1. Introduction

Light-emitting diodes (LEDs) have found their way into many applications in daily life, from warm-white LEDs for energy-efficient lighting to display backlights for smartphones, tablets and televisions. The importance of white LEDs (WLEDs) was recognized with awarding the 2014 Nobel Prize in Physics for the invention of efficient blue LEDs, which have enabled bright and energy-saving white light sources [1]. To obtain white light, part of the blue light from the LED chip is converted by color-converting luminescent materials. The inorganic materials used for this are impurity-doped wide band gap materials, so-called phosphors, or colloidal quantum dots (QDs). QDs are semiconductor nanocrystals that show a tunable, spectrally-narrow emission and are suitable for solution processing [2]. Phosphor materials consist of an inorganic host material modified with dopants, either rare-earth ions such as Eu²⁺ and Ce³⁺ or transition-metal ions such as Mn²⁺ and Mn⁴⁺. Rare-earth ions are typically used in the classic LED phosphors, with Y₃Al₅O₁₂:Ce³⁺ (YAG:Ce) as its workhorse [3]. Eu²⁺ is a dopant allowing to tune the emission spectrum by changing the composition of the host compound, for instance from green emission in SrGa₂S₄:Eu²⁺ [4,5] and SrSi₂O₂N₂:Eu²⁺ [6] to red emission in Sr₂Si₅N₈:Eu²⁺ [6,7], (Ca,Sr)S:Eu²⁺ [4,8], CaAlSiN₃:Eu²⁺ [9] and Sr[LiAl₃N₄]:Eu²⁺ [10]. Cost, environmental and supply issues for the rare-earth elements have recently triggered a tendency to develop LED phosphors which are free of rare-earth elements. Transition-metal dopants such as Mn have come into interest in materials like CaZnOS:Mn²⁺ and K₂SiF₆:Mn⁴⁺ as red LED phosphors [11–14].

The necessary conditions for a phosphor to be suitable for LED applications were determined and discussed at length by Smet et al. [3]. First of all, the emission spectrum has to be suitable for the targeted applications. For lighting, this implies a high color rendering for the resulting white light source, which can be represented on the color quality scale (CQS) [15] or using the color rendering index (CRI) [16]. For display applications, saturated colors are needed to produce a large color gamut or high color purity. The Recommendation ITU-R BT-2020 (Rec. 2020) for ultra-high definition television by the International Telecommunication Union, proposes the use of monochromatic light at 630 nm, 532 nm and 467 nm for the red, green and blue (RGB) primary colors [17], a requirement which clearly cannot be met using either phosphors or QDs. However, narrow-band color filters can be used to obtain the nearly monochromatic light. The better the emission spectrum of the phosphors matches to the color filters, the lower the filtering losses are. Secondly, for a phosphor to be suitable for LED applications, it needs to have a high internal quantum efficiency (IQE), which remains stable at elevated temperature and provides a high luminous efficacy of the radiation (LER, a measure for the average eye sensitivity for the spectrum, specified in lm/W). Thirdly, a high absorption of blue light is needed, yielding a high external quantum efficiency (EQE). The EQE is defined as the ratio of the number of emitted photons to the number of incident photons, while the IQE considers only the absorbed photons [18]. Lastly, a sufficiently fast decay is needed to avoid unwanted saturation effects when using higher excitation fluxes. Only when these four conditions are simultaneously fulfilled, a phosphor can be considered as a valid candidate for LED applications [3]. K₂SiF₆:Mn⁴⁺ was not yet included by Smet et al., but this review will discuss whether it meets all conditions for LED applications. To compare the performance of different red-emitting phosphors, some popular phosphors were selected here. As a benchmark, the commercially-available europium-based nitrides Sr₂Si₅N₈:Eu²⁺ [19–21] and CaAlSiN₃:Eu²⁺ [22] were used. In the latter two cases, the broadband emission originates from parity-allowed, interconfigurational 4f⁶5d^1$\to$4f⁷ electric-dipole transitions. As rare-earth-free alternatives, we selected the manganese-based phosphors K₂SiF₆:Mn⁴⁺ [14,23] and CaZnOS:Mn²⁺ [11,12,24]. Both K₂SiF₆:Mn⁴⁺ and CaZnOS:Mn²⁺ have d-d transitions, which are partially spin-allowed for Mn⁴⁺ and spin-forbidden for Mn²⁺. Upon incorporation in the host compound, the spin and parity selection rules can be somewhat relaxed, like in the case of the tetrahedrally coordinated CaZnOS:Mn²⁺ [11]. When Mn⁴⁺ is compared to Mn²⁺, decay times are similar (in the ms range, typical for d-d transitions), but Mn⁴⁺ is the preferred valence state for LED applications, due to better excitation properties and narrower emission bands. In K₂SiF₆:Mn⁴⁺, the intraconfigurational 3d³ transitions of the Mn⁴⁺ ion are responsible for the luminescence, yielding narrow line emission, rather than the broadband emission of the 3d⁵ configuration of Mn²⁺. For further comparison, a phosphor based on the Eu³⁺ ion, showing intraconfigurational 4f⁶ transitions was added. Eu³⁺-based phosphors like CaYAl₃O₇:Eu³⁺ are often suggested for LED applications, despite their difficult excitation with near-UV to blue light [25]. The IQE of the materials was not mentioned in this comparison as this depends largely on the synthesis conditions and it can be safely assumed that there are no fundamental restrictions for any of these materials at room temperature, as deduced from the high temperatures at which thermal quenching (TQ) sets in. Although the thermal quenching temperature T_1/2 is more often used to quantify thermal quenching, the Eu³⁺ doped red phosphors show a slow thermal quenching. Since no measurements are available up to half of the initial intensity, here the temperature at which 75% of the initial intensity is reached is compared.

Figure 1 shows the performance of the selected phosphors on the enumerated points. K₂SiF₆:Mn⁴⁺ reaches relatively high values on the color quality scale (CQS), color purity and absorption strength. The absorption strength of blue light (A) is too low for CaZnOS:Mn²⁺ and for the Eu³⁺-based phosphor, compared to the Eu²⁺-based benchmark materials, while K₂SiF₆:Mn⁴⁺ reaches a reasonably high absorption, however still below the level of the Eu²⁺-doped materials benefitting from the parity allowed 4f-5d transitions. Thermal quenching in K₂SiF₆:Mn⁴⁺ sets in earlier compared to the Eu-doped materials, but a high emission intensity is maintained at common operating temperatures of LEDs. CaAlSiN₃:Eu²⁺ and K₂SiF₆:Mn⁴⁺ have the highest color purity of the selected red phosphors due to a peak wavelength in the 630–650 nm range. A lower color purity is reached with peak emission in the 610–620 nm range. The emission spectrum of CaZnOS:Mn²⁺ is less suitable for display applications as too much color filtering is necessary to achieve the Rec. 2020 standard for the red primary color [17]. Mn⁴⁺-based phosphors are more suited for this purpose, provided that their absorption strength can still be increased. The phosphors based on Mn²⁺, Mn⁴⁺ and Eu³⁺ suffer from saturation effects at moderate photon fluxes, given their slow decay due to the forbidden intraconfigurational transitions. Applying a remote-phosphor LED design can circumvent this problem to a certain extent by removing the phosphor from direct contact with the LED chip. The larger area of the remote phosphor plate relaxes the saturation effects at moderate photon fluxes [2,14,26]. The color point of K₂SiF₆:Mn⁴⁺, close to that of monochromatic light of 630 nm, is ideal for using K₂SiF₆:Mn⁴⁺ in combination with YAG:Ce to create warm-white LEDs with good color rendering, while keeping the LER at an acceptable level. Because of these promising LED applications, this review gives an overview of the reports on K₂SiF₆:Mn⁴⁺ phosphors, focusing on synthesis, optical and structural properties.

 

Fig. 1 Performance of K₂SiF₆:Mn⁴⁺ as LED phosphor for lighting or displays, compared with the benchmark phosphors Sr₂Si₅N₈:Eu²⁺ and CaAlSiN₃:Eu²⁺, the broader emitting CaZnOS:Mn²⁺ and the line-emitter CaYAl₃O₇:Eu³⁺. The indication “A” shows the absorption of blue light (465 nm) while “TQ” gives the temperature at which the emission intensity is at 75% of the initial intensity. γ-flux expresses the inverse decay constant τ⁻¹. The lower this value, the more saturation effects are expected when elevating the excitation intensity. “LER” and “CQS” show the luminous efficacy of the radiation of the phosphor's emission and the color quality scale of a white LED (CCT = 4000 K) based on this phosphor (combined with a blue pump LED at 465 nm and the yellow phosphor Y₃Al₅O₁₂:Ce³⁺). Finally, the indication “color purity” gives the transmission of the phosphor's emitted light tuned to the Rec. 2020 recommendation.

Download Full Size | PDF

Lately, several reviews covered the recent developments concerning red LED phosphors. Li et al. [27] covered both rare-earth and Mn²⁺- and Mn⁴⁺-activated phosphors and concluded that the synthesis difficulties of Eu²⁺-doped nitride phosphors and the safety hazards of Mn⁴⁺-doped fluoride phosphors need to be overcome, but moving away from rare-earth dopants can have beneficial environmental and economic effects. Chen et al. [28] reviewed Mn⁴⁺-activated solids for WLED applications, including both fluoride and oxide phosphors, and concluded that a blueshift and increased absorption in the blue region are necessary for oxide phosphors and that thermal stability and humidity resistance as well as quantum efficiency (QE) are still points of concern for fluoride phosphors. A more extensive review on Mn⁴⁺-doped luminescent materials by Zhou et al. [29] concluded that Mn⁴⁺-doped alkaline aluminates suffer from weak UV–blue absorption and efficiency. K₂SiF₆ and K₂TiF₆ doped with Mn⁴⁺ were reported to need improvement in terms of chemical and thermal stability and in environmental safety of the synthesis process. Recently, Liu and Nguyen [30] reviewed different synthesis processes of Mn⁴⁺ hexafluoride phosphors, as well as the optoelectronic properties and applications in LEDs. They show that the fluoride phosphors are promising red LED phosphors, but improvement needs to be made on QE and scalable synthesis methods that are environmentally friendly. Coatings can improve the chemical and thermal stability of the fluoride phosphors.

The current review focuses purely on K₂SiF₆:Mn⁴⁺ and gives an extensive overview of all performed synthesis routes and properties of the resulting phosphors. K₂SiF₆:Mn⁴⁺ is selected because it was the first red Mn⁴⁺-doped fluoride phosphor, first reported in 1973 [31]. K₂SiF₆ is one of the most promising hexafluoride hosts, yielding a slightly higher LER upon doping with Mn⁴⁺ than K₂TiF₆ [32] and K₂SiF₆:Mn⁴⁺ has a 30% higher EQE than KNaSiF₆:Mn⁴⁺ [30]. In this review, special attention is given to the electronic structure of K₂SiF₆:Mn⁴⁺ and its spectroscopic parameters, lining up the diverse values that have been published over the years. The evaluation of several analysis methods, including lesser-used ones like electron paramagnetic resonance spectroscopy (EPR), Raman and extended X-ray absorption fine structure spectroscopy (EXAFS), results in a thorough overview of the properties of K₂SiF₆:Mn⁴⁺, that has not been given in previous reviews.

