The synthesis of rare earthfluoride based nanoparticles

Abstract

The requirement for narrow emission profiles and systems that do not include metals perceived to be cytotoxic, has led to an increase in the reports on rare-earth (Ln) materials, with fluoride-based materials in particular attracting attention. This class of compounds includes rare-earthfluorides (LnF₃), rare-earth oxyfluorides (LnOF) and alkali metalrare-earthfluorides (MLnF₄, where M represents an alkali metal). Here we will describe the main reactions outlining key developments in roughly chronological order.

---

Paula Rahman

Paula Rahman graduated from Imperial College with a MSci in Chemistry in 2005. She then started her PhD in the synthesis of rare-earth nanomaterials in the Physics department at King’s College London.

Mark Green

Mark Green obtained his BSc in chemistry from Manchester Metropolitan University in 1995 and his PhD in nanomaterials chemistry from Imperial College London in 1999. This was followed by post-doctoral appointments at Imperial and Oxford, before taking a position as senior scientist at Oxonica Ltd. He took up a lectureship in Physics at King’s College London in 2004, where he is currently a senior lecturer.

---

Multiple precursor routes

Nanoparticle synthesis, as a major discipline in materials chemistry, has emerged over the last 16 years, and can, in part, be traced back to the seminal reports on the organometallic routes to high-quality semiconductor quantum dots.¹ The main thrust of research has indeed been towards novel semiconducting nanostructures, where high quality materials are now routinely prepared by the thermolysis of inorganic and organometallic precursors in high-boiling-point solvents, with engineered ligands and surfactant systems optimised to control shape, size and optical quality. From this chemistry, further advances were made in the preparation of metals and metal oxides, to the point where now, almost all solid-state materials can be manufactured using chemistry that originated from the 1993 paper by Murray et al.¹ One of the key driving forces in the preparation of, in particular, semiconductor quantum dots, was the development of biological labelling as a real and useful application that could be translated to biology labs. One major limitation in semiconductor quantum dots in biology is their perceived cytotoxicity, and other materials are being investigated as potential replacements—rare-earthfluorides appear to be one such material system.

The rare-earthfluorides are expected to have the same properties as the rare-earth ions, as external ligand fields do not tend to affect the Ln³⁺ ion properties. These properties include very sharp emission lines when excited, and paramagnetic properties. No nanoscale effects change their emission characteristics but it is possible that their magnetic properties will change when the particles are small enough. Another useful property that rare-earthfluorides demonstrate is up-conversion or down-conversion when doped appropriately. Up-conversion is where a material can be excited by two low-energy photons and emit one high-energy photon, down-conversion is the opposite process where one high-energy photon is converted to two low-energy photons.

Rare earth fluoride compounds have been synthesised using most of the available wet chemical routes, with the hydrothermal method being the most investigated. A very basic hydrothermal reaction would involve mixing a rare-earth precursor with a fluoride precursor in aqueous solution, heating it and then separating out the precipitated product (often heating is not required). In order to obtain a more crystalline product, the reagents would be placed in an autoclave (often Teflon-lined to prevent fluoride ions attacking glass) and then sealed and heated.

The growth of the product can also be controlled by the addition of a ligand that binds to the surface of the growing nanoparticles, allowing slow and directed growth. Addition of a ligand often makes the products soluble in either polar or non-polar solvents, depending on the ligand. For example, the polyol method uses a long-chain alcohol as the solvent and as a ligand, the long chain allowing high reaction temperatures, and, as it coats the surface of the nanoparticles, imparts water-solubility.

Most groups, when using the hydrothermal method to synthesise Ln fluoridenanoparticles, use Ln nitrates as precursors, often synthesising the nitrate precursor in situ from oxides and nitric acid. Fewer groups used chlorides and fewer still the oxides directly. The choice of fluoride precursor varies much more than Ln precursors. Groups mainly use HF, NH₄F and NH₄HF₂ when preparing fluorides of the type LnF₃ and NaF or KF for fluorides of the type MLnF_4. Other, less frequently used, precursors include NaBF₄, KBF₄ and 1-butyl, 2-methylimidazolium tetrafluoroborate (an ionic liquid). Occasionally, more than one of these precursors is used, especially when making MLnF₄. The pH of these syntheses is often an important factor in determining the mode of growth and the final product. Groups often use NaOH, KCl, NH₄OH and NH₃ to adjust the pH appropriately. As with most chemical reactions, temperature is a key factor in determining the outcome of a hydrothermal synthesis of LnF₃nanoparticles. Usually, higher temperatures lead to more crystalline particles, but also increase the rate of reaction. Another key factor, of course, is reaction time. Longer reactions tend to give larger nanoparticles and also provide conditions for Ostwald ripening. Some investigations into novel synthetic techniques have also been performed, including microwave irradiation, sonication and the use of ionic liquids. Several microemulsion methods will also be discussed briefly.

The first reported synthesis of a rare-earthfluoride nanomaterial was in 2001 by Zhang et al.² The material was investigated as a potential additive for paraffin when used as a heavy duty lubricant, in which rare-earth-based complexes and micron-sized particles had already been investigated as additives. No heating was required and a ligand was added, resulting in a surface-coated nanoparticle. The nanoparticles produced were relatively polydispersed (Fig. 1) and a succinimide coating allowed the LaF₃nanoparticles to be dissolved easily in paraffin .

Fig. 1 TEM image of LaF₃ nanoclusters. Reproduced with permission from reference 2. Copyright Elsevier, 2001.

A similar report on nanoparticles of LaF₃ was made by Zhou et al. in 2001.³ The synthesis was similar to that of the Zhang et al.,² although this time using pyridinium di-n-hexaoctyldithiophosphate. The group made small (ca. 6 nm) nanoparticles that were dispersible in organic solvents. The particles were in the hexagonal phase but poorly crystalline.

