Future prospects of fluoride based upconversion nanoparticles for emerging applications in biomedical and energy harvesting

Rare earth doped/codoped phosphors have been extensively studied for different types of applications based on their intense luminescence features. For this, researchers have tried to choose the inorganic host matrices having both a low phonon cut-off frequency and a high refractive index. Many articles have been published on oxide based phosphor materials, but due to their high cut-off phonon frequency, use of these materials is restricted for optical based applications. This is why additional research has been carried out on ﬂ uoride based host materials because of their low phonon frequencies, low composition degradation, and high quantum ef ﬁ ciency. In this paper, the authors review the rare earth ﬂ uoride based host nano- and micromaterials for different applications and discuss possible mechanisms. © 2018 Author(s). All article content, except where otherwise noted, is licensed under a Creative Commons Attribution (CC BY) license (http://creativecommons. org/licenses/by/4.0/). https://doi.org/10.1116/1.5044596


I. INTRODUCTION
Upconversion (UC) phosphor materials doped/codoped with rare earth elements have attracted the interest of researchers due to their multipurpose applications. [1][2][3][4][5][6][7][8] The biological application of UC materials is possible due to key material properties made possible by the fluoride based phosphor materials. 4,6 Thus, due to their high luminescence intensity, low phonon frequency, and high refractive index, fluoride based UC phosphor materials are preferred to oxide based host materials. Core-shell design as well as surface functionalization of nanophosphors was developed to engineer various medical probes for therapy purposes. Several review articles were recently published on this subject. [9][10][11] Furthermore, with the appropriate light irradiation photosensitizers, the absorbed optical energy is transferred to surrounding oxygen molecules, generating cytotoxic singlet oxygen or reactive oxygen species to kill cancer cells. Chen et al. 12 developed NaGdF 4 :Yb:Er upconversion nanoparticle (UCNPs) to be used as part of a multifunctional platform. They coated these UCNPs with bovine serum albumin (BSA) and obtained UCNP@BSA nanoparticles in the process with great solubility and stability in physiological environments. They loaded two types of dye molecules, including a photosensitizer, Rose Bengal (RB), and a near infrared (NIR) absorbing dye, IR825, into the BSA layer of the UCNP@BSA nanoparticles. In the UCNP@BSA-RB&; IR825 system, the RB absorbs green light emitted from the UCNPs under 980 nm excitation to induce photodynamic cancer cell killing, while the IR825, whose absorbance shows no overlap with the UC excitation and emission wavelengths, offers nanoparticles strong photothermal performance under 808 nm laser irradiation. The obtained dual-dye loaded nanoparticles were able to kill cancer via combined photothermal and photodynamic therapies, both of which are induced by NIR light with high tissue penetration, in a synergistic manner both in vitro and in vivo. Additionally, the paramagnetic and optical properties of Gd 3+ -doped UCNPs can further be used for in vivo dual modal imaging. 12 This type of research was reviewed by Kim et al. 13 Fluoride materials such as NaYF 4 /NaGdF 4 : Er 3+ /Yb 3+ are very efficient materials, which exhibit green radiation after infrared excitation. In the NaYF 4 host, the site of Y 3+ is substituted by the rare earth dopant ions (Er 3+ , Ho 3+ , or Yb 3+ ). The UC emission of such a fluoride phosphor material has been reported by Kano et al. 14 They determined the UC efficiency of this phosphor material with different doping concentrations of Er 3+ /Yb 3+ . The optimized emission intensity was recorded for Er 3+ /Yb 3+ codoped SrWO 4 phosphor by Pandey et al. 15 have also shown the UC emission efficiency of a variety of phosphor samples by codoping with Yb, Er, Yb, and Tm. They used a model to show the efficiency and the UC mechanism evolved in phosphor material. It was also reported that the UC efficiency was reduced by several orders of magnitude after incorporation of impurity phases. A detailed study on the efficiency of this phosphor material has been reported by Thoma et al., 16 in which they described the phase diagram of NaF-Y, Nd, Lu, or Gd. At high temperatures, they found irreversible melting of the hexagonal phases, which is responsible for the change of the hexagonal phase of the sample to a cubic phase at high temperatures. The structural characterization of the hexagonal form of NaNdF 4 has also been reported by Burns et al. 17 In addition to NaNdF 4 , other fluoride hosts such as NaGdF 4 and NaEuF 4 have also been reported by different research groups 18,19 including their method of synthesis and enhanced UC emission intensity.
The UC mechanism is a very interesting branch of optics in which large numbers of photons are emitted after excitation with lower energy photons. The mechanisms of UC phenomena have been studied by numerous research groups since its discovery in 1959. 20,21