2. Electronic structure

The specific red line emission seen in K₂SiF₆:Mn⁴⁺ around 630 nm is caused by 3d³–3d³ transitions in Mn⁴⁺. The 3d³ electron configuration of Mn⁴⁺ results in 120 possible distributions of the three electrons over the ten available single-particle states of the 3d shell. The electronic Coulomb repulsion gives rise to a splitting of the 120-fold degenerate configuration in eight LS terms that can be found for the free Mn⁴⁺ ion, the magnitude of this splitting is characterized by the Racah parameters B and C [33]. These terms are further split by the crystal field (CF), characterized by so-called CF parameters [33]. The symmetry of the CF dictates the number of required parameters and the remaining degeneracies of the electronic energy levels. For an octahedral crystal field, only one CF parameter, 10Dq, is required and the splitting is described by the Tanabe–Sugano diagram (Fig. 2(a)). Due to its high effective positive charge, Mn⁴⁺ experiences a strong CF, as specified by the value of 10Dq/B. In this case, the emitting level has ²E_g symmetry, resulting in a sharp emission line of the spin-forbidden transition towards the ⁴A_2g ground state. Two broad absorption bands are present from the spin-allowed ⁴A_2g$\to$⁴T_2g (⁴F) and ⁴A_2g$\to$⁴T_1g (⁴F) transitions [34]. The third spin-allowed absorption band, from the ⁴A_2g$\to$⁴T_1g (⁴P) transition is located at higher energies and can overlap with host-lattice absorption or ligand–Mn⁴⁺ charge transfer [35]. The width of the excitation bands is explained by the strength of the electron–vibrational interaction between the electronic states of Mn⁴⁺ ions and crystal-lattice vibrations [35,36]. The emitting transition, ²E_g$\to$⁴A_2g, corresponds to a mere spin-flip of a single electron, showing a small electron–phonon coupling and hence an emission spectrum featuring a zero-phonon line (ZPL) along with a few Stokes and anti-Stokes lines.

 

Fig. 2 Illustration of how the emission and excitation spectrum are governed by the electronic structure of the octahedral [MnF₆]^2- center. (a) Tanabe–Sugano diagram for a 3d³ system in the octahedral symmetry, showing how the free-ion terms (shown in the extreme left part of the figure) split in the crystal field. Green lines indicate spin quartets, red lines indicate spin doublets. This diagram indicates the energies of the spectroscopic transitions. The black vertical line corresponds to the case of Mn⁴⁺ in K₂SiF₆. (b) Simplified configurational coordinate diagram showing the potential energy surfaces that play a role in the radiative transitions. This diagram qualitatively indicates the spectral shape for the different transitions. (c) Room temperature excitation (green line) and emission (red line) spectrum of K₂SiF₆:Mn⁴⁺. The approximate location of the zero-phonon line (ZPL) is indicated (adapted with permission from [14], copyright 2016, The Electrochemical Society).

Download Full Size | PDF

Different hosts can be compared according to their spectroscopic parameters. These can be obtained by a tedious fitting procedure involving different terms in an effective Hamiltonian. Next to the common Racah parameters B and C; the CF parameters; additional interactions are sometimes accounted for. These can include low symmetry components of the crystal field or the 3d spin-orbit coupling. The Trees parameters α and β are known to improve the correspondence between experimental and calculated spectra as they correct in an effective way for the single configurational description of the formalism. In the case of Mn⁴⁺ care should be taken to avoid overparameterization as only a few experimental energies are available. As can be seen in Fig. 2(b), it is designated to use the ²E_g$\to$⁴A_2g zero-phonon line for the CF fit, while band maxima are better used for the ⁴A_2g$\to$⁴T_2g (⁴F) and ⁴A_2g$\to$⁴T_1g (⁴F) transitions. The reason for this is the different offsets between the minima of the potential energy surfaces. A CF Hamiltonian does not account for electron-vibrational coupling and represents only one geometry of the luminescent center, corresponding to a fixed Q-value in Fig. 2(b).

Within the effective Hamiltonian formalism, numerically fitting is certainly the most practical approach, providing freedom in selecting which interactions are taken into account. If the Hamiltonian is however restricted to the octahedral crystal field and when only the Coulomb interaction (characterized by B and C) is taken into account, a simpler approach is possible where analytical formulas relate the parameters B, C and 10Dq to transition energies. The ratio 10Dq/B, as used in the Tanabe–Sugano diagram, is often directly obtained from the location of the ⁴T_2g (⁴F) level. The value of B can be obtained in addition from the location of a second energy level. The full Hamiltonian is block diagonal according to the irreducible representations (irreps) of the octahedral point group O_h. Diagonalization of the 2x2 submatrix, corresponding to ⁴T_1g, directly yields the energies of the ⁴T_1g (⁴F) (lowest eigenvalue) and ⁴T_1g (⁴P) levels. Subtracting the energy of ⁴T_2g (⁴F), which constitutes a 1x1 block, yields the following formula, derived in [37]:

Crystal field strengths 10Dq and Racah parameters B, C were empirically determined for Mn⁴⁺ in many host compounds. The value of 10Dq for Mn⁴⁺ is known to vary from 18180 cm⁻¹ in SrTiO₃ [38], to 24000 cm⁻¹ in 3.5MgO·0.5MgF₂·GeO_2, a fluorogermanate host, in which the octahedral symmetry is slightly distorted due to the Jahn–Teller effect [39]. The location of the emitting ²E_g state, and hence the color of the luminescence is largely independent of 10Dq, as evidenced by the flat line in the Tanabe–Sugano diagram (Fig. 2(a)) [35,38,39]. Still, the position of the ZPL of the emitting ²E_g$\to$⁴A_2g transition can vary from 16207 cm⁻¹ in Na₂SiF₆ to 13827 cm⁻¹ in SrTiO₃ [35,38,40]. This variation is primarily determined by the values of the Racah parameters B and C. When Mn⁴⁺ is incorporated in a crystal or molecule, the Racah parameters are reduced with respect to the free ion values B₀ = 1160 cm⁻¹ and C₀ = 4303 cm⁻¹ [41]. This reduction is commonly referred to as the nephelauxetic effect, a designation first introduced by Schäffer and Jørgensen who introduced a nephelauxetic parameter,

Based on a thorough literature survey of Mn⁴⁺ spectra, Brik and Srivastava altered Jørgensen’s nephelauxetic parameter:

 

Fig. 3 Dependence of energy of the Mn^{4+ 2}E_g level on the new nephelauxetic ratio β₁ for fluoride hosts and oxide hosts and reported values for K₂SiF₆:Mn⁴⁺, adapted from [47]. The shaded red area corresponds to a peak wavelength < 650 nm, the given linear fit is −142.83 + 15544.02β₁, with ± σ = 365 cm⁻¹. Inset: The reported values of the Racah parameters B and C for K₂SiF₆:Mn⁴⁺, which were used to calculate the nephelauxetic parameter β₁ show a linear correlation. The given linear fit is 5179.5-2.256B.

Download Full Size | PDF

Table 1. Spectroscopic parameters for K₂SiF₆:Mn⁴⁺.

View Table | View all tables in this article

Table 2. Values of zero-phonon lines for optical relevant d-d transitions in K₂SiF₆:Mn⁴⁺ from experiments and empirical calculations.

View Table | View all tables in this article

The location of the energy levels of K₂SiF₆:Mn⁴⁺ has been determined using experimental spectra, Franck–Condon fits and CF calculations, but the obtained values vary. In general, values from CF calculations are at higher energies than the experimental values. The B value of K₂SiF₆:Mn⁴⁺ varies from 351 to 770 cm⁻¹, where the lower values originate from experimental ZPL energies and the higher ones come from CF calculated energies. The value of 770 cm⁻¹ is most often referred to [14,34,36,47] and corresponds only with the CF calculated energy levels by Brik and Srivastava [35], according to Eqs. (1,2). Other publications use a value for the ZPL of the ⁴A_2g$\to$⁴T_1g (⁴F) transition that is up to 7500 cm⁻¹ lower, causing the variation of B. Such a wide distribution of the location of the ZPL is remarkable, and has consequences for other spectroscopic parameters. The 10Dq/B ratio, which shows the position of the transitions in the Tanabe–Sugano diagram, is between 26.4 and 58.1, due to the large variation in B. The C value varies from 3433 to 4405 cm⁻¹, with lower values from CF calculated energies and higher values from experimental ZPL energies. When the ²E_g$\to$⁴A_2g ZPL is observed in experimental spectra, often at lower temperatures (10–90 K), it is located at 1.99–2.003 eV. Although various data are cross-referenced, the calculated parameters often differ, which can be caused by the choice for ZPL or centered energy values for the excitation bands. When Eqs. (1,2) are applied to the CF-calculated energy levels by Brik and Srivastava [35], the obtained 10Dq/B ratio is 30.2 with C = 3440 cm⁻¹. Despite the different values for energy levels that are used in calculations, a clear linear correlation is found if all the empirical B and C values that are found for K₂SiF₆:Mn⁴⁺ are compared, as can be seen in the inset of Fig. 3. Nevertheless, for the different B and C values, a rather constant β₁ of 1.039 ± 0.018 is found, favoring the use of a single β₁ parameter.

The previous calculations were empirical, based on experimental photoluminescence spectra, typically measured at low temperature. On the other hand, it is possible to obtain information on structural, electronic and optical properties from so-called first principles methods. Novita et al. calculated transition energies starting from DV-Xα molecular orbitals that were combined in many-body states [64]. Correlation effects were introduced by a method described in [65]. The obtained results are promising, describing the influence of the host compound in a qualitative correct way and yielding ZPL energies that are 308, 170 and 450 meV different from the experimental values at 10 K [14].

Density functional theory (DFT) was applied to the study of K₂SiF₆ and K₂SiF₆:Mn⁴⁺ at LDA level [48,51], GGA level [48,51,55] or with a hybrid functional [55]. Ground state geometries were found that are in line with experimental results. Calculated energies of the direct band gap of K₂SiF₆ amount to 8.092 eV (LDA) [48] and 7.588 eV (GGA) [48]. The experimental determination of band gaps above 5 eV is not straightforward, due to absorption of the setup. The only published experimental band gap is 5.6 eV [34], which seems underestimated since DFT calculations usually underestimate band gap values. The Kohn-Sham density of states (DOS) for K₂SiF₆:Mn⁴⁺ is shown in Fig. 4. The top of the valence band is mainly composed of F-2p orbitals while the conduction band bottom is composed of Si and K orbitals. Upon adding the Mn dopant to the supercell, Kohn-Sham single particle levels with symmetry labels t_2g and e_g are introduced in the band gap. These levels correspond to molecular orbitals composed of Mn-3d and F-2p character, indicating the existence of Mn-F hybridization [55]. The electronic band structure of K₂SiF₆, calculated with LDA and GGA is shown in Fig. 4(c), showing band gaps clearly higher than the mentioned experimental value [48].

 

Fig. 4 The Kohn-Sham density of states (DOS) for K₂SiF₆:Mn⁴⁺ calculated using DFT-PBE, ground state t_2g²e_g¹ (⁴A_2g) (a) and excited state t_2g²e_g¹ (²E_g) (b), reproduced from [55] with permission of the Royal Society of Chemistry. (c) The calculated band structure of K₂SiF₆. The GGA- and LDA-calculated electronic bands are shown by the solid and dotted lines, respectively. The Fermi level is set at zero energy. The coordinates of the special points of the Brillouin zone are (in terms of the reciprocal lattice unit vectors): W(1/2, 1/4, 3/4); L(1/2, 1/2, 1/2); G(0,0,0); X(1/2, 0, 1/2); K(3/8, 3/8, 3/4), reproduced from [48] by permission of The Electrochemical Society.

Download Full Size | PDF

Even though the use of DFT is essentially limited to ground state properties, the nature of the emitting transition in Mn⁴⁺-doped phosphors allows to estimate the energy of the ZPL by means of Δ-SCF, i.e. by forcing the system into an alternative occupation of single-particle levels. Du calculated the difference in total energy between the t_2g³ (corresponding to the ground state ⁴A_2g (⁴F)) and t_2g²e_g¹ (corresponding to the emitting level ²E_g (²G)) occupations of the Mn-3d shell [55]. This technique yields ZPL energies that are systematically too low, 1.02 eV and 1.29 eV in the case of K₂SiF₆:Mn⁴⁺ using the PBE and a HSE06 functional respectively. However, the systematic behavior as a function of the covalence of the chemical bond, empirically encoded in the parameter β or β₁, is found, paving the way for a semi-empirical method to predict the emission color of Mn⁴⁺-doped phosphors [55]. A small Jahn–Teller distortion towards a tetragonal symmetry was shown in the DOS of the excited ²E_g state in Fig. 4(b). Such a small deformation of the octahedral complex of the [MnF₆]^2- defect clusters can change the energy level degeneracy, which can influence the luminescence behavior.