In 2004 the Haase group developed a nanoparticle synthesis for NaLnF₄nanoparticles,⁴ the main aim of which was to make efficient upconvertors that would be transparent when dispersed in solution. This synthesis used N-(2-hydroxyethyl)-ethylenediamine as the solvent and ligand, which also complexed to the rare-earth precursor, and 1-adamantanecarboxylic acid as a ligand. The resulting particles were around 15 nm and crystalline (Fig. 2); unfortunately the particles also had a large size distribution and had ill-defined shapes. The particles formed a transparent solution when dispersed in dimethylsulfoxide (Fig. 2). In 2007 a similar synthesis was performed⁵ but the nanoparticles were modified to be soluble in water. This was done by surface modification after the initial synthesis of the particles with 1-hydroxyethane 1,1-diphosphonic acid. The group found that this surface modification actually made up-conversion efficiency higher.

Fig. 2 TEM image of doped NaYF₄ (left picture) and photograph of the dispersed nanoparticles in DMSO illuminated under UV light (right picture). Reproduced with permission from reference 4. Copyright Wiley VCH, 2004.

Karbowiak et al. investigated the synthesis of nanocrystalline NaGdF₄ and KGdF₄.^6,7 They reported the first hydrothermal synthesis of an alkali metalrare-earthfluoride in 2004, which was doped with Eu³⁺ and the luminescence properties were investigated. The synthesis involved dissolving Ln oxides in concentrated HCl to which NaF or KF was added. The resulting precipitate was aged at 90 °C for up to 10 days. The report on KGdF₄ shows no TEM images but shows data consistent with a nanocrystalline material as ascertained from the X-ray diffraction (XRD) patterns. The NaGdF₄ results do have TEM images which show that the material consists of nanoparticles rather than a bulk material with nano-sized domains. The nanoparticles were uneven in size and shape and agglomerated, but were crystalline. The size and shape of the particles were ill-defined most likely because no ligand was added to the reaction to control growth rate and prevent agglomeration. The Karbowiak group reported more routes in 2005^8,9—the first report investigated various methods for the synthesis of NaGdF₄ including a reverse micelle method and a solid-state reaction.⁹ The results showed that the co-precipitation method previously described was preferable for this type of material. Overall the method was relatively cheap and simple with no compromise in optical properties. The second report described doped KGdF₄.⁸ The main investigation concentrated on the phase transitions of the nanoparticles when heated and the resulting changes in optical properties.

The Li group have looked in-depth into the synthesis of rare-earthfluoride materials including LnF₃ and NaLnF₄. The fluoride products obtained by Wang et al.^10,11 were hollow structures (named inorganic fullerene-like particles by the group) rather than solid particles. This synthesis was a pH-dependent, hydrothermal synthesis where nitrate precursors were formed in situ from Ln₂O₃ and nitric acid. The fluoride source (F-source) was NH₄F and hydroxides were used to change the pH of the solution. Temperatures varied from 80–180 °C and reaction times between 12–24 h. The results showed that an increase in temperature led to an increase in particle size, as expected. The fluoride particles only became hollow after heat treatment in an autoclave at pH between 4 and 5. No further mention of the effect of pH on the products was made and the group did not report a mechanism for the formation of the hollow particles. The products were non-uniform in size but were crystalline. The group developed a new synthetic method, mentioned first in 2005¹² and then in more detail in 2006¹³ based on liquid-solid-solution (LSS) growth, which was employed to synthesise a range of rare-earthfluoride materials (Fig. 3). This hydrothermal method used sodium linoleate, linoleate acid and ethanol added to aqueous solutions of the precursors, rare-earthnitrates and NaF or NH₄HF₂. Reaction temperatures of 100–200 °C for about 8–10 h were used. The resulting nanoparticles varied in shape and size but were generally monodisperse and crystalline. The group discovered that nanoparticle shape depended on the ionic radius of the rare-earth ion which led to the theory that the growth mode changed with ionic size.

Fig. 3 TEM images of NaYF₄ (A, B, C), CeF₃ (D), PrF₃ (E), NdF₃ (F), LaF₃ (G, H), NaYb₂F₇ (I). Reproduced with permission from reference 13. Copyright American Chemical Society, 2006.

Li carried extensive research into the synthesis of NaLnF₄-type nanoparticles. This class of material is of some considerable interest as when doped appropriately, the materials can undergo either up-conversion or down-conversion. In 2005¹⁴ the group used a hydrothermal process with cetyltrimethylammonium bromide (CTAB) and ethylenediaminetetraacetic acid (EDTA) as size- and shape-control agents for the synthesis of NaYF₄. The particles exhibited various shapes and sizes depending on which ligand was added, for example, with CTAB, nanorods were formed, but with EDTA spherical shapes formed. The particles were relatively uniform and the XRD patterns suggest a high level of crystallinity. In 2006¹⁵ attempts to make NaCeF₄ using a similar synthesis showed that at lower temperatures (less than 180 °C) the formation of CeF₃ was favoured, but at higher temperatures the target material was made. Other factors such as the pH of the reactant solution also affected the formation of all the rare-earthfluorides.

In another report¹⁶ the Li group used a solvothermal method to make NaYF₄nanocrystals, where oleic acid and alcohol were added as co-solvents to water along with NaOH. The precursors were Ln nitrates and NaF. Temperatures and reaction times were similar to that of the 2003 syntheses of LnF₃.^10,11 The resulting particles were highly monodispersed and also highly crystalline. This work was further expanded to include a range of reaction parameters.¹⁷ The method yielded many different shapes and sizes for NaYF₄, including nanorods, hexagonal nanoplates, rectangular particles and square particles, all of which were crystalline and monodisperse (Fig. 4). The mechanism of the formation of the differently shaped particles was also investigated and it was found that alcohol chain length, the ratio of F⁻ to Y³⁺, reaction time and temperature were all factors in product shape and size.

Fig. 4 TEM images of NaYF₄nanoparticles doped with 71.4% Eu³⁺ (A), 5% Eu³⁺ (B), 71.4% Tb³⁺ (C), 5% Tb³⁺ (D). Reproduced with permission from reference 17. Copyright American Chemical Society, 2007.