II. UC MECHANISM OF RARE EARTH DOPED PHOSPHORS
There are many reviews and research articles related to upconverting materials and to the different mechanisms leading to UC and Auzel's review of 2004 was the first extensive review in which UC luminescence microscopy for biological applications is mentioned. 22 In this section, we briefly describe UC mechanisms and describe the fundamental differences between them and the two-photon excitation (TPE) as well as the second-harmonic generation (SHG) nonlinear processes.
There are many multiphoton processes, which generate photons whose energy exceeds that of the irradiation photons. Figure 1 shows a schematic diagram differentiating three basic processes: the SHG, simultaneous two-photon absorption (TPA), also called TPE when the exited state gives rise to radiative emission of a photon, and UC processes in which intermediate metastable levels of active ions have a determinant role. SHG is a second order nonlinear optical process which requires phase matching and symmetry restrictions. The output photon wavelength is accurately half of the input photon wavelength. SHG is commonly used for coherent generation of visible light and one of the most known nonlinear optical materials for such an application is KH 2 PO 4 . In a rare earth doped luminescent material, population of a real excited state of the active ion via the simultaneous absorption of two photons is a third order nonlinear process, which is a result of third order optical susceptibility of the material. For biological applications, TPE microscopy is an alternative to confocal microscopy with clear advantages, for example, due to better transparency of tissues in the IR compared to visible light, high-resolution imaging in intact thick animal or human organs. More details about the SHG and TPE microscopy can be read in the references. 23,24 While TPE needs simultaneous absorption of two photons without an intermediate state and has a Femto second lifetime, the UC processes involve sequential absorption of two (or more) photons using real intermediate metastable states. The efficiency of these processes is thus higher compared to that of the TPE.
The UC phenomenon was first proposed and published by Bloembergen in 1959. 25 There are several important phenomena taking part in the UC processes, which are depicted in Fig. 2. [26][27][28][29][30] These include process of (a) ground state absorption and excited state absorption (ESA), (b) energy transfer upconversion (ETU), (c) a TPA, (d) cooperative sensitization upconversion (CSU), (e) cross relaxation (CR), and (f ) photon avalanche (PA). Cooperative UC is an inefficient process, which is not included in the UC of nanoparticles at this stage. These processes are briefly discussed below.

A. Excited state absorption
Since it occurs via sequential absorptions of the pump photons, this mechanism takes place in a multilevel ion system, which is the case for the 4f n ground electronic configuration of rare earth ions that is illustrated in Fig. 2(a). It is a single ion process which can be efficient if there is an equal separation in the ladderlike structure (energy difference between E 1 and G equal to the energy difference between E 2 and E 1 ) and if E 1 is a metastable state for the reservoir capability. The ion is excited by a first pump photon from the ground state G level to the metastable state E 1 and then another pump photon promotes the ion from this intermediate state E 1 to a higher state E 2 . Therefore, UC emission at shorter wavelength than the one of the excitation source will take place from the higher lying state E 2 to the ground state G. For efficient ESA, the ladderlike system of energy states of the active ions is necessary. It is well known that trivalent rare earth ions embedded into low-energy phonon materials, which is the case for fluoride materials, have such a ladder type of energy levels with metastable states. 22 Due to the single ion characteristic of ESA, the efficiency of this process is dependent on the dopant concentration.

B. Energy transfer UC
Unlike the ESA process, which involves only one ion, a nonradiative energy transfer between two ions exists and plays a key role for ETU, as illustrated in Fig. 2(b). In the case of ETU due to successive energy transfers, the ion 1 is excited by a first pump photon absorption from the ground state G to the metastable state E 1 . Then, due to the interaction between ion 1 (called sensitizer or donor) and ion 2 (called activator or acceptor), the energy accumulated by the sensitizer is transferred to the activator, which is thus excited in its E1 state. Further excitation of ion 1 into its E 1 state  may lead to energy transfer where both ions 1 and 2 are initially in their excited state E 1 promoting the acceptor (ion 2) in the higher lying state E 2 and the sensitizer (ion 1) in the ground state. UC emission will then take place from the higher lying state E 2 to the ground state G of the activator. In the case of ETU due to a cross relaxation energy transfer, each of the two neighboring ions 1 and 2 is promoted to the E 1 excited state via absorption of pump photons, and then energy transfer promotes the acceptor (ion 2) in the higher lying state E 2 and the sensitizer (ion 1) in the ground state. There are two different kinds of nonradiative energy. In one case, the energy difference between G and E 1 is equal to the energy difference between E 1 and E 2 states: it is a resonant energy transfer. If an energy mismatch exists between these two transitions, the energy transfer would be phonon assisted. The distance between sensitizer and activator plays an important role and needs to be considered in the efficiency of the ETU process, which may easily be calculated by using the concentration of the dopant. The ETU process is the most essential part for the UC as it utilizes the ion pairs as sensitizer and activator. It is considered that Eu 3+ /Yb 3+ , Tm 3+ /Yb 3+ , Er 3+ /Yb 3+ , and Ho 3+ / Yb 3+ , which act as sensitizer and activator pairs, are the most frequently 31-34 used ones for the enhanced UC emission intensity. The sensitizer Yb 3+ ion has a large absorption cross section as compared to others. For fluoride nanoparticles, the optimized concentration of the Yb 3+ ion can be kept from 10% to 100% without evoking cross relaxations, because Yb 3+ has only two energy levels in its electronic structure. Yb 3+ ion may be excited by using pump photons in the wavelength range 975/980 nm depending on the host material.