3. Synthesis

Inorganic phosphors for LED applications are typically synthesized using energy consuming solid-state synthesis methods. Starting materials, such as metal oxides, nitrides or fluorides, are mixed together in powder form and baked for several hours at high temperatures up to 1600 °C [5,66]. While most LED phosphors are oxides or nitrides, K₂SiF₆:Mn⁴⁺ is the odd one out as a fluoride. Solid-state synthesis methods for fluorides have mainly been applied in lanthanide-containing fluorides such as NaYF₄ [67], but have not been applied on Mn⁴⁺-doped fluorides. All synthesis methods for Mn⁴⁺-doped solid fluorides use aqueous or anhydrous HF or even elemental F₂ as a fluorine source [68]. Although precautions have to be taken when using HF, it can be handled safely using the right protective equipment and avoiding glass labware. Possible synthesis methods using HF include etching, cocrystallization and (co)precipitation, and will be discussed here. A schematic overview of the four main synthesis methods is given in Fig. 5.

 

Fig. 5 Schematic overview of the main possible methods for K₂SiF₆:Mn⁴⁺ synthesis.

Download Full Size | PDF

Adachi and Takahashi proposed K₂SiF₆:Mn⁴⁺ as a phosphor for LEDs in 2008 [69]. Their synthesis was based on etching Si wafers in an solution of KMnO₄ in 25% HF, possibly with the addition of KHF₂ or MnF₂ to the etching solution [34,49,53,58,69,70]. The etching synthesis is schematically displayed in Fig. 5. Alternative Si sources were also explored, using crushed quartz schist [52] or powdered silica glass [62] with the same etching solutions. The chemical reaction of the formation of K₂SiF₆:Mn⁴⁺ can be represented by (2−2x) SiO₂ + 4 KMnO₄ + 12 HF $\to$ 2 K₂Si_1−xMn_xF₆ + (4−2x) MnO₂ + 6 H₂O + 3 O₂ [62]. An overview of the properties of phosphors synthesized using etching is given in Table 6 in the appendix. Although pure phosphor materials can be produced with high quantum efficiencies, the use of Si wafers as starting products makes this synthesis method rather expensive and slow due to the limited surface area available for reaction. Syntheses based on Si powder as a pure Si source, which would increase the surface area, have not been reported.

Although the interest of the LED market in K₂SiF₆:Mn⁴⁺ started in 2008, the red-emitting luminescent material was discovered 35 years earlier. The first description of the synthesis of K₂SiF₆:Mn⁴⁺ was given by Paulusz in 1973 [31]. It was prepared by recrystallizing mixtures of K₂MnF₆ and the host-lattice material K₂SiF₆. The host material can be synthesized using a precipitation method in HF based on SiO₂ and KF as starting materials [71]. Because K₂SiF₆ is a rather stable fluoride, it is now commercially available (CAS Number 16871-90-2) and it has also been applied in the electrodeposition of silicon [72].

K₂MnF₆ still has to be synthesized in-house, because it quickly decomposes under influence of water and/or temperature. It is produced according to the method of Bode et al. [73], using a redox reaction in a solution of KHF₂ and KMnO₄ in 40% HF. While the solution is cooled and stirred, H₂O₂ is added dropwise, after which the purple solution turns brown, indicating the reduction of Mn⁷⁺ to Mn⁴⁺, and a golden-yellow precipitate forms. The latter is washed with a KHF₂ solution in HF and with HF and then dried, yielding single phase K₂MnF₆. Several more recent publications use a similar cocrystallization synthesis for the preparation of K₂SiF₆:Mn⁴⁺, though with small adjustments like a higher HF concentration and drying at 70 °C [13,74]. For K₂TiF₆:Mn⁴⁺, it has been shown that during a cocrystallization synthesis in a small volume of HF solvent, a cation-exchange reaction takes place for the undissolved portion of K₂TiF₆ host while the mixture is heated to 70 °C, as is schematically shown in Fig. 6 [75].

 

Fig. 6 Synthesis of Mn⁴⁺-activated red phosphors using a cocrystallization synthesis. Photographs of the HF solution with dissolved K₂MnF₆ crystals (a) and the same solution containing K₂TiF₆ powders after cation exchange reaction for 3 min (b). (c) Schematic illustration of the cation exchange procedure for synthesizing Mn⁴⁺-activated fluoride compounds, reprinted by permission from Macmillan Publishers Ltd [75].

Download Full Size | PDF

In coprecipitation methods, K₂MnF₆ is dissolved in a solution of SiO₂ or H₂SiF₆ and KF or KHF₂ in HF, from which a K₂SiF₆:Mn⁴⁺ precipitate forms, as illustrated in Fig. 7 [63,64,76–85]. Other reagents can include (NH₄)₂SiF₆ [79], SiO₂ [80], H₂SiF₆ [82,84], SiF₄ [81], KHF₂ [80,84], and various potassium salts [79,81,82]. The difference with cocrystallization methods is that all precursors are fully dissolved before the phosphor precipitates, while cocrystallization occurs in a slurry. The solution is stirred at room temperature or in a water bath with a fixed temperature between −16 °C and 60 °C [51,57,61]. The reactions can be performed at different temperature regimes such as cooling a heated phosphor solution (>60 °C) to temperatures <30 °C [79], cooling to 0 °C [80,81] or by evaporating the solvent at 25–120 °C [79,81]. The precipitates are washed with 20% HF and/or ethanol or acetone and dried at room temperature, 70 °C or 100 °C. An overview of the properties of phosphors synthesized using coprecipitation is given in Table 7 in the appendix. Although it is not widely addressed in publications, the addition rate of the different precursors can be critical to obtain high quantum efficiencies, since the phosphor precipitation starts immediately when the Si precursor comes into contact with the K precursor and K₂MnF₆ [86]. Park et al. synthesized their K₂SiF₆:Mn⁴⁺ phosphor for LED applications by adding Si to a Mn-compound solution, after which a K-containing solution was added in intervals [85].

 

Fig. 7 Schematic illustration of a coprecipitation synthesis of K₂SiF₆:Mn⁴⁺. Appearance of the phosphor under day light and UV illumination (left).

Download Full Size | PDF

Besides the coprecipitation methods, where K₂MnF₆ is synthesized separately, a simple precipitation synthesis is possible from HF solution, which is illustrated in Fig. 8. For this, SiO₂ is dissolved in 25–48% HF to form a solution of H₂SiF₆, possibly under mild heating (60 °C). To this solution, KMnO₄ and/or KF or K₂CO₃ are added as K and Mn sources. Dropwise addition of H₂O₂ or mild heating (50 °C) of the solution can speed up the redox reaction that turns the solution from purple to colorless and forms the yellow K₂SiF₆:Mn⁴⁺ precipitate at room temperature or in an ice bath. Alternatively, the precipitate can be formed by addition of acetone as a nonsolvent. The precipitate is filtered, washed with 20% HF and/or ethanol or methanol and/or water and dried at various temperatures reaching from room temperature to 80 °C [14,23,56,71,87–93]. In the presence of H₂O₂, the chemical reaction of the formation of K₂SiF₆:Mn⁴⁺ can be represented by (2−2x) SiO₂ + 4 KMnO₄ + (20−4x) HF + (4−2x) H₂O₂ $\to$ 2 K₂Si_1−xMn_xF₆ + (4−2x) MnO₂ + (14−4x) H₂O + (7−2x) O₂. An overview of the properties of phosphors synthesized using precipitation methods is given in Table 8 in the appendix.

 

Fig. 8 Schematic illustration of a simple precipitation synthesis of K₂SiF₆:Mn⁴⁺ at 0 °C. Appearance of K₂SiF₆:Mn⁴⁺ under day light and UV illumination (right) reproduced with permission from [14], copyright 2016, The Electrochemical Society.

Download Full Size | PDF

Alternatively, a hydrothermal synthesis in an autoclave at 120 °C can also be performed [71]. The reaction process is similar to the simple precipitation method, but high temperatures and high pressures can improve the quality of the phosphor material. Although solvothermal synthesis methods are often used to improve the quality of the end product at lower temperatures, hydrothermal synthesis using HF increases the safety risks, so it is not easily scalable. Shape control during the synthesis, forming K₂SiF₆:Mn⁴⁺ nanorods, can be obtained using 100–500 nm SiO₂ particles and potassium oleate as reagents together with KMnO₄ in HF [94]. The amount of highly corrosive HF used can be reduced by using a mixed solution of glacial acetic acid and 40% HF as solvent for a precipitation synthesis at 50 °C [95]. A “HF-free” synthesis can be performed hydrothermally based on soluble Mn⁴⁺ in the form of Mn(HPO₄)₂ in phosphoric acid, mixed with SiO₂ and KHF₂, which are kept at 180 °C in an autoclave [96]. Although HF is not used as a solvent here, the reaction of KHF₂ with the phosphoric acid causes in situ formation of HF. Taking the high temperature and pressure of the hydrothermal synthesis into account, caution still has to be taken concerning proper handling of HF. K₂SiF₆:Mn⁴⁺ is commercially produced by General Electric, Mitshubishi Co, Force4 Corporation and Denka Company [97–100], but the different synthesis methods are still subject to research. When some of the HF in the synthesis is replaced by H₂SiF₆, a low-HF process compared to [79] is possible [83]. Some variation in the synthesis of K₂MnF₆ is possible, including an electrolysis-aided precipitation reaction of MnF₂ and KF in HF [80] and the removal of impurities by filtration of a (nearly) saturated solution of K₂MnF₆ in aqueous HF followed by precipitation induced by decreased temperature or nonsolvents [82]. An overview of the properties of phosphors synthesized using these alternative synthesis methods is given in Table 9 in the appendix. A true HF-free synthesis might come up in the future with the use of ionic liquids. In a microwave-assisted synthesis with [Bmim]PF₆ as the solvent and fluorine source, undoped K₂SiF₆ nanoparticles have recently been produced [101].

The phosphor powders can be synthesized using a one or two-step reaction, followed by filtration or evaporation to remove the excess HF solvent, washing to remove traces of HF and drying to obtain the final phosphor powder. Post-synthesis treatment is still subject to research, but can include coatings and/or annealing steps.

K₂SiF₆:Mn⁴⁺ is prone to moisture-induced degradation which is suspected to be a hydrolysis reaction of surface Mn⁴⁺ ions to MnO₂ [84]. The undoped K₂SiF₆ host can be used as a coating to minimize the contact of Mn⁴⁺ with water [79,102]. A coating of the phosphor is provided by contacting the phosphor with a saturated solution of the undoped K₂SiF₆ host and 40–70 w% HF at 20–50 °C [79]. Yoshida et al. prepared a similar undoped-host coating, by mixing phosphor particles with a solution of H₂SiF₆, to which H₂O₂ and a solution of KHF₂ are added [84]. Besides the undoped host material, other complex fluorides with the composition A₃[MF₆] (A = Na, K, Rb and M = Ga, In, Sc, Y, Gd) [103] or hydrophobic organic functional groups like alkyl phosphates [63], oleic acid [100], carboxyl, amines, quarternary ammonium salts, alkyl betaines, alkoxysilanes or fluorinated polymers can be used as coating materials to improve the humidity resistance of K₂SiF₆:Mn⁴⁺ [102,104]. The influence of the coatings will be discussed further, in the section on chemical stability.

Some mentions have been made of annealing treatments with an atmosphere that involves a F-containing oxidizing agent. An annealing procedure at 300–475 °C in a 20/80 F₂/N₂ atmosphere for 8 hours can be applied before and/or after the coating treatment [83,105]. In this way, it should be possible to get all Mn ions in a + 4 valence state, even if Mn²⁺ or Mn³⁺ have been formed during the synthesis. The presence of fluorine in the reaction atmosphere is essential, since a fluorine source has to be present to form stable [MnF₆]^2- complex anions. In an oxidizing atmosphere lacking fluorine, manganese oxides will be favored over fluorides. Caution has to be taken when performing such annealing treatments, since F₂ can form HF in contact with water, which can be still present in the powder if it was not fully dried. A heat treatment at 600 °C in a fluoride-containing atmosphere, possibly mixed with N₂ of Ar makes the K₂SiF₆:Mn⁴⁺ phosphor unlikely to degrade when used in high power WLEDs for long time [106].