Lezhnina et al. investigated two different types of syntheses,¹⁸ both based on previously reported methods. The first, which is based on Li's method,¹⁰ was modified by initially adjusting the pH to 7 followed by decreasing the pH after the reaction. The EuF₃ products were highly crystalline and cuboidal in shape. This is in comparison to the original synthesis of PrF₃ and LaF₃ which also gave crystalline products, although roughly spherical particles. The second method is based on the method reported by Eiden-Assmann,¹⁹ but used NaF, ethylene glycol and water instead of KF, diethylene glycol and ethanol. The resultant EuF₃ particles were non-spherical in shape, non-uniform in size but crystalline, which matched well with the original synthesis.

Lemyre and Ritcey²⁰ used an unusual method to make YF₃nanoparticles; a microemulsion method based on previously reported syntheses for silica²¹ and Y₂O₃²² which gave particles that were amorphous and ill-defined in shape. The original method used two separate solutions of the precursors; Lemyre and Ritcey however, adjusted the method to only use one solution containing YCl₃ and NH₄HF₂ dissolved in water, igepal CO520 and cyclohexane, which were added together while stirring then sonicated. The resultant particles were highly crystalline, monodisperse and evenly shaped (Fig. 5). Lemyre applied this method to ErF₃ and YbF₃ and reported similar results showing promise as an effective room-temperature method for the synthesis of rare-earthfluorides.

Fig. 5 TEM images of YF₃nanoparticles using the one-solution method. Reproduced with permission from reference 20. Copyright American Chemical Society, 2005.

Zhang reported the synthesis of water-soluble LaF₃, NaYF₄ and NaGdF₄nanoparticles in 2006.²³ The group used a pH-dependent synthesis to make water-soluble particles. Chitosan, a naturally occurring biopolymer, made the particles soluble with functional groups that could be used to attach to other molecules. The method was a pH-dependent, hydrothermal synthesis which used NH₄F as the F-source. Rare-earthchloride precursors were used and ammonia was used to adjust pH. The temperature and time were interesting, only 75 °C for 2 h—these conditions were relatively mild compared to similar syntheses.¹⁰ The particles were crystalline but were disperse in size and shape. Some quick studies into the toxicity of the nanoparticles in the presence of live cells of HT-29 and BSA were favourable, showing no more cell death than a control sample. Zhang's group also investigated the preparation of NaYF₄nanoparticles and the up-conversion properties when doped.²⁴ The group used polyvinylpyrrolidone (PVP) as a ligand to give the nanoparticles good water solubility as well as being a size-control agent. The synthesis also used ethylene glycol and water as solvents, NaCl and rare-earthnitrates as precursors and NH₄F as the F-source. The temperature of the reaction was 160 °C and reaction time 2 h. The NaYF₄ particles were monodisperse in size and crystalline but varied in shape. It was also found that coating the doped nanoparticles in a silica shell did not affect the luminescence properties. Another method for making water-soluble NaYF₄nanoparticles with polyethylenimine (PEI) as the capping agent was also reported, utilising NaCl, NH₄F and Ln chloride as precursors, and water and ethanol as solvents. The reaction time was 24 h and the temperature was 200 °C, significantly longer and higher than the previously described synthesis. The particles were more disperse in size and shape, however. A quick study was performed on the biocompatibility of the nanoparticles with human cells, HT-29, and found that the particles were not toxic. The same synthesis was employed for NaGdF₄nanoparticles except the reaction time was significantly reduced from 24 to 5 h.²⁵ The resulting particles were even in shape and size and also crystalline. Although NaGdF₄nanoparticles have been doped previously, Fig. 6 shows their emissive properties.

Fig. 6 Photo of NaGdF₄ particles dispersed in solution and illuminated under UV, the particles have been doped with Tb³⁺, Eu³⁺, Sm³⁺ and Dy³⁺ for (a), (b), (c) and (d) respectively. Reproduced with permission from reference 25. Copyright Institute of Physics Publishing, 2007.

Two papers from the Zhu^26,27 group reported a synthesis for EuF₃ and LaF₃. In this synthesis, rare-earthnitrates were first made in situ and then mixed with a solution of KBF₄. The mixture was sonicated or stirred but not heated. This resulted in particles which were not quite nano-sized but were small enough to warrant interest. When stirred, disc-shaped particles were formed. The EuF₃, when sonicated, resulted in a very interesting ‘flower’ shape, around 1 micron in diameter.

A similar reaction described in another paper by the same group²⁸ was used to synthesise LaCeF₃nanoparticles. In this case the F-source was changed to NH₄F. The particles were not as interesting in morphology as the KBF₄ reactions, however, as they formed approximately spherical shapes.

Lin's group reported two different syntheses for CeF₃, a polyol method²⁹ and a hydrothermal method.³⁰ The polyol method used Ln nitrate precursors and NH₄F in diethylene glycol (DEG) at 200 °C for 1 h. The hydrothermal method used chloride precursors with NaF. Trisodium citrate was added to form a complex with the Ln ion and acted as a ligand after the nanoparticles had formed. This solution was autoclaved at 180 °C for 24 h. The polyol method resulted in approximately spherical particles which were crystalline. The particles were very small, around 7 nm. The hydrothermal method yielded larger crystalline particles, all of which had flat plate-like structures. The nanoplates tended to have at least one dimension below 100 nm, with larger diameters. The trisodium citrate is the reason the plate-like shapes are formed as the ligand binds strongly to the (0001) surface slowing growth in that direction.