C. Cooperative sensitization UC
This process involves a 3 ion system in which ion 1 and ion 3 [ Fig. 2(d)] are similar. Both are excited through absorption of pump photons. Then, as ion 1 and ion 3 act together in the elementary process of sensitization of ion 2 (energy transfer to ion 2), the process is qualified as cooperative: the energy accumulated by these two ions is given in one transfer to the activator (ion 2), which reaches a higher excited state. UC emission will then take place from this higher lying state to the ground state G of the activator. The CSU's efficiency is normally much lower than that of both ETU and ESA. 22 The CSU mechanisms have been published in detail and may be found in Refs. 3, 26, and 35-37.

D. Cross relaxation
Cross relaxation is an energy transfer phenomena with an ion to ion interaction as shown in Fig. 2(e). In this interaction, ion 1 transfers the excited energy to ion 2 via (Ion 2) G + (Ion 1) E 2 → (Ion 2) E 1 + (Ion 1) E 1 . In this process, ion 2 and ion 1 may be similar or dissimilar, and ion 2 can be in the excited state as well. The efficiency of the CR process is very closely connected to the dopant concentration. The CR process is one of the reasons for the concentration quenching mechanism, which is intentionally used for tuning the output color UC and also to construct a photon avalanche process. [35][36][37]

E. Photon avalanche
Photon avalanche is the most complex among all UC processes. It is illustrated in Fig. 2(f ). This process leads to UC emission when an excitation pump power above a certain threshold is used. A much less intense UC fluorescence is obtained below this threshold value, while above the pump threshold, the UC luminescence intensity increases by orders of magnitude. In this process, the pump photon is not resonant with transition from the ground state G of ions 1 or 2; it is only resonant with the transition E 1 → E 2 of the UC emitting ion 2. If one ion 2 is promoted to its metastable state E 1 , regardless of the mechanism of population of this state, and the pump photons are in resonant with the transition E 1 → E 2 , it may be further excited, via ESA, to its E 2 state. So, if the CR energy transfer schematized in Fig. 2(e) is efficient, this ion as well as one of its neighbor ions will be promoted in the metastable state E 1 . Two ions are now available for E 1 → E 2 ESA. Then, by the same feeding process, four ions will be further in state E 1 and so on. If the energy selfstorage in state E 1 is efficient enough compared to energy losses of this level, the avalanche effect will start as soon as

III. FLUORIDE BASED HOST
The UC luminescence mechanism has usually been studied in different host matrices, such as halides (iodides, bromides, and chlorides), fluorides, oxides, vanadates, phosphates, and oxysulfides. Ideally, host matrices should be nonhygroscopic for good stability in air and in water and they should have low phonon frequencies, which minimize the losses due to nonradiative relaxation pathways. Halides such as iodides, bromides, and chlorides contain low phonon frequencies (≤300 cm −1 ) but are hygroscopic and have limited use. The oxides' host matrices have relatively large phonon frequencies (larger than 500 cm −1 ) and good chemical stability. In comparison to other host matrices, fluorides have quite low phonon frequency (between 300 and 500 cm −1 ), fine chemical stability of samples, and are easily dispersed in a colloidal state with water as well as several nonpolar solvents such as hexane, ethanol, chloroform, toluene, and dimethyl sulfoxide. So, fluoride host materials, mainly alkaline rare earth tetrafluoride (AREF 4 ), are the most popular hosts for UC emission. [33][34][35][36][37][38] Ledoux et al. 38 have recently given quite a complete overview of fluoride host characteristics and advantages for efficient UC luminescence.

IV. SYNTHESIS
The synthesis of UC phosphor is the major challenge to the production of monodispersed NPs. These NPs exhibit a narrow size distribution and are highly soluble in an aqueous solution for the purpose of biological applications. The nucleation and growth process of the NPs are dependent on the crystallization method. The related nucleation growth process has been investigated by Victor Lamer; consequently, it is known as the LaMer mechanism. 39,40 The thermodynamic and kinetic description of the fundamental particle growth has been published in reference. 41 This research shows that NPs are mostly in the composition of the host lattice and agglomerated according to this parameter.
Among the existing kinds of UC host materials, fluorides have been established to be an ideal host for UC emissions with a high quantum yield due to the low phonon frequency and very high stability. To date, the NaYF 4 doped with Yb 3+ /Er 3+ , Yb 3+ /Ho 3+ , and Yb 3+ /Tm 3+ are the most efficient UC materials to produce blue, green, and red emission. 42,43 Recently, many research groups have focused on the preparation methods/techniques, surface modification, and application of UCNPs in biological fields. [41][42][43] There are many methods to synthesize NPs; however, basically three methods are very common for fluoride based UCNPs synthesis and they are described as coprecipitation, hydrothermal and thermal decomposition methods.  Others synthetic strategies such as microemulsion synthesis, microwave-assisted synthesis, and ionic liquid synthesis have been also discussed recently. 42