4. Structural properties

From a structural point of view, there is little variation in reported crystal structures of K₂SiF₆:Mn⁴⁺. Apart from Nguyen et al., who report a hexagonal P321 crystal structure for K₂SiF₆:Mn⁴⁺ [63], all XRD measurements show the cubic Fm$\overline{3}$m structure with the lattice constant a = 8.13 Å. The cubic structure of K₂SiF₆:Mn⁴⁺ is apparent in the XRD pattern (Fig. 9). Cubic K₂SiF₆:Mn⁴⁺ is thermodynamically more stable than the hexagonal phase, with calculated formation enthalpies of −5.65322 × 10³ eV for the cubic phase and −5.65311 × 10³ eV for the hexagonal phase, leading to a difference of 110 meV per unit cell [61].

 

Fig. 9 XRD patterns of a K₂SiF₆:Mn⁴⁺ phosphor (b) and the reference pattern (ICSD 29407) for K₂SiF₆ (a) reproduced with permission from [14], copyright 2016, The Electrochemical Society.

Download Full Size | PDF

Introducing a dopant into a host crystal can lead to small alterations of the crystal structure, especially when the dopant ion has a different ionic radius than the ion it replaces. In K₂SiF₆:Mn⁴⁺ the Mn–F distance is expected to be larger than the Si–F distance since the ionic radius of Mn⁴⁺ (0.53 Å) is larger than that of Si⁴⁺ (0.40 Å) [107]. The shortest reported Mn–F bond length in K₂SiF₆:Mn⁴⁺ is 1.6829 Å [34,35], which is actually the Si–F bond length of 1.683(2) Å in K₂SiF₆ [108], wrongly reported as the Mn–F bond length. From EXAFS measurements, the Si–F bond length in K₂SiF₆:Mn⁴⁺ is calculated as 1.66 Å [51] and the Mn–F bond length in K₂MnF₆ is calculated as 1.809 Å [64]. The Mn–F bond length in K₂MnF₆ as obtained from XRD measurements is 1.79 Å [109]. When the Mn–F bond length is estimated according to the increased ionic radius, a bond length of 1.745 Å for Mn⁴⁺ in a K₂SiF₆ host results in an estimated expansion of the bond length of 3.9% [54]. With a Mn–F bond length in K₂SiF₆:Mn⁴⁺ of 1.807 Å calculated from EXAFS measurements, a lattice expansion of 7.37% is calculated for all A₂SiF₆ hosts [64], which is in accordance with the Mn–F bond length in K₂MnF₆. Although there is a lattice expansion when introducing the Mn dopant [64], no effect of this lattice expansion is seen in the diffraction patterns for low doping concentrations, when comparing the XRD pattern of K₂SiF₆:Mn⁴⁺ with the K₂SiF₆ reference pattern (Fig. 9).

The synthesized particles are typically µm sized, varying from 2 to 100 µm. Drymilling [13], laser ablation [70], or a shape-controlled synthesis [94] can reduce the particle size, but even if the crystallites are submicron sized (62–110 nm after laser ablation), the particle size will remain in the µm range due to cluster formation. The particles are mainly octahedral and cubic [14,51,95], but can also be decahedrons [96] or nanorods [94]. The body color of the phosphor is either white, light-yellow, yellow or orange, but cannot be directly linked to dopant concentration or QE, since those are not consistently reported. For a red LED phosphor, a yellow or orange body color is beneficial, since it indicates a good absorption of blue light. Only in the shape-controlled synthesis by Li et al., a pink body color was mentioned [94], which could be caused by the unintentional presence of KMnF₄, which has a darker body color than K₂MnF₆. An overview of the structural and luminescence properties of all published K₂SiF₆:Mn⁴⁺ phosphors are presented in Table 6, Table 7, Table 8, and Table 9 in the appendix.

More advanced structural characterization techniques such as EPR, Raman and EXAFS are occasionally reported for K₂SiF₆:Mn⁴⁺. In X-band EPR, Mn⁴⁺ shows 6 hyperfine components between 300 and 400 mT due to the ½$\to$−½ transition of Mn⁴⁺ with a d³ electronic configuration (Fig. 10). Similar EPR spectra have been reported for BaSnF₆:Mn⁴⁺ [50] and for ZnGeF₆·H₂O:Mn⁴⁺ [110]. Mn²⁺ and Mn⁴⁺ are both EPR-active and result in similar hyperfine structures, but the difference between both can be determined from a difference in g value. From the differential absorption value in EPR, it can be found that in a so-called HF-free synthesis, the actual dopant concentration is controlled by the amount of KHF₂ starting materials [96]. This can be expected since KHF₂ acts as the F source, and the Mn⁴⁺ ion can only be stabilized in the [MnF₆]^2- complex anion. EXAFS on K₂SiF₆:Mn⁴⁺ has only been used to calculate the Si–F and Mn–F bond length as discussed before [51,64]. Raman spectroscopy shows a shift of the peak at 117.6 cm⁻¹ (external vibrations involving the motion of the entire [SiF₆]^2- system) [57] and a shift of the peak at 473.3 cm⁻¹ (e_g stretching of Si–F in [SiF₆]^2-), both to lower wavenumbers, due to a slight structural distortion after Mn⁴⁺ doping [71]. This is consistent with the larger bond lengths found for Mn–F compared to Si–F. No changes in the Raman spectra were observed before and after compression to 50 kbar [57] or for a synthesis at −16 °C [61]. In Fourier transform infrared spectroscopy (FTIR), K₂SiF₆:Mn⁴⁺ shows characteristic absorption peaks at 740 cm⁻¹ and 484 cm⁻¹, which are assigned to Si–F bond vibrations [71,100], but no influence of the Mn dopant was found.

 

Fig. 10 X-band EPR spectra (a, b) of K₂SiF₆:Mn⁴⁺ synthesized with different Si:F ratios and differential EPR absorption intensities (c) reprinted with permission from [96]. Copyright 2016 American Chemical Society.

Download Full Size | PDF

4.1 Pressure dependence

The effects of the application of external pressure on the K₂SiF₆ crystal and the K₂SiF₆:Mn⁴⁺ phosphor in particular are not well-known. A discrete-variational multi-electron (DVME) study on the multiplet energy levels of cubic A₂MF₆:Mn⁴⁺ (A = K, Rb, Cs; M = Mn, Si, Ge) crystals showed that the Mn–F bond length decreases for increasing pressures, as has been observed previously in ruby [64]. This is consistent with ab initio calculations on the effect of pressure on the structural properties of K₂SiF₆ [48]. The calculated band gap grows nonlinearly with pressure while the lattice constant a and the Si–F distance have a linear pressure dependence. Both decrease, but the Si–F bond length decreases at a 30–40 times smaller rate compared to the lattice constant [48]. At atmospheric pressure, the calculated band gaps are still bigger than the once-mentioned experimental band gap, contrary to the normally underestimation of calculated band gaps.

The calculated absorption spectra of K₂SiF₆ show a blueshift at pressures from 0 to 200 kbar, as well as an enhanced absorption intensity. This can be caused by stronger overlap and mixture of the Si and F wave functions [48] and could have an influence on the absorption of K₂SiF₆:Mn⁴⁺ under pressure as well. The influence seems to be negligible though, since in K₂SiF₆:Mn⁴⁺, a slight redshift in the excitation spectrum is seen after a compression-decompression procedure up to 57 kbar [57], caused by a decrease in Racah parameters B and C. The blueshift of the K₂SiF₆ host absorption in the high-energy range (10–25 eV) [48] does not seem to influence the Mn⁴⁺ absorption in the low-energy range (1–5 eV) [57]. A pressure-induced blueshift of the Mn⁴⁺ absorption has been observed in the lower symmetric fluoride phosphor KNaSiF₆:Mn⁴⁺ [76].

At hydrostatic pressures above 9 kbar, an additional line in the emission spectrum of K₂SiF₆:Mn⁴⁺ appears at 624 nm. This line is attributed to the ZPL of the ²E_g$\to$⁴A_2g transition which is not visible in room-temperature emission spectra due to the high cubic crystal symmetry of K₂SiF₆. Further increasing the pressure up to 220 kbar does not change the spectrum apart from a redshift of all lines equal to −0.1 nm/kbar as can be seen in Fig. 11. Upon releasing pressure, all observed lines are returning to their previous positions and the ZPL remains visible at 622 nm even at ambient pressure [57]. Since the presence of the ZPL is induced by a reduced crystal symmetry, the presence of the ZPL after the compression-decompression procedure shows that the change in crystal structure persists after the release of pressure. The pressure-induced redshift of the emission has also been observed in KNaSiF₆:Mn⁴⁺ [76].

 

Fig. 11 Emission spectra of K₂SiF₆:Mn⁴⁺ excited at 442 nm under varying pressures, reprinted from [57], with permission of AIP Publishing.

Download Full Size | PDF

The decay time shortens from 9 to 6 ms after a compression-decompression procedure. The pressure breaks the local symmetry of the environment of the Mn⁴⁺ ion, activating the ZPL and partly releasing the restrictions of the spin-forbidden ²E_g$\to$⁴A_2g transition. This influence on the decay time is not seen in the lower symmetric KNaSiF₆:Mn⁴⁺, where decay times increase from 6 to 12 ms under pressure up to 270 kbar [76]. Although a hexagonal crystal structure exists for K₂SiF₆ exists, phase transitions are not expected. Both XRD and Raman spectroscopy confirm that the cubic phase is maintained after a compression-decompression procedure, with peak broadening as its only effect [57].

5. Luminescent properties

All reported luminescence excitation and emission spectra of K₂SiF₆:Mn⁴⁺are similar. The emission spectrum of K₂SiF₆:Mn⁴⁺ was first reported in 1973 and was recorded at 90 K, see Fig. 12 [31]. At least two excitation bands are reported, at 340–360 nm and 450–470 nm, with the latter one being the most intense band. The excitation bands are related to the spin-allowed ⁴A_2g$\to$⁴T_1g (⁴F) and ⁴A_2g$\to$⁴T_2g (⁴F) transitions. The full width at half maximum (FWHM) of these excitation bands is around 41–54 nm at room temperature.

 

Fig. 12 Emission spectrum at 90 K of K₂SiF₆ doped with 1 mol% Mn as published by Paulusz in 1973. Weak emission at shorter and larger wavelengths is shown on an enlarged scale (dashed lines) [31].

Download Full Size | PDF

A third excitation band at lower wavelengths, around 240–260 nm is rarely reported, but is visible in Fig. 13. The intensity of this UV excitation band is rather low compared to the other excitation bands and most reported spectra only start at 300 nm. The third excitation band is assigned to the ⁴A_2g$\to$⁴T_1g (⁴P) transition, but can be hidden by charge-transfer transitions or host–lattice absorption [35]. The excitation and emission spectra recorded at 10 K in Fig. 13 clearly show the vibronic fine structure [14]. In diffuse reflectance ultraviolet-visible spectra (DRS), the strongest absorption peak is located at 455 nm [34,91], while the undoped host K₂SiF₆ shows no absorption at wavelengths larger than 250 nm [34].

 

Fig. 13 Excitation (black) and emission (red) spectrum of K₂SiF₆:Mn⁴⁺ at 10 K (left) and 300 K (right), adapted with permission from [14], copyright 2016, The Electrochemical Society. The electronic transitions are labeled in the 10 K spectrum, the vibrational modes of the emission are labeled in the spectrum at 300 K. The dopant concentration is 1.8%.

