analysis (SI: Figure S7–S12). The crystallographic data are not only consistent with the bulk metals, they also differ clearly from the corresponding metal oxides (SI: Table S2).

Figure 4.

Electron microscopy of the as‐prepared La⁰, Ce⁰, Eu⁰, Tm⁰, Y⁰, Sc⁰ with overview TEM images (red dots indicate single nanoparticles) and HRTEM images with lattice fringes, see SI for more data: Figure S7–S12).

Figure 5.

Size distribution of the as‐prepared La⁰, Ce⁰, Eu⁰, Tm⁰, Y⁰, and Sc⁰ nanoparticles (statistical evaluation of 250 nanoparticles on TEM images).

With La⁰ as a large and Sc⁰ as the smallest rare‐earth metal, the synthesis strategy is suitable to obtain all rare‐earth metals as nanoparticles in the liquid phase. Having high‐quality nanoparticles as a new form of rare‐earth metals available for the first time, we also got a first glimpse on their reactivity and their use as starting materials in chemical synthesis. Under inert conditions (argon, nitrogen, vacuum), the as‐prepared suspensions (THF, toluene) and powder samples are chemically very stable. Attention must be paid if the nanoparticles get into contact with oxidizing agents. Again, we take the Sm⁰ nanoparticles as example. Here, direct contact of a powder sample (≈20 mg) to air causes an immediate reaction with sting flame (Figure 6 a).

Figure 6.

Reactivity of the as‐prepared Sm⁰ nanoparticles: a) uncontrolled, violent reaction in air (20 mg Sm⁰ powder), b) controlled reactions with NH₃, N₂H₄, S₈, I₂, EtOH, HCp at 25 °C in the liquid phase (Sm⁰ suspensions in THF).

Beside the uncontrolled reaction of the as‐prepared Sm⁰ powders in air, we have exemplarily reacted the Sm⁰ nanoparticles with NH₃, N₂H₄, S₈, I₂, ethanol (EtOH), and cyclopentadiene (HCp) at room temperature (25 °C) in the liquid phase (Figure 6 b, SI: Figure S13–S26). The respective reactions are already indicated by the change of the deep black Sm⁰ suspension to colorless, yellowish or red solutions/suspensions and result in different inorganic and metal‐organic ocmpounds, including Sm(NH₂)₃(THF), Sm(N₂H₃)₃(THF), SmS_1.8, SmI₃(THF)₃, Sm(OEt)₃, and Sm(Cp)₃ (SI: Figure S14–S26). After separation from the liquid‐phase, Sm(NH₂)₃(THF) and SmS_1.8 can be thermally converted to SmN and Sm₂S₃ (SI: Figure S15,S16,S19,S20). Depending on the respective reaction and compound, nanoparticles (SmS_1.8, SI: Figure S18) or even single crystals (SmI₃(THF)₃, Sm(Cp)₃, SI: Table S3, Figure S21,S22,S25) can be obtained. Moreover, Sm(N₂H₃)₃(THF) and SmI₃(THF)₃ are already new compounds (SI: Figure S17,S22). The reaction of the Sm⁰ nanoparticles with N₂H₄ resulted in an unusual metal hydrazide, which is even more reactive than the intial Sm⁰ (see SI). Such reaction was yet only reported for alkali metals^[15] and points to the reactivity of the rare‐earth metal nanoparticles. Even more interesting than each of the aforementioned follow‐up reactions, however, might be the general option to use the as‐prepared rare‐earth metal nanoparticles in a controlled manner as starting materials in chemical syntheses near room temperature (≤100 °C) and/or instantaneously in the liquid phase.

For all kinds of liquid‐phase syntheses, it is a matter of course that the as‐prepared nanoparticles contain at least the solvent as surface capping. In this regard, FT‐IR spectra show THF to be adsorbed on the surface of the as‐prepared nanoparticles (SI: Figure S13). Taking the high reactivity of the rare‐earth metal nanoparticles into account, however, it is a surprise as such that the nanoparticles are chemically stable in suspension. Moreover, a small‐sized, low‐weight molecule like THF as the surface capping is beneficial for follow‐up reactions as it can be easily exchanged by other, stronger‐binding ligands, which then can be chosen independent from the original nanoparticle synthesis. Our exemplary reactions already indicate products that do not contain THF (e.g., SmS_1.8, Sm(OEt)₃, Sm(Cp)₃).

In sum, a reliable liquid‐phase synthesis of rare‐earth metal nanoparticles is shown for the first time with La⁰, Ce⁰, Sm⁰, Eu⁰, Tm⁰, Y⁰, and Sc⁰ as examples. The synthesis uses metal halides as low‐cost starting materials, THF as conventional solvent, and [LiNaph] as a powerful, easy‐to‐handle reducing agent. The as‐prepared rare‐earth metal nanoparticles exhibit very small sizes (1–3 nm) and narrow size distributions (±0.3 nm). They are monocrystalline and well‐dispersible in THF or toluene. Whereas the rare‐earth metal nanoparticles are chemically stable under inert conditions, they show heavy reactivity, for instance, in contact to air. However, the reactivity can be well‐controlled, so that the rare‐earth metal nanoparticles can be exemplarily reacted with NH₃, N₂H₄, S₈, I₂, EtOH and HCp to obtain different inorganic and metal‐organic compounds at room temperature and instantaneously in the liquid phase. Having the first liquid‐phase synthesis of rare‐earth metal nanoparticles established, these nanoparticles might become most interesting as starting materials to prepare new and/or high‐purity rare‐earth metal compounds as well as rare‐earth‐metal‐based materials.

Conflict of interest

The authors declare no conflict of interest.

Supporting information

As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer reviewed and may be re‐organized for online delivery, but are not copy‐edited or typeset. Technical support issues arising from supporting information (other than missing files) should be addressed to the authors.

Supporting Information

D. Bartenbach, O. Wenzel, R. Popescu, L.-P. Faden, A. Reiß, M. Kaiser, A. Zimina, J.-D. Grunwaldt, D. Gerthsen, C. Feldmann, Angew. Chem. Int. Ed. 2021, 60, 17373.