A. Coprecipitation method
The coprecipitation method is the easiest method as compared to other approaches because of their favorable reaction conditions, low expense for required equipment, effortless protocols, and slow reaction times. Yi at el. 44 was the first to report on the coprecipitation synthesis method for synthesizing NaYF 4 :Yb 3+ /Er 3+ phosphor NPs, using ethylenediaminetetraacetic acid (EDTA), Fig. 3. In this method, particle size is controlled by varying the relative molar amount of the EDTA.
With the preparation of rare earth doped/codoped NaYF 4 phosphors by the coprecipitation method, annealing treatment is usually required to enhance the UC luminescence intensity that in turn allows the nanoparticles to aggregate and grow larger. Additionally, the capping reagents such as EDTA can become carbonized during annealing. This reduces the hydrophilicity of NPs. Moreover, by using a silica coating on the surface of the sample, the hydrophilic nature of the material may be enhanced, and it leads to an increase in the size of the NP. Therefore, the NaYF 4 phosphor synthesized by the coprecipitation technique has a limited application in the biomedical sciences. 45

B. Hydrothermal method
The hydrothermal technique is a chemical synthesis process carried out in a sealed atmosphere at high temperature and pressure. In a typical hydrothermal synthesis process specific reaction vessels, Teflon lined autoclaves are used for the sealed atmosphere reaction. It is an efficient and suitable process for preparing UCNPs and produces diverse controllable architectures and morphologies. For illustration, a variety of shapes of NaYF 4 crystals such as rod, prism, tube, disk, and octadecahedral shapes have been obtained by using this preparation technique. [46][47][48] Shang et al. synthesized different shapes of rare earth fluoride micro/nanocrystals. 49 They reported that different molar ratios of oleic acid (OA)/ligand (Ln 3+ ) and F − /Ln 3+ have effects on the shapes of the crystals with changes from nanorods to hexagonals. The advantages of adopting the hydrothermal route for synthesizing high quality UCNPs are (a) the use of simple equipment and the overall process is also simple (b) at a relatively low reaction temperature (commonly around 200°C), (c) easy control of the structure, size, and morphology of the NPs, and (d) a high degree of purity yield. However, generally, the hydrophilicity of the UCNPs synthesized using the hydrothermal technique has not been high enough due to the presence of hydrophobic organic ligands (e.g., OA) on the surface of the nanoparticles.
Kumar et al. have synthesized different shapes of rare earth doped fluoride micro/nanocrystals (YF 3 :Tm 3+ /Yb 3+ /Ca 2+ ). 50 They have reported on the doping effect of Ca 2+ on the YF 3 : Tm 3+ /Yb 3+ phosphor material. After increasing the concentration of the Ca 2+ ions, the shape and size of the fluoride based host phosphor particles were transformed from a dumbbell to a pyramid shape (Fig. 4). In a typical preparation method, Y 2 O 3 , Tm 2 O 3 , Yb 2 O 3 , and CaCO 3 were used as precursor materials. Concentrated HCl was added to these weighed materials and heated up to 80°C. After a while, 20 ml of oleic acid along with 8 ml of ethanol was added to this solution and stirred for 15 min. Finally, to this solution NH 4 F was added as an aqueous solution and stirred for another 30 min. This solution was kept in a Teflon lined autoclave and heated at 190°C for 24 h. After cooling down, the solution was washed with ethanol three times in order to remove the organic impurities. The obtained raw material was dried over night for further characterization and measurements.