Download Full Size | PDF

The emission spectrum of K₂SiF₆:Mn⁴⁺ does not depend on the synthesis conditions. The number of reported emission peaks can differ however, but this is mainly because of non-resolved close-lying peaks. When the emission spectrum is measured at a nm resolution, mostly five or six peaks are reported. The main vibronic modes coupled to the ²E_g$\to$⁴A_2g transition are the antisymmetric v₃(T_1u), v₄(T_1u) and v₆(T_2u).

There is a small divergence in the reported peak wavelength of the emission, with values reported from 629 to 634 nm. The ZPL for the ²E_g$\to$⁴A_2g transition is located between 619 and 630 nm, but is usually not present in the emission spectrum due to the cubic crystal symmetry of K₂SiF₆:Mn⁴⁺. When the ZPL is present in experimental spectra, it is either measured at low intensity in a hexagonal crystal structure [63], after pressure was applied on the material [57], or at low (10–90 K) temperatures [14,31,53].

It is reported that the emission intensity of phosphors that were prepared by etching synthesis of crushed quartz schist is 1–2 orders lower than from Si wafer synthesis [52]. The QE is however better suited to assess the luminescence intensities because the excitation intensity and absorption characteristic are then properly accounted for. Due to the nanoporous surface of the phosphor prepared from quartz schist, increased scattering might lower the absorption, lowering the EQE. Though quantum efficiencies are not always consistently reported, values up to 90% of IQE and 69% of EQE have been reached by coprecipitation synthesis [80].

Controlling the dopant concentration seems to be a continuous issue in the synthesis of K₂SiF₆:Mn⁴⁺. According to Lyons et al., only 2–8 mol% of Mn is incorporated in the phosphor when 8–12 mol% of Mn is used during the synthesis [79]. Even for cocrystallization and coprecipitation methods, the initial synthesis concentration of Mn is always higher than the final measured dopant concentration. While no side products are supposed to form, surface clusters of Mn can significantly lower the dopant concentration in the final product. When etching wafers, the Mn concentration detected using X-ray photoelectron spectroscopy (XPS) is <1 mol% [49]. In a precipitation synthesis, typically an excess of KMnO₄ is present, but the dopant concentration as detected by energy-dispersive X-ray spectroscopy (EDX) is <1–3.6 mol% [14,63,71,87]. With atomic absorption spectroscopy (AAS), lower concentrations such as 0.87 mol% of Mn can be detected [95]. Although not all present Mn is incorporated, the concentration of KMnO₄ in the synthesis influences the dopant incorporation, as is shown by Yeo et al., where a range of 15–30 mmol of KMnO₄ in 70 mL HF results in a Mn concentration of 2.08–3.77 mol% as calculated from inductively coupled plasma mass spectrometry (ICP-MS) [92]. A longer, six-hour synthesis can result in a dopant concentration of 13 mol% [89]. In a coprecipitation synthesis, the concentration that is measured is mostly lower than the intended concentration. When aiming for 0.01 mol%, ICP-MS analysis showed a nondetectable amount (<0.004 mol%) and 2.46 mol% was detected when aiming for 10 mol%. Moreover, XPS showed that most of the present Mn was Mn³⁺, leading to a concentration of 0.59 mol% of Mn⁴⁺ instead of the aimed 10 mol% [77]. Mn³⁺ can influence the absorption but will not contribute to the red emission of the phosphor. Although the + 4 valence state is expected to be stable in the [MnF₆]^2- complex anion, variation in synthesis conditions can influence the valence state of dopants and introduce Mn³⁺ contamination. By using cation exchange in a cocrystallization reaction, it is possible to create a range of different Mn concentrations from 0.72 to 6.50 mol% in K₂TiF₆:Mn [75]. It has been shown that concentration quenching becomes significant around 3.3 mol% Mn in K₂SiF₆, but this can be improved to above 7.6 mol% with reduced defect concentrations [97].

Full data sets on the luminescence properties of K₂SiF₆:Mn⁴⁺from different sources, including luminescence lifetime, quantum efficiency and thermal quenching, are hardly available. From the data presented in Table 6 to Table 9 in the appendix, some trends can however be discerned. The reported decay times at room temperature, are τ = 8.1 ± 1.0 ms, when averaged over all reports. In reports of precipitation synthesis, a slightly lower average of τ = 7.7 ± 0.5 ms is found. This can indicate an increase in non-radiative transitions due to synthesis-induced defects. Mostly mono-exponential decay profiles are reported, but a second order decay was found at increased temperatures. The luminescent-lifetime measurements in Fig. 14 show a mono-exponential decay at 220 K with a decay time τ = 10.5 ms. With increasing temperature (in the 295–450 K range), a bi-exponential decay is required for the fit since a second, faster decay component emerges. The time constant of the slow component decreases with increasing temperature, reaching τ = 8.1 ms at room temperature and τ = 4.9 ms at 450 K [14]. Even when a mono-exponential decay is found, the decay time decreases with increasing temperature, decreasing from 8.2 ms at room temperature to 2.4 ms at 470 K [56]. The absorption rate changes very little with temperature, so the radiative and non-radiative transition probabilities can be calculated from the quantum efficiency and decay times. Since the emission intensity does not drop until 420 K, as will be discussed further, this shows that the non-radiative transition probability remains stable from 290 to 420 K, while the radiative transition probability increases [56]. The relatively slow decay can cause saturation of the excited state and high excitation power have shown to lead to intensity saturation starting at 25–40 W/cm² [14,97].

 

Fig. 14 Decay-profile measurements (dots) and fit (lines) of the luminescence intensity of K₂SiF₆:Mn⁴⁺ at 450 K (a), 295 K (b) and 220 K (c) reproduced with permission from [14], copyright 2016, The Electrochemical Society.

Download Full Size | PDF

For the samples where both the IQE and the particle size are reported, we see two sets of data in Fig. 15 (red). The lowest IQE values are reported for smaller particle sizes (4–10 µm), while somewhat bigger particles (10–40 µm) result in IQE values of 70–90%. When comparing the decay time to the reported particle size in Fig. 15 (green), no clear trend can be seen. There are several publications where the particles are 10 µm, but even for an equal particle size, there is a spread of decay times from 7.1 to 9.7 ms. Although the data set is not complete, one can conclude that variations in the decay time are not correlated to the particle size, but merely point to other influences, such as defects.

 

Fig. 15 Internal quantum efficiency (red) and decay time (green) as function of particle size. The according data can be found in Table 6–Table 9 in the Appendix.

Download Full Size | PDF

6. Applications

With emission essentially situated between 600 and 650 nm, the red-emitting K₂SiF₆:Mn⁴⁺ phosphor can be applied in both LEDs for display backlights and in warm-white LEDs for lighting applications. Both types of applications call for efficient red phosphors, but the required specifications are different. K₂SiF₆:Mn⁴⁺ is included as a possible red phosphor in several patents for LED applications, including LED televisions and edge-lit displays which demand a large color gamut and for silicon-resin sheets for illumination applications [111–122]. K₂SiF₆:Mn⁴⁺ has even been proposed for use in head lamps for cars, although the saturation behavior of K₂SiF₆:Mn⁴⁺ at high excitation fluxes seems incompatible with the high-brightness requirements [116]. Besides lighting applications, K₂SiF₆:Mn⁴⁺ has been proposed as a downshifting phosphor for solar cell applications [93] and its stable emission can be used for magnetic field sensing and calibration [123]. For structural optimization of white LEDs, K₂SiF₆:Mn⁴⁺ has been proposed as temperature sensor, due to the linear dependence of the anti-Stokes to Stokes emission intensity ratio [56]. Nevertheless displays and general lighting are the two major applications of this phosphor, justifying a more detailed discussion of the phosphor requirements.

6.1 Displays

Display applications require saturated colors to achieve the widest possible color gamut. To achieve the Rec. 2020 standard, the primary RGB colors are ideally composed of monochromatic light located at respectively 630 nm, 532 nm and 467 nm [17]. Rec. 2020 improves the color gamut requirements compared to the 1953 National Television Standards Committee (NTSC) color gamut by 34%. Using broad-emitting phosphors, narrow-band color filters are needed to obtain the mandatory nearly-monochromatic light. This reduces the total efficiency of the system, filtering away light from the phosphor emission, introducing a trade-off between color saturation and energy efficiency. The narrow emission spectrum of K₂SiF₆:Mn⁴⁺ is an advantage for display applications, since less emission needs to be filtered out and the peak emission is located at the right position of 630 nm. In contrast to Eu²⁺-based phosphors with comparable peak wavelength, K₂SiF₆:Mn⁴⁺ hardly contains emission beyond 650nm, where the eye sensitivity steeply drops. When K₂SiF₆:Mn⁴⁺ is combined with a blue LED at 467 nm, only efficient narrow green emission around 532 nm is needed to complete the full color gamut. Narrow-band green phosphors are not yet available; β-SiAlON:Eu²⁺ and SrGa₂S₄:Eu²⁺ both have a FWHM of about 50 nm [5,124], but color filters can be applied to obtain a narrow green emission by cutting the tails of the emission band. Semiconductor nanocrystals (or quantum dots) are the main alternatives in display applications, thanks to their tunable and narrow emission [2]. Although initially based on CdS/CdSe core-shell QDs, cadmium free QDs start to arise, such as InP or the lead halide perovskites [2,125]. The properties obtained when K₂SiF₆:Mn⁴⁺ is combined with green phosphors or QDs and a blue LED chip are presented in Table 3, aiming for a high color gamut. When ideal monochromatic RGB light is used according to the Rec. 2020 standard, the color gamut reaches up to 133.9% NTSC in the CIE 1931 color space. Combinations of K₂SiF₆:Mn⁴⁺ and green phosphors or QDs reach color gamuts of 85–127% NTSC, as can be seen in Table 3, with the highest value close to the ideal limit, using CH₃NH₃PbBr₃/NaNO₃ perovskite QDs. Interestingly, the current application of QDs is situated almost uniquely in display applications, characterized by less stringent conditions on temperature, flux and lifetime compared to general lighting. K₂SiF₆:Mn⁴⁺, having similar application restrictions, is in the same playing field as the QDs.

Table 3. Lighting properties of K₂SiF₆:Mn⁴⁺ with green phosphors or QDs and a blue LED suited for displays.

View Table | View all tables in this article

In display applications, red QDs such as CdSe/CdS, InP/ZnSe or lead-halide perovskites will be the main competitors of K₂SiF₆:Mn⁴⁺ [2,126–128]. It will be interesting to see how the competition between QDs and phosphors evolves in the coming years. While QDs offer superior, strongly tunable, narrow-band emission compared to the common broad-band emitting Eu²⁺ or Ce³⁺ phosphors, their long-term (photo)stability currently limits their use to applications. Especially in display applications, where driving conditions are relatively benign and where their spectral properties are most beneficial, QDs are limited by their stability. The emergence of Mn⁴⁺-based red phosphors can be considered disruptive for the application of QDs, as they offer similar spectral properties as QDs, while maintaining basic properties of impurity-doped phosphors, such as a stable emission colour and a stable intensity up to fairly high temperatures. The main disadvantages of Mn⁴⁺ are the need for fluoride chemistry and often, similar to QDs, stability issues which have not yet been fully understood. Hence it is clear that the outcome of this competition is not a priori clear. For general lighting applications, characterized by much higher photon fluxes for extended durations, QDs cannot yet compete with impurity-doped bulk phosphors, although the currently developed approaches to allow their use in display applications can eventually pave the way for use in general lighting. Similarly, it are interesting times to see which share Mn⁴⁺-doped phosphors will acquire in the market of general lighting, currently being solely occupied by Eu²⁺ and Ce³⁺-doped phosphors.

6.2 Lighting

Although high color gamuts can be reached by combining K₂SiF₆:Mn⁴⁺ with green phosphors or QDs, these phosphor combinations cannot straightforwardly be transferred to lighting applications. Although the display solution of combining narrow band red, green and blue components leads to white light and a wide range of possible displayed colors, it falls short for lighting applications, where color rendering and color temperature are key factors. Narrow RGB white LEDs lack yellow emission, hindering the optimal rendering of yellow objects.