The van Veggel group took the synthesis reported by Dang's group³ and improved it. In this report,³¹ particles that were polydisperse in size and shape, but crystalline, were reported. The synthesis was modified to include a ligand known to impart better solubility; ammonium di-n-octadecyldithiophosphate. The group also made doped LaF₃nanoparticles and investigated the luminescence properties of LaF₃ doped with various other Ln³⁺ ions. The same group also formulated a synthesis which rendered doped LaF₃nanoparticles soluble in water.³² The synthesis was similar to the previous methods with the exception of the ligands; the group tried out several different ligands all with the general formula RO(PO₃)²⁻ where R was PEG-based or –NH₂-based. These ligands acted as a coating on the nanoparticles which controlled growth during the reaction, imparting solubility and preventing agglomeration when in water whilst the particles remained luminescent. The group also tried another synthetic method for making doped LaF₃nanoparticles—a pH dependent co-precipitation method.³³ The ligand used was citrate from citric acid, and the Ln precursors were nitrates, the F-source was NaF and the solvent varied from water to a water–methanol combination. The pH was adjusted using NH₄OH. The reaction conditions were mild; 75 °C and 2 h. From the AFM images reported, the particles were polydisperse in size but the XRD pattern suggested a high level of crystallinity. However, this synthesis was also used to make GdF₃ particles,³⁴ which were disperse in size and shape. The GdF₃nanoparticles were investigated for use as NMR and MRI contrast agents with promising results. The group foresees GdF₃nanoparticles in general use, as it is relatively easy to modify the surface ligands to contain a range of functional groups.

Huang et al. started investigating the synthesis of LnF₃ in 2004.³⁵ The first synthesis made wire-shaped LnF₃, nanosize in two dimensions but several hundred nanometres long in the other. This method was interesting as the synthesis first made very long, thin wires of NH₄LnF₄ (Ln = Nd, Sm, Eu, Gd and Tb) which were then calcined to form LnF₃. This was a solvothermal reaction in methanol with Ln nitrate precursors and NH₄F as the F-source. The LnF₃nanorods were not regular in shape, but were however crystalline. The same group also reported a room temperature synthesis of NaEuF₄nanoparticles in the same year.³⁶ The synthesis was very simple; europium chloride hydrate and NaF were dissolved and stirred in water. The ratio of europium to sodium and reaction time determined the product; either NaEuF₄ or Na₅Eu₉F₃₂. The resultant products were regular in size but were uneven in shape. High-resolution optical microscopy however suggests a high level of crystallinity.

The Chen group investigated this type of reaction more thoroughly. One paper reported the effect of F:Eu ratio when NaBF₄ was used as the F-source.³⁷ The shapes of the nanoparticles changed from nanodiscs, which self-aligned to nanospindles (prolate spheroids). The particles again were uneven but regular in size. A second paper³⁸ investigated the effect of the F-source on the synthesis of EuF₃. Changing the source changed not only the crystal phase of the product, but the shape as well. Potassium fluoride, for example, led to hexagonal EuF₃ particles in crescent-moon shapes while CsF gave orthorhombic nanowires of EuF₃. This reaction is one of the more versatile syntheses for rare-earthfluorides. The process required no heating, only two precursors, and water. Control over morphology could be changed by adjusting the precursor ratio or by using alternative F-sources.

Chen's group also looked at another water-based, room temperature synthesis of NaYF₄.³⁹ The group first complexed the rare-earth-chloride starting material with EDTA and injected the solution rapidly into a separate NaF solution. This, in theory, meant that nucleation occurred over a very small period of time and so limiting the initial size distribution. The resulting particles were uniform in shape and size, most being spherical in shape with diameters in the range 30–45 nm. The most interesting result of this investigation was the effect of the EDTA:Ln³⁺ ratio on particle size. As the relative amount of EDTA was increased, the resultant particles decreased in diameter. The group also found that annealing the samples partly changed the crystal structure from cubic to hexagonal, the amount of hexagonal phase increasing with annealing temperature. The same group looked at a related synthesis for NaYF₄, which was similar to the organometallic-type reactions typically used in semiconductor quantum dot synthesis.⁴⁰ The rare-earth precursor was made into an oleate complex, dissolved in 1-octadecene and mixed with NaF in octadecene. This reaction was performed at 260 °C rather than at room temperature like the previously reported synthesis. The resulting particles were hexagonal-shaped and in the hexagonal phase (Fig. 7) or spherical-shaped in the cubic phase depending on the reaction temperature. The cubic shape formed at the lower temperature of 210 °C and the hexagonal shape formed at 260 °C. The oleic acid bound to the complex also acted as a ligand which imparted solubility to the resultant particles, which did not occur with the EDTA-complexed particles.

Fig. 7 TEM images of hexagonal NaYF₄:Yb, Er nanoplates formed at 260 °C. Reproduced with permission from reference 40. Copyright American Chemical Society, 2006.

Fan et al. came up with a very simple reaction towards doped GdF₃nanoparticles;⁴¹ Ln chlorides were dissolved in ethanol and then NH₄F was added. The solution was heated using an autoclave at temperatures varying from 60–180 °C for 2 h. The particles were irregular in shape, regular in size and more crystalline at higher reaction temperatures. The researchers also investigated the up-conversion properties of GdF₃ doped with Er³⁺ and Yb³⁺ and found that the intensity of emission was much lower than the equivalent bulk material.

Jacob et al. investigated the synthesis of metal fluoridenanoparticles,⁴² including LaF₃ and YF₃. This is one of the few reported syntheses that used ionic liquids and/or a microwave to heat the reaction. The ionic liquid, 1-butyl-3-methylimidazolium tetrafluoroborate (BMIBF₄), acted as the solvent and as the F-source. The synthesis was relatively simple; the rare-earthnitrate and ionic liquid were heated in a domestic microwave for just 5 min. The resulting products were highly crystalline but were disperse in shape and size and exhibited ‘needle’ morphologies.

Kumar et al. used another solvothermal method;⁴³ a solution containing Ln nitrates in water was added drop-wise to an ethylene glycol and water solution of NH₄F and then refluxed. The products were also heat treated after synthesis. The LaF₃nanocrystals were highly crystalline but uneven in shape and size.

Ma et al. developed another microwave synthetic pathway, combined with a pH-adjusted hydrothermal method.⁴⁴Rare-earthnitrates were formed in situ; NaF was added and the pH adjusted to between 4 and 5. The solution was then refluxed in a microwave for 20 min. The resulting particles were around 30 nm in diameter and hollow, approximately spherical and highly crystalline (Fig. 8). The group also analysed the resulting particles before the microwave heating and found that prior to heating, the particles were small, aggregated and solid (not hollow).