C. Thermal decomposition methods
In this method, the metal trifluoroacetate precursor is thermally decomposed to provide the desired metal fluorides. The thermal decomposition synthesis route is one of the most generally used strategies for the synthesis of UCNPs and has been implemented in the production of AREF 4 UCNPs such as LiYF 4 , 51 NaYF 4 , 52,53 NaGdF 4 , 54,55 NaLuF 4 , 56 and KGdF 4 . 57 This method has also been utilized to produce lanthanide-doped fluorescent NPs, which is based on other host matrix, such as GdF 3 58 and CaF 2 . 59 The thermal decomposition method also provides the best control of particle size and structure with a comparatively short reaction time and also produces a high quality yield with monodispersed materials. 60 The technique makes use of thermolysis of organometallic precursors, generally lanthanum trifluoroacetates/metal. Commercially obtainable Lanthanum trifluoroacetates precursors are generally prepared by reacting Lanthanum oxides with trifluoroacetic acid. 61 Normally, an elevated temperature (boiling) solvent such as 1-octadecene (ODE) solution of surface ligands is heated to slightly above the decomposition temperature of the precursors (over 300°C). The solution of the requisite metal or lanthanum trifluoroacetates is introduced into the heated ligand solution for the reaction. Moreover, the reactants are quickly released due to the fast nucleation process. 62 With the aid of surface ligands such as oleylamine (OM) or OA, the particle size of the synthesized samples is controlled. 60 These ligands include polar top groups which coordinate to the growing particle throughout the reaction, along with extensive hydrocarbon tails which extend out to the bulk solution to give the requisite solubility. 60,63 In addition, the enhancement of the particle size is controlled by using surface ligands. The repulsive interaction technique is used for the optimization of the surface ligands to stop the agglomeration of the particles. The use of air susceptible metal precursors is required resulting in the need to frequently purge the system with an inert gas, as well as due to the toxic mutable products that are produce by the reaction. 64 The sample is collected following precipitation by centrifugation and then mixed with organic polar solvents and colloidally monodispersed precursors, which allows the sample to be kept for prolonged periods of time. 61 Yuan et al. 65 have also reported core cell modified structure by varying the concentration of silica/silver to get the improved upconversion emission. Figure 5 depicts the synthesis steps of cubic NaYF 4 :Yb 3+ /Er 3+ UCNPs, and by changing the concentration of reagents, reaction time, and reaction temperature, various shapes and sizes of NaYF 4 :Yb 3+ /Er 3+ UCNPs can be obtained. 64 A modified preparation method has also been available, which takes advantage of the exceptional solubility properties of polyethylene glycol (PEG). In this modified method, polyethylenimine capping ligands were used, instead of the predictable OM or OA capping ligands, which are easily dispersed in aqueous media. In Figure 6, the shapes and sizes of synthesized UC phosphors are shown with respect to different annealing time intervals. 66 In general, the thermal decomposition technique is used for the preparation of the hexagonal phase that is lanthanum doped (such as beta-NaYF 4 ) UCNPs with the particles diameter size ranging in the order of 15-40 nm. This method requires well controlled reaction temperatures, organic solvents, and moderate air sensitive precursors along with an inert atmosphere. The reproducibility of UCNPs is a practical challenge in the temperature range of thermal decomposition. For colloidal stability, nonpolar capping ligands are required as well as surface modification for biological applications. 67