For lighting applications, a full coverage of the visible spectrum is desired to accurately render the color of objects, like for a broad-emitting black-body radiating light source, such as the sun or an incandescent lamp. For energy-efficient lighting however, all invisible radiation, beyond the eye-sensitivity curve, should be omitted. Therefore, the combination of a blue LED with a broad-band yellow phosphor and a narrow red emission can be tailored to specific lighting specifications, such as the correlated color temperature (CCT) and the color rendering index (CRI). When K₂SiF₆:Mn⁴⁺ is combined with a LED (350–550 nm) and additional phosphors, CCT values of 2500–4500 K and CRI values of 70–95 can be reached [79].

YAG:Ce is a classic LED phosphor lacking emission in the red spectral region, making it ideally suited to be combined with the red-emitting K₂SiF₆:Mn⁴⁺. Due to the large offset between the excitation and emission bands in the spectrum of K₂SiF₆:Mn⁴⁺, there is essentially no reabsorption of YAG:Ce emission by the red phosphor. Therefore, the pc-WLED can be simulated by weighted addition of the emission spectra of a blue LED, YAG:Ce and K₂SiF₆:Mn⁴⁺. In Fig. 16, x is the K₂SiF₆:Mn⁴⁺ content, while 1−x is the YAG:Ce content in the total phosphor amount. The phosphor amount can be an actual wt% of phosphor added to the LED chip or a (simulated) ratio of converted photons. The emission spectra are normalized on the amount of emitted photons, implying an IQE of 100%. For each x, the D_UV value (i.e. the deviation from the black body locus) is minimized for a fixed CCT by varying the ratio between the blue emission from the LED and the yellow and red emission from the phosphor blend. When increasing the amount of K₂SiF₆:Mn⁴⁺ in the phosphor mixture, the CCT in Fig. 16 can be tuned from 8376 K for pure YAG:Ce to 2700 K for 61.7% YAG:Ce. Both the LER and the CRI in Fig. 16 reach a maximum value for x = 0.27. It is important that in a wide range of CCT values (4750–2700 K), a sufficiently high CRI can be obtained (>80) while the LER remains above 320 lm/W, which is not the case when red phosphors with a much broader bandwidth are used. In that case, lowering the CCT results in a relatively larger spill of photons beyond 650 nm, where the eye sensitivity is low. The optimal combination of 63% YAG:Ce and 27% K₂SiF₆:Mn⁴⁺ results in a CCT of 3466 K, a CRI of 90 and a LER of 337 lm/W. When only YAG:Ce is used in a pc-WLED, the range of available CCT values is lower. In Fig. 16, the CRI values for YAG:Ce are projected on the CCT values of a YAG:Ce–K₂SiF₆:Mn⁴⁺ pc-WLED, and they show a dramatic drop with decreasing color temperature. This shows the need for a red phosphor for warm-white light in YAG:Ce pc-WLEDs.

 

Fig. 16 Range of correlated color temperature (CCT) (blue), radiant luminous efficacy (LER) (green) and color rendering index (CRI) (red) for a pc-WLED with YAG:Ce and K₂SiF₆:Mn⁴⁺ phosphors (adapted with permission from [14], copyright 2016, The Electrochemical Society.) CRI values (yellow) for a pc-WLED with YAG:Ce are projected on the x-axis for equal CCT values. The black lines are guides to the eye for a CCT of 6500 K, 4500 K, 3000 K and 2800 K.

Download Full Size | PDF

Apart from simulations, experimental pc-WLED setups are also used to evaluate the properties of phosphor combinations. The lighting properties in Table 4 are from an experimental YAG:Ce–K₂SiF₆:Mn⁴⁺ pc-WLED setup or from simulations thereof. The difference between the simulated data in Fig. 16 and the experimental data is that perfect IQE values are assumed in simulations, while absorption and scattering effects can lower the QE and thereby luminescent output in experimental setups. In general, the values for CCT and CRI in Table 4 are in accordance with the simulations in Fig. 16. These combinations are all best suited for lighting applications, aiming for efficient white light. The obtained CCT values reach from 2700 K (warm white) over 3900 K (cool white) to 5296 K (blue-white). This overview shows the broad possibilities with K₂SiF₆:Mn⁴⁺ upon combination with YAG:Ce. Comparing the efficacy for the different phosphor combinations is difficult, because the overall efficacy is largely determined by the electrical-to-optical conversion efficiency of the blue pumping LED. This is reflected in the efficacy values given in Table 4, reaching from 36.7 lm/W to 129.3 lm/W. For YAG:Ce combined with K₂SiF₆:Mn⁴⁺ a CRI value of 63.9–94 is reached for an average phosphor amount of one-third of K₂SiF₆:Mn⁴⁺ and two-thirds of YAG:Ce.

Table 4. Lighting properties of K₂SiF₆:Mn⁴⁺ with YAG:Ce and a blue LED suited for lighting.

View Table | View all tables in this article

Although good lighting properties can be obtained using YAG:Ce, K₂SiF₆:Mn⁴⁺ can also be combined with other phosphors or QDs to improve the lighting properties even further. The lighting properties from experimental phosphor/QD–K₂SiF₆:Mn⁴⁺ pc-WLED setups or from simulations thereof are presented in Table 5. LuAG:Ce is very similar to YAG:Ce, but substitution of the Y ion for Lu causes a blueshift of the Ce³⁺ emission, which allows to narrow the cyan gap [136]. Adding K₂SiF₆:Mn⁴⁺ will increase the WLED emission in the red region, but the blueshifted Ce³⁺ emission will be less intense in the orange region, resulting in lower CRI values for warm-white CCT values. Red phosphors with a broader emission band like Sr₂Si₅N₈:Eu²⁺ profit more from the use of LuAG:Ce since the broad emission covers the orange region [135]. The gap in the orange region can be filled by adding (Sr,Ca)AlSiN₃:Eu²⁺ to the mix of LuAG:Ce and K₂SiF₆:Mn⁴⁺, resulting in an CRI of 97. Although the CRI values are higher for the combination of BAM:Eu²⁺ and SrSi₂N₂O₂:Eu²⁺ with K₂SiF₆:Mn⁴⁺ compared to the LuAG–K₂SiF₆:Mn⁴⁺ phosphor combination, UV-LED excitation is necessary for BAM:Eu²⁺, severely worsening the overall electrical–optical efficiency [18]. Even though it is possible to excite K₂SiF₆:Mn⁴⁺ with UV-light around 360 nm, Fig. 13 shows that excitation with blue light is more efficient. Moreover, the choice for BAM:Eu²⁺ is not ideal, since its emission at 450 nm is efficiently absorbed by K₂SiF₆:Mn⁴⁺, further reducing the total conversion efficiency. High CRI values have been reported by GE without mentioning the composition of the yellow phosphor used in the phosphor blend [97,99], and by using alkaline earth silicate phosphors doped with Eu²⁺ [23].

Table 5. Lighting properties of K₂SiF₆:Mn⁴⁺ combined with phosphors and a blue LED suited for lighting.

View Table | View all tables in this article

The broadness and peak position of the red-emitting phosphor are more critical in the warm white than in the cool white LED from the viewpoint of optimizing human perception of color. The LER of Mn⁴⁺-doped fluoride phosphors increases with the decrease of local symmetry because of higher ZPL emission resulting in a slight blueshift of the total emission. The optimization of color rendition is a trade-off, for Mn⁴⁺-doped fluoride phosphors with <30 nm bandwidths, the blueshifted color is better for energy efficiency while the redshifted color is beneficial for color rendition. Apart from a good color rendering and a high efficiency, a long lifetime is expected for pc-WLEDs, up to 50,000 hours [86]. To reach this long lifetime, the phosphor has to be environmentally stable during this lifetime.

7. Thermal stability

Thermal properties are important to take into account for LED phosphors, since some heating is inevitable. There are several heat sources that influence the red phosphor in a WLED [138]. The first heat source, the Stokes losses, are intrinsic to all phosphor materials and are most prominent in red phosphors. After the absorption of a blue photon, the phosphor relaxes to the emitting excited state before a red photon is finally emitted, leading to the dissipation of roughly 750 meV for every converted photon. The second heat source is the blue LED chip, as the conversion of electrical power into photons is not fully efficient, i.e. when the IQE is less 100% [3]. Those combined effects can easily lead to temperatures up to 450 K at the phosphor level. To increase the performance of the phosphor in high-power LEDs at elevated operating temperatures, a remote-phosphor approach can be used. By separating the phosphor layer from the LED chip, the operating temperature of the phosphor can be lowered due to the lower excitation flux [14]. The thermal design should be such that the Stokes losses from the conversion process, which amount to 30% of the energy of the incident blue photon flux, are adequately removed from the remote phosphor plate by conducting heat sinks [26,139].

The overall thermal stability of the luminescent properties of K₂SiF₆:Mn⁴⁺ phosphors is good and better than certain red Eu²⁺-doped phosphors, such as SrS:Eu²⁺ [140]. The absorption as measured by DRS is stable in the measured range from 293 to 353 K [56]. The integrated emission intensity (Fig. 17) shows an abnormal enhancement up to 300 K [52,56,62,69] rather than the typical thermal quenching behavior often found in phosphors [58]. The intensity increase is caused by the combination of the expected peak broadening and the increase in peak intensity of the anti-Stokes emission lines upon increasing temperature [56]. The non-radiative transition probability remains stable from 290 to 420 K, while the radiative transition probability increases. Therefore, the increase in emission intensity with increasing temperature should be attributed to enhanced phonon-assisted radiative transitions which increase the QE. The stronger electron–phonon interaction at higher temperatures results in increased radiative transition probabilities for the spin- and parity-forbidden ²E_g$\to$⁴A_2g transition [56].

 

Fig. 17 Integrated emission intensity of K₂SiF₆:Mn⁴⁺ as a function of temperature, adapted with permission from [14], copyright 2016, The Electrochemical Society. The dashed line is the fit from [52] for the increasing emission intensity in the low temperature range.

Download Full Size | PDF

As a side note, the thermal stability up to room temperature can be utilized as a probe for high magnetic field sensing over the 5–300 K temperature range. Application of a magnetic field up to 35 T induces a Zeeman splitting into subpeaks with a linear splitting energy. With a wide range of possible excitation sources, this broadens the scope of magnetic field sensing and magnetic field calibration under a pulsed high magnetic field [123].

Although the integrated emission intensity initially increases, at a certain temperature thermal quenching takes over. A stable emission intensity is reported up to 400–470 K [14,49,51,53,88,94,97,98], and in some cases the intensity is even still increasing up to 440–453 K [91,96]. In phosphors from etching synthesis, an increase in emission intensity starts already at 300 K [52,62,69]. The reported quenching energies, calculated from the Arrhenius equation, reach from 0.81 to 1.09 eV [49,53,56,63,88], with quenching attributed to thermally activated cross-over from the parabola of the ²E_g excited state to the ⁴A_2g ground state [49,53]. Coating the phosphor seems to increases the thermal stability [63]. Only Wang et al. find a lower value of 0.2 eV for the quenching energy, with a quenching behavior that starts already at room temperature [78]. The reported quenching temperatures T_1/2 reach from 430 to 520 K [14,31,49,78]. A minimal thermal quenching (<5%) is reported at 423 K [13,97], while other samples show a 24% decrease of the initial emission intensity at this temperature [78]. When coated with an alkyl phosphate coating, an improved thermal stability up to 523 K is reported, compared to a 6% intensity loss for the uncoated phosphor at this temperature [63].