Fig. 8 TEM image of PrF₃nanoparticles after microwave heating. Reproduced with permission from reference 44. Copyright Elsevier, 2007.

Miao et al. used an aqueous room-temperature reaction with Ln nitrate precursors and a NaBF₄ F-source to synthesise EuF₃.⁴⁵ This reaction could be modified by changing the ratios of the precursors to give various morphologies such as discs, rods and rings. The rings were composed of aggregated smaller particles consistently throughout the sample. The same synthesis was used to make LaF₃ nanobundles and spherical nanoparticles.

Nuñez and Ocaña⁴⁶ used a very similar ionic liquid to Jacob et al.,⁴²1-butyl, 2-methylimidazolium tetrafluoroborate (BMIMBF₄) to make EuF₃, TbF₃ and YF₃nanoparticles. The ionic liquid was also used as a solvent as well as F-source. Three different forms of Ln precursor were used, nitrate, acetate (OAc) and acetylacetonate (acac). The solvent system was ethylene glycol, the reaction temperature was 120 °C and the reaction time was 15 h. The ytterbium precursors were investigated to find which was most effective. The group found that Y(acac)₃ and Y(OAc)₃ yielded particles which were monodisperse and crystalline. The nitrate precursor gave particles which were irregular in shape and size. Both acac and OAc precursors resulted in flat, rhombohedral-shaped nanoparticles but the acac particles were larger. The acac precursors were used for Eu and Tb and gave approximately spherical and rhombohedral shapes respectively. These particles were less crystalline, although this is more likely due to their small size.

A paper by Fan Zhang et al.⁴⁷ described the synthesis of NaYF₄ particles, and although most results obtained are not small enough to be considered nanoparticles the morphologies are very different from any other reported synthesis. The synthesis used oleic acid, water and ethanol as solvents and NaF and Y(NO₃)₃ as precursors. NaOH was used to create basic conditions. Reactant solutions were heated in an autoclave at temperatures between 130–230 °C for up to 24 h. Different shapes and sizes were obtained by adjusting relative amounts of the reactants and solvents, (Fig. 9) for the unusual flower-decorated hexagonal discs formed in this synthesis.

Fig. 9 Electron micrographs and electron diffraction patterns of NaYF₄ flower-decorated hexagonal discs and hexagonal rods. Reproduced with permission from reference 47. Copyright Wiley VCH, 2007.

Zhang et al. developed another microwave method, using a microemulsion technique.⁴⁸ This method was performed by the preparation of two separate solutions of HF and YCl₃ in cetyltrimethylammonium bromide, cyclohexane and n-pentanol, combining the two and microwaving. The resulting YF₃ particles were large and in the shape of bundles, each strand being itself on the nanoscale. The heating was performed in a domestic microwave for only 3 s as the products required annealing in order to become well-crystallised.

Ansari et al. used a solvothermal method with pyridine as a ligand to synthesise TbF₃nanoparticles.⁴⁹ The precursors, sodium fluoride and terbium nitrate were dissolved in water and ethanol and mixed; the combined solution was kept at 70 °C for 2 h. The particles were approximately rod-shaped but uneven in size and aggregated into small bundles. The TbF₃nanorods showed strong green emission under UV excitation.

The trifluoroacetate precursor method

One interesting route to rare-earthfluorides is based on the decomposition of rare-earth trifluoroacetate salts (LnTFA). In the following section the main reports using trifluoroacetate (TFA) precursors to synthesise rare-earthfluorides alkali rare-earthfluoride and rare-earth oxyfluorides (LnOF) are summarised.

Roberts published a paper in 1961⁵⁰ detailing the preparation of lanthanum and neodymium trifluoroacetate salts, their properties and importantly, the decomposition pathway. After heating, the solid residues were analysed using elemental analysis and the dehydrated forms of rare-earth trifluorides were identified. The decomposition pathway was also investigated by analysing the gaseous products given off when heated, and a reaction was suggested as shown in Fig. 10.

Fig. 10 The decomposition products of neodymium trifluoroacetate.

Roberts found that the trifluoroacetate salts slowly decomposed into the fluoride of the metal when heated above 110 °C and rapidly decomposed when heated to between 250 and 300 °C. If further heating continued to 700 °C the metal oxyfluoride formed. This report suggested that the trifluoroacetate salts of metals could potentially be used as a single source precursor for metal fluorides or metal oxyfluorides.

To the best of our knowledge the first group to use the trifluoroacetate-based precursor to synthesis rare-earthfluoridenanoparticles was Yan's group.⁵¹ Using lanthanum trifluoroacetate and a mixture of oleic acid and octadecene as solvent, the group made trigonal LaF₃ triangular nanoplates which were highly monodisperse and had the tendency to self-align both side-down and face-down (Fig. 11). The nanoparticles were also highly crystalline and dispersible in organic solvents. The reaction temperature and time were 280 °C and 1 h. Not long after Yan's group published the use of trifluoroacetate salts as precursors for rare-earthfluoridenanoparticles, another group from Moscow State University published the synthesis of europium fluoride.^52,53 This synthesis used trioctylphosphine oxide (TOPO) as the solvent and capping ligand and so gave different morphologies to Yan's method. These particles were small and were approximately spherical in shape, although not as crystalline as those produced by Yan's group.

Fig. 11 TEM images of self-aligned LaF₃ nanoplates. Reproduced with permission from reference 51. Copyright American Chemical Society, 2005.

The reaction temperatures of both syntheses were different, in fact the TOPO experiment had a higher temperature (310 °C in comparison to 280 °C). It would be expected that the higher reaction temperature would lead to a more crystalline product, but in these reactions this is not the case. TOPO, it seems, is not a good choice of solvent for the growth of rare-earth particles from the degradation of trifluoroacetate precursors (however a paper by Shan et al. in 2007⁵⁴ also used TOPO, but with more favourable results).