V. APPLICATIONS OF FLUORIDE BASED UC PHOSPHORS
The different RE phosphor based compounds are easily synthesized as discussed earlier and have numerous applications. These phosphors are nontoxic and biocompatible compounds. The RE doped phosphors give emissions from the UV to NIR range and therefore can be used in photonics, temperature sensing, chemotherapy, and biological imaging. The luminescence mechanisms of these materials are superior to those of conventional and endogenous fluorescent markers, which enable UC bioimaging as a matchless visualizing tool without any autofluorescence arising out of the biological tissues. The RE UC nanoparticles possess attractive structural, thermal, chemical, and optical features, such as narrow band width emission lines, long lifetimes, strong photostability, and thermal stability. Therefore, the research is concentrated on discovery and improving materials and their properties for especially UC bioimaging as biological luminescence marker applications.  Among the different phosphors, the fluoride based RE doped phosphors are very attractive and have versatile utility in fields such as photonics, solar cells, temperature sensing, cancer therapy, chemotherapy, and biological imaging. The photonics and solar cell applications arise due to downconversion (DC) and quantum cutting (QC) processes. [88][89][90][91][92][93] For example, Guo et al. have synthesized Ho 3+ , Yb 3+ codoped BaGdF 5 nanoparticles by the hydrothermal method and reported downshifting and quantum cutting emissions due to first and second order energy transfers. The emission of NIR photons observed at 980 and 1180 nm is due to cooperative downconversion and cross relaxation energy transfer processes. The theoretical quantum efficiency is estimated to be 192% for this material. 88 When the Ho 3+ , Yb 3+ codoped BaGdF 5 nanoparticles are excited with 450 nm, the Ho 3+ ions are promoted to its 5 F 1 state, which relax to the 5 F 3 state nonradiatively. Due to cooperative energy transfer (CET) and CR processes, it gives Stoke emissions in the visible and NIR regions centered at 980 and 1180 nm due to the Ho 3+ and Yb 3+ ions. The QC process due to Ho 3+ ions in the NIR region is verified by time resolved spectroscopy. The NIR emissions at 980 and 1180 nm are due to the 2 F 5/2 → 2 F 7/2 manifolds of Yb 3+ ions and the 5 I 6 → 5 I 8 transition of the Ho 3+ ions, respectively. Thus, it has also been proposed that these nanoparticles open the door for designing ultraefficient photonic devices in low band gap solar cells and thermophotovoltaic energy conversion, etc.
Zhang et al. have synthesized Y 6(1−x) O 5 F 8 :xYb 3+ , 1% Ho 3+ microarchitectures by the hydrothermal method and reported QC emissions along with UC emission due to CETs. The emission of NIR photons observed at 980 and 1180 nm is due to cooperative downconversion and cross relaxation energy transfer processes. 89 The schematic presentation of the energy levels of Ho 3+ and Yb 3+ ions is shown in 450 nm radiation, the Ho 3+ ions are promoted to the 3 H 6 and 5 F 1 states, which relax to the 5 G 4 and 5 F 4 states nonradiatively. Due to CET and CR processes, Stoke emissions are given in the visible and NIR regions centered at 980 and 1180 nm due to the Ho 3+ and Yb 3+ ions. The NIR emissions at 980 and 1180 nm are a result of the 2 F 5/2 → 2 F 7/2 transition (Yb 3+ ions) and the 5 I 6 → 5 I 8 transition (Ho 3+ ions), respectively. They also emit visible photons upon 360 and 450 nm excitations.
The photoluminescence excitation (PLE) and PL spectra of different concentrations of Yb 3+ ions (0, 5, 10, and 15 mol. %) in Y 6 O 5 F 8 :Yb 3+ /Ho 3+ phosphor are shown in Fig. 8. The PLE spectrum, shown as an inset in Fig. 8(b), was measured by determining the emission at 980 nm, whereas the PL spectra were monitored upon excitation with 360 nm and 450 nm sources. The inset image in Fig. 8(a) shows that quantum cutting can be obtained by exciting the sample with 360 nm, and the same was obtained for an excitation of 450 nm. 89 The emission spectra show that the DC intensity of the visible regime decreases continuously on increasing the concentrations of the Yb 3+ ions, whereas that of NIR emission is an optimum at 10 mol. % [see Figs. 8(a) and 8(b)]. This occurs due to CET from Ho 3+ to Yb 3+ ions. This NIR emission may be used to increase the efficiency of the crystalline solar cells.
There is enormous interest in the spectral conversion for quantum cutting processes using three photons. In an interesting work, Yu et al. observed near infrared quantum cutting through a three step process (three photons) in Tm 3+ doped β-NaYF 4 nanophosphor. 93 The sample was excited using 470 nm, which promotes the ions from the ground state ( 3 H 6 ) to the excited state ( 1 G 4 ). When the sample is excited with a 470 nm source, the emission of three NIR photons occurs at 1165, 1466, and 1800 nm via three different emission processes. When the sample is further excited by a 798 nm source, the Tm 3+ ions are promoted to the 3 H 6 state and give rise to the emission of two NIR photons at 1466 and 1800 nm along with another NIR photon at 808 nm. This excitation process increases the population of the ions in the 3 H 6 state and enhances the intensity of these NIR photons. The three step emission processes are summarized to be (a) the first-step, 1 G 4 → 3 H 4 at 1165 nm, (b) the secondstep, 3 H 4 → 3 F 4 at 1466 nm, and (c) the third-step, 3 F 4 → 3 H 6 at 1800 nm. The photoluminescence emission spectra of the Tm 3+ ion have been observed on excitation with 470, 798, and 976 nm wavelengths.