8. Chemical stability

Although thermal quenching is fairly limited, chemical stability still is an issue for K₂SiF₆:Mn⁴⁺ phosphors, which is often overseen. Thermal decomposition of the phosphor is reported in the 638–673 K range [14,71]. Degradation tests at high temperature and high humidity (HTHH) conditions, typically 85 °C and 85% relative humidity, are reported occasionally and are a good way to consider stability of phosphor materials over the long expected lifetime of 50.000 hours in LEDs. The yellow phosphor can darken when exposed to HTHH conditions due to hydrolysis of [MnF₆]^2- groups at the surface to MnO₂, which has a black body color [97,100]. The combination of a blue LED and HTHH conditions can cause a drop of 23% in total luminous flux in a WLED with β-SiAlON:Eu²⁺ and K₂SiF₆:Mn⁴⁺ as phosphor materials [98]. A K₂SiF₆:Mn⁴⁺ powder commercially available from Mitsubishi Co. maintains 76.2% of its initial emission intensity after 1000 hours at HTHH although when boiled in water, it quickly turns brown within 1 hour [98].

An undoped K₂SiF₆ coating improves the phosphor stability in HTHH conditions, inducing a loss in QE of only 5% after 48 hours, compared to a 33% loss for an uncoated phosphor [79]. A K₂SiF₆ coating on a commercially available K₂SiF₆:Mn⁴⁺ powder showed that it protected against hydrolysis of surface groups, since the body color darkening at HTHH conditions was reduced [97]. An alkyl phosphate coating improves the stability of K₂SiF₆:Mn⁴⁺, both when immersed in water for 60 minutes and for 30 days at HTHH conditions. After 60 minutes in water, the emission of the coated sample was halved, while only 9% of the initial emission intensity of the uncoated sample remained. After 30 days of HTHH conditions, the relative IQE of the uncoated sample was 57%, while the coated sample still reached 85% of its initial IQE. A phosphor-converted WLED (pc-WLED) setup with β-SiAlON:Eu and commercially-available K₂SiF₆:Mn⁴⁺ reached 50% of its initial intensity after 3660 hours at HTHH, while the uncoated K₂SiF₆:Mn⁴⁺ took 4627 hours and the alkyl-phosphate-coated K₂SiF₆:Mn⁴⁺ took 8159 hours to reach 50% of the initial intensity [63]. After 12 hours at HTHH conditions, the IQE of a phosphor with a low Mn concentration of 1.21 mol% remained stable, while higher doped samples showed a drastic efficiency decrease [85]. A phosphor sample coated with 10–20 nm oleic acid stayed stable when immersed in water for 30 minutes and showed discoloration due to hydrolysis of [MnF₆]^2- to MnO₂ after 4 hours in water in 25% of the powder. After 15 days in water, the red emission still remained, while in the uncoated K₂SiF₆:Mn⁴⁺, the Mn ions were fully hydrolized to MnO₂ [100]. In a WLED setup, the phosphor coated with oleic acid remained at 85% of the initial emission intensity after 450 hours at HTHH conditions, while the commercial phosphor from Denka Corporation lost 23% of its initial emission intensity, although it also contained a protective coating [100].

Murphy et al. improved the color stability of their phosphor by contacting the fluoride phosphor with a solution of H₂SiF₆ and possibly K₂SiF₆ at 20–70 °C for 5–60 minutes. In comparison, samples that were treated with the H₂SiF₆/K₂SiF₆ solution, either with or without F₂ annealing, showed the best balance of properties after ageing at HTHH conditions and subjection to high-flux light conditions. The annealing especially improved resistance against laser damage, which also showed in the control samples. Upon treatment, XPS measurements showed a decrease in surface Mn as well as a reduced C and O concentration [83,105]. For the K₂SiF₆ coating, a coating with a thickness of 1.2 µm on a particle of 16 µm improved quantum efficiencies, increasing the IQE from 84.7% to 87.6% and the EQE from 58.8% to 60.2% [102]. Although the QE increase is only small, it proves that the benefits in improved moisture stability have no drawback on the QE. The stability of the phosphor in HTHH conditions increased with increasing coating thicknesses. The thinnest organic coated particles reached an IQE of 87.1% and an EQE of 60.4%. After 15 hours at HTHH conditions, the IQE and EQE of both coated and uncoated phosphors dropped, but the fluoride-coated particles kept 30% higher efficiencies compared to uncoated phosphors [102]. Complex fluorides with the composition A₃[MF₆] have a lower water solubility compared to K₂SiF₆:Mn⁴⁺ [103], which makes them useful as moisture protecting coatings.

9. Conclusions and perspectives

In the past years, interest in K₂SiF₆:Mn⁴⁺ as a narrow-band red phosphor for phosphor converted LEDs increased, as transition-metal dopants can replace the traditional rare-earth dopants. The spectroscopic properties of the material were widely investigated, although there is some disagreement on the obtained spectroscopic parameters and the location of the zero phonon lines. The ZPL of the emission has a low intensity and can only be observed in some low-temperature or high-pressure conditions, causing some uncertainty on spectroscopic parameters that are based on ZPL values, such as B and Dq. First principles calculations have only been performed to a limited extent, although they show promising results that approach experimental values up to a few hundred meV. A consensus on the band gap value for K₂SiF₆:Mn⁴⁺ based on calculations and a single experiment has not been reached yet.

The synthesis methods have evolved over the years, after the initial use of etching of various Si sources. Now coprecipitation and precipitation synthesis methods are preferred since they are efficient and more cost-effective than etching. Hydrothermal synthesis methods have only been used seldomly, due to the safety hazards with HF at elevated pressures and temperatures. A HF-free synthesis would strongly decrease the safety hazards associated to the current synthesis methods. As a first step in the direction of a HF-free synthesis of K₂SiF₆:Mn⁴⁺, an ionic liquid synthesis has been explored, but has not yet resulted in a phosphor synthesis. Since the fluorine ions are a crucial part of the crystal lattice and because of their sensitivity to oxidation and hydrolysis, extreme acidic environments and a fluorine excess seem necessary synthesis conditions. Even solution synthesis methods that claim to be HF free produce HF in situ. Further progress in synthesis methods is mainly expected in post-processing, including coatings and annealing steps. Coating methods include organic coatings and coatings with the undoped host material K₂SiF₆. Coatings have proven to be beneficial in terms of moisture stability and QE. Annealing treatments, common for many non-fluoride phosphors, have been proposed, but the fluorine-containing atmosphere must be approached with caution, since it can easily form HF in contact with water. Structural studies of post-synthesis methods can result in an optimized, stable phosphor, compatible with the long lifetime of pc-WLED applications.

All reported synthesis methods result in similar structural properties for the cubic crystal lattice. There is only a small variation in the luminescence spectra, mostly caused by instrumentation issues. Overall it is clear that the narrow emission around 630 nm and the broad excitation bands around 460 nm are ideally suited for pc-WLED and backlighting applications. Unfortunately, other important parameters, such as QE, decay times and effective dopant concentrations are not consistently measured and/or reported. Reported dopant concentrations should be treated with caution, since they can reach the instrumentation sensitivity limit. From the available overview, it can be seen that somewhat bigger particles (10–40 µm) yield higher quantum efficiencies. Although quantum efficiencies up to 90% are reported, most reported values are lower. Controlling the dopant concentration is a problem in many syntheses, since dopant concentrations are rather low and increasing the Mn concentration during the synthesis does not necessarily increase the dopant concentration. More research opportunities may lie in advanced structural analyses using EPR or EXAFS measurements to detect variations in incorporation and valence states. Note that only Mn⁴⁺ contributes to the red emission, while Mn³⁺ contamination can influence the absorption. Increasing the absorption of Mn⁴⁺ by increasing the dopant concentration and controlling the + 4 valence state will increase the EQE as this is crucial for optimal performance in LED applications. Since the initial Mn concentrations during the synthesis are not corresponding to the final dopant concentration, it is important to measure the dopant concentration using sensitive methods like XPS, EDX or ICP-MS.

The emission wavelength of K₂SiF₆:Mn⁴⁺ coincides with the ideal red color point at 630 nm for display applications. High color gamuts (up to 127% of the NTSC) can be reached when K₂SiF₆:Mn⁴⁺ is combined with green perovskite quantum dots. In lighting applications, K₂SiF₆:Mn⁴⁺ can be combined with YAG:Ce or other LED phosphors such as LuAG:Ce, SrSi₂N₂O₂:Eu or La₃Si₆N₁₁:Ce to form warm-white LEDs. Efficacies up to 137 lm/W and CRI values up to 95 have been reported. Depending on the phosphor mixture, a wide range of color temperatures can be reached. This shows the potential for K₂SiF₆:Mn⁴⁺ as a LED phosphor for both lighting and displays, when long-term stability and high quantum efficiencies are carefully mastered.

The luminescence of K₂SiF₆:Mn⁴⁺ is stable up to the application temperatures in LEDs. In case of high-brightness applications, a remote-phosphor setup can prevent saturation effects at high photon fluxes, related to the long lifetime of the Mn⁴⁺ excited state. Still, the unsatisfactory stability at high temperature and high humidity conditions is an issue to be solved for long-term, reliable operation. Several applications of coatings and annealing have been tested and seem to improve the stability of the phosphor material. Further improvement of the coatings and annealing in possible post-synthesis treatments will broaden the application range of K₂SiF₆:Mn⁴⁺ in pc-WLEDs with a long lifetime.

Appendix

The reported phosphor properties are represented in Table 6–Table 9, ordered according to the different synthesis methods.

Table 6. Phosphor properties resulting from etching synthesis

View Table | View all tables in this article

Table 7. Phosphor properties resulting from coprecipitation synthesis

View Table | View all tables in this article

Table 8. Phosphor properties resulting from precipitation synthesis

View Table | View all tables in this article

Table 9. Phosphor properties resulting from other synthesis methods

View Table | View all tables in this article

Funding

Innovation by Science and Technology (IWT) (SBO130030).

References and links

1. A. B. Nobel Media, “The 2014 Nobel Prize in Physics - Press Release,” http://www.nobelprize.org/nobel_prizes/physics/laureates/2014/press.html.

2. S. Abe, J. J. Joos, L. I. Martin, Z. Hens, and P. F. Smet, “Hybrid remote quantum dot/powder phosphor designs for display backlights,” Light Sci. Appl. 6(6), e16271 (2017). [CrossRef]  

3. P. F. Smet, A. B. Parmentier, and D. Poelman, “Selecting conversion phosphors for white light-emitting diodes,” J. Electrochem. Soc. 158(6), R37–R54 (2011). [CrossRef]  

4. H. Wu, X. Zhang, C. Guo, J. Xu, M. Wu, and Q. Su, “Three-band white light from InGaN-based blue LED chip precoated with green/red phosphors,” IEEE Photonics Technol. Lett. 17(6), 1160–1162 (2005). [CrossRef]  

5. J. J. Joos, K. W. Meert, A. B. Parmentier, D. Poelman, and P. F. Smet, “Thermal quenching and luminescence lifetime of saturated green Sr_1−xEu_xGa₂S₄ phosphors,” Opt. Mater. 34(11), 1902–1907 (2012). [CrossRef]  

6. R. Mueller-Mach, G. Mueller, M. R. Krames, H. A. Höppe, F. Stadler, W. Schnick, T. Juestel, and P. Schmidt, “Highly efficient all-nitride phosphor-converted white light emitting diode,” Phys. Status Solidi 202(9), 1727–1732 (2005). [CrossRef]  

7. T. Horikawa, X. Q. Piao, M. Fujitani, H. Hanzawa, and K. Machida, “Preparation of Sr₂Si₅N₈:Eu²⁺ phosphors using various novel reducing agents and their luminescent properties,” IOP Conf. Ser. Mater. Sci. Eng. 1(1), 12024 (2009). [CrossRef]  

8. Y. Hu, W. Zhuang, H. Ye, S. Zhang, Y. Fang, and X. Huang, “Preparation and luminescent properties of (Ca_1-x,Sr_x)S:Eu²⁺ red-emitting phosphor for white LED,” J. Lumin. 111(3), 139–145 (2005). [CrossRef]  

9. K. Uheda, N. Hirosaki, Y. Yamamoto, A. Naito, T. Nakajima, and H. Yamamoto, “Luminescence Properties of a Red Phosphor, CaAlSiN₃:Eu²⁺, for White Light-Emitting Diodes,” Electrochem. Solid-State Lett. 9(4), H22 (2006). [CrossRef]  