Yan's group went on to develop and explore the use of metal TFA salts as precursors to rare-earthfluorides (LnF₃), rare-earth oxyfluorides (LnOF) and sodium rare-earthfluorides (NaLnF₄). A report published in 2006⁵⁵ thoroughly described the conditions needed to make NaLnF₄ (Ln = Pr–Lu, Y). Using two precursors; sodium trifluoroacetate and rare-earth trifluoroacetate and a combination of three solvents; oleic acid, oleylamine (coordinating solvents) and octadecene (non-coordinating), the group obtained many different shapes and sizes (Fig. 12). By exploring the results when the reaction parameters were changed, the group managed to neatly summarise the conditions needed to synthesise the cubic and hexagonal phases of NaLnF₄ in terms of free energy (Fig. 13). It is worth noting that the group have shown that the cubic phase of NaLnF₄ is stable in the nanoscale regime and by allowing nanocrystals to grow, the change from cubic to hexagonal phases was observed.

Fig. 12 TEM and HRTEM (insets) images of NaYF₄nanorods (a) and (b), NaYF₄ nanoplates (c), NaNdF₄nanorods (d), NaEuF₄nanorods (e), NaHoF₄ nanoplates (f). Reproduced with permission from reference 55. Copyright American Chemical Society, 2005.

Fig. 13 Schematic diagrams of the free energy for the synthesis of NaLaF₄nanoparticles. Where Group I are Pr to Nd, II are SM to Tb and III are Dy to Lu, Y. Reproduced with permission from reference 55. Copyright American Chemical Society, 2005.

The mechanism of the reaction was also examined and it was found that the nucleation time occurred over less than 1 min for α-NaYF₄, which partly explains why the final products are so monodisperse.⁵⁶ The up-conversion properties of doped NaYF₄ were also investigated⁵⁷ as a factor of crystal phase, surface ligand and surface coating and the results showed that phase change can result in different up-conversion pathways. The cubic/α-phase doped with Yb³⁺ and Er³⁺, favours green, which is a two-photon process and the hexagonal/β-phase favours red, which is a three-photon process.

Yan's group also explored, using the same detailed approach, the synthetic conditions needed to make LnOF and LnF₃ (Ln = La to Lu, Y) using LnTFA and the three solvents already mentioned.⁵⁸ They found that the solvent combination was important when it came to the synthesis of LnF₃ or LnOF; the most important conclusion being that oleylamine is a fluorophile and encourages formation of LnOF rather than LnF₃. The group again demonstrated excellent control over morphology of the nanoparticles and the ability of the product to self-assemble over a long range, Fig. 14, demonstrates the range of shapes and self-alignment possibilities. The group looked further into the synthesis of the oxyfluorides,⁵⁹ specifically doped LaOF and GdOF, using TFA precursors. The group detailed the conditions needed for different shapes and sizes and also investigated the optical properties of the particles. The group found that LaOF is a better host for green emission (Ce³⁺ and Tb³⁺ doped) but GdOF is better for red emission (Eu³⁺ doped).

Fig. 14 TEM and HRTEM (insets) images of GdF₃ (a), (b) and (c), YF₃ (d) and (e), HoF₃ (f) and YbF₃ (g) nanoparticles. Reproduced with permission from reference 58. Copyright Wiley VCH, 2007.

At the same time as Yan's group, another group reported the synthesis of doped NaYF₄ using TFA precursors in oleylamine at 330 °C.⁶⁰ The group made the more efficient up-converting hexagonal phase and found oleylamine enables the phase transition to the hexagonal form. The particles were also relatively small, which had been difficult to achieve before even when using other synthetic approaches. As pointed out by Yi and Chow,⁶⁰ this is much better for use in biological applications, where particles smaller than 5.5 nm are preferred for renal clearance.⁶¹ The synthesis did not require the cubic phase of NaYF₄ as a starting material, just the appropriate TFA salts. The group also found that the up-conversion efficiency was around one order of magnitude less than that of the bulk. Additionally the particles were made water soluble by exchanging ligand with a PEG diacid.

Yi and Chow also investigated the effect of growing a shell of NaYF₄ on doped NaYF₄nanoparticles and found a significant improvement in up-conversion emission.⁶² The group also developed another method to make the particles water-soluble by attaching a poly(acrylic acid) in addition to the shell, which had the advantage of offering the ability to attach to biomolecules. Further work investigated the synthesis of alkali metalrare-earthfluorides with alternatives to sodium.⁶³ The paper describes the synthesis of LiYF₄ and BaYF₄ using TFA precursors, except for the barium source which was barium acetoacetonate (barium trifluoroacetate was not soluble in the solvents). The synthesis of LiYF₄ also involved adding trifluoroacetic acid in addition to the TFA precursors. The products exhibited good up-conversion emission when doped, and the particles themselves were monodisperse and crystalline. A brief study showed that the trifluoroacetic acid concentration and reaction temperatures were the more important influences on the outcome of the reactions, for example, the presence of trifluoroacetic acid and the relatively high temperature of 340 °C were necessary for the formation of pure LiYF₄, otherwise YOF was formed as a side product.

In a communication, Capobianco reported using TFA precursors for the synthesis of doped, cubic NaYF₄⁶⁴ based on the report by Zhang et al.⁵¹ The particles themselves were polydisperse in size but crystalline and displayed red, green and blue up-conversion luminescence when doped. The group extended the synthesis⁶⁴ and modified it to yield NaYF₄ particles that were highly monodisperse, self-assembled and crystalline (Fig. 15).⁶⁵ The change to the synthesis simply modified the method of delivery of the Ln precursors, which were first dissolved in the appropriate solvents and added slowly through a cannula into another solution at the required reaction temperature containing the rest of the precursors. Furthermore, Capobianco's group showed⁶⁶ that coatingNaGdF₄ (doped with Ce³⁺ and Tb³⁺) in a NaYF₄ shell prevents the oxidation of the Ce³⁺ to Ce⁴⁺ and improved luminescence from Tb³⁺. In this system, NaGdF₄ was the host material and the Ce³⁺ acted as a sensitiser which absorbed UV light and transferred the energy to Tb³⁺ allowing the particles to emit green fluorescence. The synthetic method itself was not however novel, but yielded high-quality particles.