Interestingly, in another set of experiments, the Tm 3+ /Yb 3+ codoped phosphors also emits the NIR emission (Stoke emission) at 1800 nm from Tm 3+ ion on excitation with a 976 nm source. This is due to nonradiative energy transfer from Yb 3+ to Tm 3+ ions. The Tm 3+ /Yb 3+ codoped sample was excited with a 976 nm diode laser, which promotes the Yb 3+ ions to its excited state ( 2 F 5/2 ) state from the ground state ( 2 F 7/2 ). These ions transfer their energy to the Tm 3+ ion in the 3 H 4 state, which ultimately populates the 3 H 4 state. As a result, the emission of NIR photon is possible at 1800 nm in the Tm 3+ /Yb 3+ codoped sample, which matches the thirdstep emission process via the 3 F 4 → 3 H 6 transition. The generation of these three photons, that is, 1165, 1466, and 1800 nm, through sequential depopulation of the ions from 1 G 4 state leads to their applications not only in the field of low band gap solar cells but also thermovoltaic spectral conversion. 88,93 Zhu et al. have prepared monodispersed Pr 3+ /Yb 3+ codoped NaGdF 4 @NaYF 4 active-core/inert-shell nanoparticles and studied the NIR quantum cutting luminescence in this material. In this case, the quantum cutting arises due to CET from Pr 3+ to Yb 3+ ions. They have introduced an extra Yb 3+ ion into the shell in order to prevent the adverse effect raised by concentration quenching. It forms an active-core/ active-shell environment and improves the intensity of NIR emission. The schematic arrangement of the excitation and emission processes in the two cases is shown in Fig. 9. 94 The Pr 3+ /Yb 3+ codoped NaGdF 4 @NaYF 4 active-core/ inert-shell nanophosphors were excited by a 445 nm source and the spectra thus obtained are given in Fig. 10. The Pr 3+ ion gave an intense blue emission at 483 nm due to the 3 P 0 → 3 H 4 transition. When codoped with Yb 3+ ions, the emission intensity at 483 nm decreased notably, whereas an intense NIR emission centered at 979 nm occurred due to the 2 F 5/2 → 2 F 7/2 transition (Yb 3+ ion). Since the Yb 3+ ions cannot be excited by 445 nm, therefore, the NIR emission occurred due to the CET process. Figure 10 also shows an increase in the intensity of the NIR band due to the addition of the Yb 3+ ion into the shell. Wang et al. have also reported quantum cutting in Eu 3+ doped NaYbF 4 nanotubes through cooperative DC energy transfer, and cross relaxation processes. The sample was excited by using a 393 nm source, which promotes the Eu 3+ ions from the 7 F 0 to 5 L 6 level. The ions in this state relax nonradiatively to populate different low lying states. As a result, a large number of radiative transitions took place at 465, 488, 510, 555, 591, 615, 650, and 695 nm wavelengths. Some of the Eu 3+ ions present in the 5 D 2 state transfer their energy to 2 F 7/2 state of the two Yb 3+ ions cooperatively. This leads to a quantum cutting process by the broad emission of two NIR photons in the 977-1015 nm region. The cross relaxations between Eu 3+ and Yb 3+ ions also promote the ions in the upper states of the two ions, which improves the emission intensity of the NIR transition. 95 The schematic representation of CET and cross relaxation of the ions is shown in Fig. 11.
The emissions of visible and NIR photons as a function of different Eu 3+ concentrations are shown in Fig. 12. It is clear from the figure that the emission intensity increased with increasing concentration up to a certain level and then decreased considerably due to concentration quenching. The emission intensity was an optimum at 3 mol. % concentration of the Eu 3+ ion for the visible and NIR photons, respectively. The reduction in the emission intensity was found to be due to shorter distances than the critical distance between the Eu 3+ /Yb 3+ ions, respectively. Thus, the Eu 3+ doped NaYbF 4 nanotubes gave efficient quantum cutting emissions and can be used in photonic devices and solar cells. 95 On the other hand, the fluoride based RE doped nanophosphors show interesting frequency UC in which lowenergy NIR photons are converted into high energy UV/ visible photons. The RE doped fluoride nanophosphors give rise to anti-Stokes emissions in the visible and NIR regions, which are used in optical devices, temperature sensing, chemotherapy, photothermal treatment, biomarkers, biological imaging, etc. The NIR light can easily penetrate through the biological tissues and can be tuned to destroy infectious diseases in the body. Since the RE doped fluoride nanophosphors are excited by NIR (976 nm) light, the damage of tissue through the use of high energy visible even near UV photon can be avoided. The RE doped fluoride nanophosphors are nontoxic and more stable and thus a state of art with potential applications in biological and clinical laboratories. [68][69][70][71][72][73][74][75][76][77][78][79][80][81][82][83][84][85][86][87] The use of RE doped materials in different fields can be understood with the following schematic diagram (Fig. 13). 96 The RE doped fluoride nanophosphors show color tunability [97][98][99] and temperature sensing ability. 99 in these levels. This change can be monitored by observation of the emission spectra of the RE ions at different temperatures starting at room temperature. 100,101 It can be used in nanothermometry to show small variations in the temperature of human bodies at an intracellular scale. 100 For example, Dubey et al. 100 have studied the optical nanothermometry in the NaYF 4 :Er 3+ /Yb 3+ /Li + phosphors and found an enhancement in the temperature sensing in the presence of Li + ions with a maximum sensor sensitivity of ∼0.0059 K −1 . Figure 14 shows the sensor sensitivity (upper inset) and relative sensitivity (lower inset) of the Er 3+ /Yb 3+ /Li + triply ions doped NaYF 4 phosphor for the 2 102 Basically, the cell-cytotoxicity and cell-permeability of luminescent materials are very important parameters to check for their compatibility for use of such materials as bioprobes. The cytotoxicity analysis shows that the mPEG-UCNPs are less toxic with a cell viability up to 80% for an increase in the concentration from 125 to 250 μg ml −1 . The fluorescence and bright-field laser scanning confocal microscope images show the fluorescence through the intracellular region with negligible background fluorescence (Fig. 15).