Fig. 15 TEM images of NaYF₄:Er, Yb nanoparticles. Reproduced with permission from reference 65. Copyright American Chemical Society, 2007.

Liu and Chen tried a different approach to making rare-earthfluorides using TFA precursors. The group developed⁶⁷ a solvothermal method to synthesise LaF₃nanoparticles in the hexagonal phase. The experiment was based on a versatile general approach described by Wang et al.¹² and involved adding the TFA precursors dissolved in water to a mixture of sodium oleate, oleylamine and ethanol, mixing, and then heating in an autoclave for 24 h at 190 °C. While this experiment was carried out at much lower temperatures than the conventional approach for this type of synthesis, it had a significantly longer reaction time. The products themselves were crystalline and were soluble in organic solvents, but tended to be relatively large and asymmetric in shape.

In 2007, another group attempted to use TOPO as the solvent when synthesising NaYF₄ using TFA precursors.⁵⁴ The resulting particles were not only crystalline and monodisperse but also in the hexagonal phase. The particles were also small enough to be used as biological labels (around 11 nm). TOPO has a higher boiling point than the usual three solvents used in TFA syntheses so it has the potential to produce the hexagonal phases of this class of material with much more ease.

In conclusion, the various techniques to synthesise rare-earthfluoridenanoparticles have all demonstrated control over shape and size. The methods using trifluoroacetate precursors in particular have demonstrated ease of control over the morphology of the Ln fluorides using three key solvents (oleylamine, oleic acid and octadecene), whilst other factors such as temperature and precursor concentrations also allow control over size and shape. The desire for nanoparticles with no heavy metals, narrow emission profiles and a narrow size distribution has lead to a wealth of research in this materials system,^68,69 with many more applications expected in the up-coming years.

Acknowledgements

We acknowledge the EPSRC for support into rare-earth nanomaterials (EP/D501563). We thank the ACS, Wiley VCH, IoP publishing and Elsevier for kind use of the figures.

References

1. C. B. Murray, D. J. Norris and M. G. Bawendi, J. Am. Chem. Soc., 1993, 115, 8706.
2. Z. F. Zhang, L. G. Yu, W. M. Liu and Q. J. Xue, Tribol. Int., 2001, 34, 83.
3. J. F. Zhou, Z. S. Wu, Z. J. Zhang, W. M. Liu and H. X. Dang, Wear, 2001, 249, 333.
4. S. Heer, K. Kompe, H. U. Gudel and M. Haase, Adv. Mater., 2004, 16, 2102.
5. H. Schäfer, P. Ptacek, K. Kömpe and M. Haase, Chem. Mater., 2007, 19, 1396.
6. M. Karbowiak, A. Mech, A. Bednarkiewicz and W. Strek, J. Alloys Compd., 2004, 380, 321.
7. A. Mech, M. Karbowiak, L. Kepinski, A. Bednarkiewicz and W. Strek, J. Alloys Compd., 2004, 380, 315.
8. M. Karbowiak, A. Mech, L. Kepinski, W. Mielcarek and S. Hubert, J. Alloys Compd., 2005, 400, 67.
9. M. Karbowiak, A. Mech, A. Bednarkiewicz, W. Strek and L. Kepinski, J. Phys. Chem. Solids, 2005, 66, 1008.
10. X. Wang and Y. D. Li, Chem.–Eur. J., 2003, 9, 5627.
11. X. Wang and Y. D. Li, Angew. Chem., Int. Ed., 2003, 42, 3497.
12. X. Wang, J. Zhuang, Q. Peng and Y. D. Li, Nature, 2005, 437, 121.
13. X. Wang, J. Zhuang, Q. Peng and Y. D. Li, Inorg. Chem., 2006, 45, 6661.
14. J. H. Zeng, J. Su, Z. H. Li, R. X. Yan and Y. D. Li, Adv. Mater., 2005, 17, 2119.
15. J. H. Zeng, Z. H. Li, J. Su, L. Wang, R. Yan and Y. Li, Nanotechnology, 2006, 17, 3549.
16. L. Y. Wang and Y. D. Li, Nano Lett., 2006, 6, 1645.
17. L. Y. Wang and Y. D. Li, Chem. Mater., 2007, 19, 727.
18. M. M. Lezhnina, T. Justel, H. Katker, D. U. Wiechert and U. H. Kynast, Adv. Funct. Mater., 2006, 16, 935.
19. S. Eiden-Assmann and G. Maret, Mater. Res. Bull., 2004, 39, 21.
20. J. L. Lemyre and A. M. Ritcey, Chem. Mater., 2005, 17, 3040.
21. F. J. Arriagada and K. Osseo-Asare, J. Colloid Interface Sci., 1999, 211, 210.
22. M. H. Lee, S. G. Oh and S. C. Yi, J. Colloid Interface Sci., 2000, 226, 65.
23. F. Wang, Y. Zhang, X. P. Fan and M. Q. Wang, Nanotechnology, 2006, 17, 1527.
24. Z. Li and Y. Zhang, Angew. Chem., Int. Ed., 2006, 45, 7732.
25. F. Wang, X. P. Fan, M. Q. Wang and Y. Zhang, Nanotechnology, 2007, 18, 025701.
26. L. Zhu, X. M. Liu, J. Meng and X. Q. Cao, Cryst. Growth Des., 2007, 7, 2505.
27. L. Zhu, I. Meng and X. Q. Cao, Eur. J. Inorg. Chem., 2007, 3863.
28. L. Zhu, J. Meng and X. Q. Cao, J. Nanopart. Res., 2008, 10, 383.
29. Z. L. Wang, Z. W. Quan, P. Y. Jia, C. K. Lin, Y. Luo, Y. Chen, J. Fang, W. Zhou, C. J. O'Connor and J. Lin, Chem. Mater., 2006, 18, 2030.
30. C. Li, X. Liu, P. Yang, C. Zhang, H. Lian and J. Lin, J. Phys. Chem. C, 2008, 112, 2904.
31. J. W. Stouwdam and F. van Veggel, Nano Lett., 2002, 2, 733.
32. P. R. Diamente and F. van Veggel, J. Fluoresc., 2005, 15, 543.
33. V. Sudarsan, S. Sivakumar, F. van Veggel and M. Raudsepp, Chem. Mater., 2005, 17, 4736.
34. F. Evanics, P. R. Diamente, F. van Veggel, G. J. Stanisz and R. S. Prosser, Chem. Mater., 2006, 18, 2499.
35. B. Huang, Z. Liu, J. M. Hong, X. T. Chen, Z. L. Xue and X. Z. You, J. Cryst. Growth, 2005, 276, 613.
36. M. Wang, Q. L. Huang, J. M. Hong, W. H. Wu, Z. Yu, X. T. Chen and Z. L. Xue, Solid State Commun., 2005, 136, 210.
37. M. Wang, Q. L. Huang, J. M. Hong, X. T. Chen and Z. L. Xue, Cryst. Growth Des., 2006, 6, 1972.
38. M. Wang, Q. L. Huang, J. M. Hong, X. T. Chen and Z. L. Xue, Cryst. Growth Des., 2006, 6, 2169.
39. G. S. Yi, H. C. Lu, S. Y. Zhao, G. Yue, W. J. Yang, D. P. Chen and L. H. Guo, Nano Lett., 2004, 4, 2191.
40. Y. Wei, F. Q. Lu, X. R. Zhang and D. P. Chen, Chem. Mater., 2006, 18, 5733.
41. X. P. Fan, D. B. Pi, F. Wang, J. R. Qiu and M. Q. Wang, IEEE Trans. Nanotechnol., 2006, 5, 123.
42. D. S. Jacob, L. Bitton, J. Grinblat, I. Felner, Y. Koltypin and A. Gedanken, Chem. Mater., 2006, 18, 3162.
43. G. A. Kumar, C. W. Chen, J. Ballato and R. E. Riman, Chem. Mater., 2007, 19, 1523.
44. L. Ma, W. X. Chen, Y. F. Zheng, J. Zhao and Z. D. Xu, Mater. Lett., 2007, 61, 2765.
45. Z. J. Miao, Z. M. Liu, K. L. Ding, B. X. Han, S. D. Miao and G. M. An, Nanotechnology, 2007, 18, 125605.
46. N. O. Nuñez and M. Ocaña, Nanotechnology, 2007, 18, 455606.
47. F. Zhang, Y. Wan, T. Yu, F. Q. Zhang, Y. F. Shi, S. H. Xie, Y. G. Li, L. Xu, B. Tu and D. Y. Zhao, Angew. Chem., Int. Ed., 2007, 46, 7976.
48. J. S. Zhang, W. P. Qin, J. S. Zhang, Y. Wan, C. Y. Cao, Y. Jin, G. D. Wei, G. F. Wang and L. L. Wang, Chem. Res. Chin. Univ., 2007, 23, 733.
49. A. A. Ansari, N. Singh and S. P. Singh, J. Nanopart. Res., 2008, 10, 703.
50. J. E. Roberts, J. Am. Chem. Soc., 1961, 83, 1087.
51. Y. W. Zhang, X. Sun, R. Si, L. P. You and C. H. Yan, J. Am. Chem. Soc., 2005, 127, 3260.
52. N. G. Zhuravleva, A. A. Eliseev, N. A. Sapoletova, A. V. Lukashin, U. Kynast and Y. D. Tretyakov, Mater. Sci. Eng., C, 2005, 25, 549.
53. N. G. Zhuravleva, N. A. Sapoletova, A. A. Eliseev, A. V. Lukashin, Y. D. Tretyakov and U. Kynast, Opt. Mater., 2006, 28, 606.
54. J. Shan, X. Qin, N. Yao and Y. Ju, Nanotechnology, 2007, 18, 445607.
55. H. X. Mai, Y. W. Zhang, R. Si, Z. G. Yan, L. D. Sun, L. P. You and C. H. Yan, J. Am. Chem. Soc., 2006, 128, 6426.
56. H. X. Mai, Y. W. Zhang, L. D. Sun and C. H. Yan, J. Phys. Chem. C, 2007, 111, 13730.
57. H. X. Mai, Y. W. Zhang, L. D. Sun and C. H. Yan, J. Phys. Chem. C, 2007, 111, 13721.
58. X. Sun, Y. W. Zhang, Y. P. Du, Z. G. Yan, R. Si, L. P. You and C. H. Yan, Chem.–Eur. J., 2007, 13, 2320.
59. Y. P. Du, Y. W. Zhang, L. D. Sun and C. H. Yan, J. Phys. Chem. C, 2008, 112, 405.
60. G. S. Yi and G. M. Chow, Adv. Funct. Mater., 2006, 16, 2324.
61. H. Soo Choi, W. Liu, P. Misra, E. Tanaka, J. P. Zimmer, B. Itty Ipe, M. G. Bawendi and J. V. Frangioni, Nat. Biotechnol., 2007, 25, 1165.
62. G. S. Yi and G. M. Chow, Chem. Mater., 2007, 19, 341.
63. G. S. Yi, W. B. Lee and G. M. Chow, J. Nanosci. Nanotechnol., 2007, 7, 2790.
64. J. C. Boyer, F. Vetrone, L. A. Cuccia and J. A. Capobianco, J. Am. Chem. Soc., 2006, 128, 7444.
65. J. C. Boyer, L. A. Cuccia and J. A. Capobianco, Nano Lett., 2007, 7, 847.
66. J. C. Boyer, J. Gagnon, L. A. Cuccia and J. A. Capobianco, Chem. Mater., 2007, 19, 3358.
67. C. H. Liu and D. P. Chen, J. Mater. Chem., 2007, 17, 3875.
68. Z. G. Yan and C. H. Yan, J. Mater. Chem., 2008, 18, 5046.
69. J. Shen, L. D. Sun and C. H. Yua, Dalton Trans., 2008, 5687.

---