For biological tissues of the human body, the absorption window of incident photons is from around 600 to 1400 nm. The absorption coefficients of water, melanin, deoxy-, and hemoglobin are very small in this regime (Fig. 16). The photons in this wavelength window penetrate more deeply into tissue than that of photons in the 400-750 nm wavelength range. 104 The IR photons are absorbed only in a small amount by the tissues and the rest of the incident photons are scattered. Therefore, a laser beam penetrates the tissue up to a few millimeters. The crucial applications of the tissue imaging are thus based on the specific optical window such as 860 nm, 1060 nm, and 1310 nm (based on the excitation wavelength of the material). But for the UC nanophosphor, the excitation wavelength is found to be 980 nm, which is a suitable wavelength for imaging purposes. This imaging technique is better than the whole body imaging techniques such as magnetic resonance imaging, X-rays, and ultrasound, because of the simple visual inspection at a fine spot, below <100 μm penetration depth. Moreover, this technique is also better due to its high resolution.
The RE doped UC fluoride nanophosphors are also used in fluorescence resonance energy transfer sensing of drug release and imaging based chemotherapy, 103 NIR imaging and photothermal treatment, 105 thermal scanning at the cellular level, 106 in vivo imaging, computed tomography (CT) and magnetic resonance (MR) imaging, 107,108 etc. For example, the trimodal properties of core-shell BaYbF 5 : Tm@BaGdF 5 :Yb,Tm nanophosphors have been investigated by Zhang et al. and applied these nanoparticles for in vivo trimodal upconversion luminescence/CT/MR imaging. 108 The core-shell BaYbF 5 :Tm@BaGdF 5 :Yb,Tm nanophosphors are nontoxic and appear clearly in in vivo UC fluorescence imaging of a mouse with injection of the nanocrystals (Fig. 17). Moreover, the nanocrystals have the potential application in CT scanning. In the case of in vivo CT imaging, the images of Kunming mice were taken using iCT (Brilliance iCT, Philips Healthcare, Cleaveland, OH) many times before and after the injection of nanoparticles and iobitridol, respectively. A CT scan was taken before injection and after injection for 3 min, 30 min, 1 h, and 2 h. It has been observed after 3 min uptake of nanocrystals, the image quality of liver [ Fig. 18(a)] and spleen [ Fig. 18(b)] was enhanced up to 2 h. After 2 h, there was no enhancement observed. This phenomenon can easily be distinguished by using a volume rendering method [ Fig. 18(c)].
The Gd 3+ ions show magnetic properties and are present in BaYbF 5 :Tm@BaGdF 5 :Yb/ Tm-PEG nanoparticles, which enable the enhancement of the T 1  is limited due to their insensitivity to the full solar spectrum. The spectral distribution of sunlight consists of photons with wavelengths that range from ultraviolet to infrared (280-2500 nm), Fig. 20. Current PV cells, however, only utilize a relatively small fraction of the solar photons. 110 The solar radiation spectrum shows the fraction (highlighted in green in the online version) absorbed by a typical crystalline silicon-based PV cell and the spectral regions that can be utilized through quantum cutting and UC processes (highlighted in purple and red, respectively, in the online version). The function of NaYF 4 nanophosphors to eliminate the spectral variance among the incident solar spectrum, semiconductor, and the sub-bandgap character of the semiconductor materials is argued explicitly. Some very good examples of Er 3+ -doped/codoped NaYF 4 UC phosphors to enhance the responsivity in the NIR spectrum for solar cell applications are given in Table I. 109 The UC is especially relevant for amorphous Si solar cells.
The excitation wavelength of NaYF 4 phosphor with the right dopants also makes them suitable for phosphors in UV-pumped white LED devices. 118 The roles of the synthesized method and preparation conditions have severe effects on the efficiency thereof. 119 The UC properties can also be enhanced with codoping with transition metals. 120 Adusumalli et al. 121 used NaYF 4 :Ce 3+ /Tm 3+ /Mn 2+ nanocrystals as an efficient single component phosphor in order to produce intense white light. The Ce 3+ ions were used as the excitation source, and the single blue band emission of Tm 3+ ions was combined with the broadband green-yellow emission of Mn 2+ ions to produce strong white-light emission.

VI. SUMMARY AND CONCLUSIONS
This review summarizes the UC emission research in rare earth doped fluoride based phosphor materials. The SHG, TPE, and UC processes have been discussed and compared. The ESA as well as the basic energy transfer mechanisms involved in UC processes occurring in rare earth doped phosphors has also been discussed. The synthesis section includes synthesis methods, such as the coprecipitation method, hydrothermal method, and thermal decomposition method, which normally yield nanoscaled particles. The fluoride phosphors synthesized through these routes generally produce nanoparticles in the range of 10-50 nm. The nanoparticles in this range enable potential applications in fields such as solar cell, photonic devices, nanothermometry, biomarkers, and biological imaging. The NIR to visible UC process leads to applications of the rare earth doped fluoride phosphor materials in chemotherapy, photothermal treatment, theranostics, etc. It is found that these materials are also useful in medicinal and clinical laboratories for CT scan and MR imaging. Thus, the rare earth doped fluoride phosphors are very unique and promising materials for many functions and multidirectional applications. The scope of the use of these materials is endless.