10. P. Pust, V. Weiler, C. Hecht, A. Tücks, A. S. Wochnik, A.-K. Henß, D. Wiechert, C. Scheu, P. J. Schmidt, and W. Schnick, “Narrow-band red-emitting Sr[LiAl₃N₄]:Eu²⁺ as a next-generation LED-phosphor material,” Nat. Mater. 13(9), 891–896 (2014). [CrossRef]   [PubMed]  

11. C. J. Duan, A. C. A. Delsing, and H. T. Hintzen, “Photoluminescence Properties of Novel Red-Emitting Mn²⁺-Activated MZnOS (M = Ca, Ba) Phosphors,” Chem. Mater. 21(6), 1010–1016 (2009). [CrossRef]  

12. J. J. Joos, K. Lejaeghere, K. Korthout, A. Feng, D. Poelman, and P. F. Smet, “Charge transfer induced energy storage in CaZnOS:Mn - insight from experimental and computational spectroscopy,” Phys. Chem. Chem. Phys. 19(13), 9075–9085 (2017). [CrossRef]   [PubMed]  

13. A. A. Setlur, E. V. Radkov, C. S. Henderson, J.-H. Her, A. M. Srivastava, N. Karkada, M. S. Kishore, N. P. Kumar, D. Aesram, A. R. Deshpande, B. Kolodin, S. G. Ljudmil, and U. Happek, “Energy-efficient, high-color-rendering LED lamps using oxyfluoride and fluoride phosphors,” Chem. Mater. 22(13), 4076–4082 (2010). [CrossRef]  

14. H. F. Sijbom, J. J. Joos, L. I. D. J. Martin, K. Van den Eeckhout, D. Poelman, and P. F. Smet, “Luminescent behavior of the K₂SiF₆:Mn⁴⁺ red phosphor at high fluxes and at the microscopic level,” ECS J. Solid State Sci. Technol. 5(1), R3040–R3048 (2016). [CrossRef]  

15. W. Davis and Y. Ohno, “Color quality scale,” Opt. Eng. 49(3), 033602 (2010). [CrossRef]  

16. D. Nickerson and C. W. Jerome, “Color rendering of light sources: CIE method of specification and its application,” Illum. Eng. 60(4), 262 (1965).

17. Radiocommunication Sector of ITU (ITU-R), Parameter Values for Ultra-High Definition Television Systems for Production and International Programme Exchange (2015).

18. J. J. Joos, J. Botterman, and P. F. Smet, “Evaluating the use of blue phosphors in white LEDs: the case of Sr_0.25Ba_0.75Si₂O₂N₂:Eu²⁺,” J. Solid State Light. 1(1), 6 (2014). [CrossRef]  

19. Y. Q. Li, J. E. J. van Steen, J. W. H. van Krevel, G. Botty, A. C. A. Delsing, F. J. DiSalvo, G. de With, and H. T. Hintzen, “Luminescence properties of red-emitting M₂Si₅N₈:Eu²⁺ (M=Ca, Sr, Ba) LED conversion phosphors,” J. Alloys Compd. 417(1–2), 273–279 (2006). [CrossRef]  

20. X. Piao, T. Horikawa, H. Hanzawa, and K. Machida, “Characterization and luminescence properties of Sr₂Si₅N₈:Eu²⁺ phosphor for white light-emitting-diode illumination,” Appl. Phys. Lett. 88(16), 2004–2007 (2006). [CrossRef]  

21. S. E. Brinkley, N. Pfaff, K. A. Denault, Z. Zhang, H. T. Hintzen, R. Seshadri, S. Nakamura, and S. P. Denbaars, “Robust thermal performance of Sr₂Si₅N₈:Eu²⁺: An efficient red emitting phosphor for light emitting diode based white lighting,” Appl. Phys. Lett. 99, 241106 (2011). [CrossRef]  

22. X. Piao, K. Machida, T. Horikawa, H. Hanzawa, Y. Shimomura, and N. Kijima, “Preparation of CaAlSiN₃:Eu²⁺ phosphors by the self-propagating high-temperature synthesis and their luminescent properties,” Chem. Mater. 19(18), 4592–4599 (2007). [CrossRef]  

23. J. H. Oh, H. Kang, Y. J. Eo, H. K. Park, and Y. R. Do, “Synthesis of narrow-band red-emitting K₂SiF₆:Mn⁴⁺ phosphors for a deep red monochromatic LED and ultrahigh color quality warm-white LEDs,” J. Mater. Chem. C 3(3), 607–615 (2015). [CrossRef]  

24. J. C. Zhang, L. Z. Zhao, Y. Z. Long, H. Di Zhang, B. Sun, W. P. Han, X. Yan, and X. Wang, “Color manipulation of intense multiluminescence from CaZnOS:Mn²⁺ by Mn²⁺ concentration effect,” Chem. Mater. 27(21), 7481–7489 (2015). [CrossRef]  

25. S. H. Park, K. H. Lee, S. Unithrattil, H. S. Yoon, H. G. Jang, and W. Bin Im, “Melilite-structure CaYAl₃O₇:Eu³⁺ phosphor: structural and optical characteristics for near-UV LED-based white light,” J. Phys. Chem. C 116(51), 26850–26856 (2012). [CrossRef]  

26. M. Meneghini, M. Dal Lago, N. Trivellin, G. Meneghesso, and E. Zanoni, “Thermally activated degradation of remote phosphors for application in LED lighting,” IEEE Trans. Device Mater. Reliab. 13(1), 316–318 (2013). [CrossRef]  

27. J. Li, J. Yan, D. Wen, W. U. Khan, J. Shi, M. Wu, Q. Su, and P. A. Tanner, “Advanced red phosphors for white light-emitting diodes,” J. Mater. Chem. C Mater. Opt. Electron. Devices 4(37), 8611–8623 (2016). [CrossRef]  

28. D. Chen, Y. Zhou, and J. Zhong, “A review on Mn⁴⁺ activators in solids for warm white light-emitting diodes,” RSC Advances 6(89), 86285–86296 (2016). [CrossRef]  

29. Z. Zhou, N. Zhou, M. Xia, M. Yokoyama, and H. T. Hintzen, “Research progress and application prospects of transition metal Mn⁴⁺-activated luminescent materials,” J. Mater. Chem. C Mater. Opt. Electron. Devices 4(39), 9143–9161 (2016). [CrossRef]  

30. H.-D. Nguyen and R.-S. Liu, “Narrow-band red-emitting Mn⁴⁺-doped hexafluoride phosphors: synthesis, optoelectronic properties, and applications in white light-emitting diodes,” J. Mater. Chem. C Mater. Opt. Electron. Devices 4(46), 10759–10775 (2016). [CrossRef]  

31. A. Paulusz, “Efficient Mn (IV) emission in fluorine coordination,” J. Electrochem. Soc. 120(7), 942–947 (1973). [CrossRef]  

32. J. H. Oh, Y. J. Eo, H. C. Yoon, Y.-D. Huh, and Y. R. Do, “Evaluation of new color metrics: guidelines for developing narrow-band red phosphors for WLEDs,” J. Mater. Chem. C Mater. Opt. Electron. Devices 4(36), 8326–8348 (2016). [CrossRef]  

33. N. M. Avram and M. G. Brik, eds., Optical Properties of 3d-Ions in Crystals: Spectroscopy and Crystal Field Analysis (Springer, 2013).

34. T. Takahashi and S. Adachi, “Mn⁴⁺-activated red photoluminescence K₂SiF₆ phosphor,” J. Electrochem. Soc. 155(12), E183–E188 (2008). [CrossRef]  

35. M. G. Brik and A. M. Srivastava, “On the optical properties of the Mn⁴⁺ ion in solids,” J. Lumin. 133(2013), 63–72 (2013).

36. M. G. Brik, S. J. Camardello, and A. M. Srivastava, “Influence of covalency on the Mn^{4+ 2}E_g -> ⁴A_2g emission energy in crystals,” ECS J. Solid State Sci. Technol. 4(3), R39–R43 (2014). [CrossRef]  

37. B. Henderson and G. Imbusch, Optical Spectroscopy of Inorganic Solids (Oxford University Press, 1989).

38. Z. Bryknar, V. Trepakov, Z. Potůček, and L. Jastrabík, “Luminescence spectra of SrTiO₃:Mn⁴⁺,” J. Lumin. 87, 605–607 (2000). [CrossRef]  

39. G. Kemeny and C. H. Haake, “Activator center in magnesium fluorogermanate phosphors,” J. Chem. Phys. 33(3), 783–789 (1960). [CrossRef]  

40. Y. K. Xu and S. Adachi, “Properties of Na₂SiF₆:Mn⁴⁺ and Na₂GeF₆:Mn⁴⁺ red phosphors synthesized by wet chemical etching,” J. Appl. Phys. 105(1), 013525 (2009). [CrossRef]  

41. P. H. M. Uylings, A. J. J. Raassen, and J. F. Wyart, “Energies of N equivalent electrons expressed in terms of two-electron energies and independent three-electron parameters : a new complete set of orthogonal operators : II. Application to 3dN configurations,” J. Phys. B At. Mol. Phys. 17(20), 4103–4126 (1984). [CrossRef]  

42. C. E. Schäffer and C. K. Jørgensen, “The nephelauxetic series of ligands corresponding to increasing tendency of partly covalent bonding,” J. Inorg. Nucl. Chem. 8, 143–148 (1958). [CrossRef]  

43. C. K. Jørgensen, Modern Aspects of Ligand Field Theory (North-Holland Pub. Co, 1971).

44. B. Han, B. W. Wessels, and M. P. Ulmer, “Optical investigation of electronic states of Mn⁴⁺ ions in p-type GaN,” Appl. Phys. Lett. 86(4),042505 (2005). [CrossRef]  

45. F. Karel, J. Pastrňák, and J. Rosa, “Thermoluminescent measurements of the trap level spectrum in AlN:Mn⁴⁺ phosphors,” Czech. J. Phys. 19(8), 974–982 (1969). [CrossRef]  

46. V. Baslon, J. P. Harris, C. Reber, H. E. Colmer, T. A. Jackson, A. P. Forshaw, J. M. Smith, R. A. Kinney, and J. Telser, “Near-infrared ²E_g → ⁴A_2g and visible LMCT luminescence from a molecular bis-(tris(carbene)borate) manganese(IV) complex,” Can. J. Chem. 95(5), 547–552 (2017). [CrossRef]  

47. M. G. Brik, S. J. Camardello, A. M. Srivastava, N. M. Avram, and A. Suchocki, “Spin-forbidden transitions in the spectra of transition metal ions and nephelauxetic effect,” ECS J. Solid State Sci. Technol. 5(1), R3067–R3077 (2016). [CrossRef]  

48. M. G. Brik and A. M. Srivastava, “Ab initio studies of the structural, electronic, and optical properties of K₂SiF₆ single crystals at ambient and elevated hydrostatic pressure,” J. Electrochem. Soc. 159(6), J212–J216 (2012). [CrossRef]  

49. R. Kasa and S. Adachi, “Red and deep red emissions from cubic K₂SiF₆:Mn⁴⁺ and hexagonal K₂MnF₆ synthesized in HF/KMnO₄/KHF₂/Si solutions,” J. Electrochem. Soc. 159(4), J89–J95 (2012). [CrossRef]  

50. R. Hoshino, T. Nakamura, and S. Adachi, “Synthesis and photoluminescence properties of BaSnF₆:Mn⁴⁺ red phosphor,” ECS J. Solid State Sci. Technol. 5(3), R37–R43 (2016). [CrossRef]  

51. L. Wei, C. C. Lin, M. Fang, M. G. Brik, S.-F. Hu, H. Jiao, and R. Liu, “A low-temperature co-precipitation approach to synthesize fluoride phosphors K₂MF₆:Mn⁴⁺ (M = Ge, Si) for white LED applications,” J. Mater. Chem. C Mater. Opt. Electron. Devices 3(8), 1655–1660 (2015). [CrossRef]  

52. S. Adachi and T. Takahashi, “Direct synthesis of K₂SiF₆:Mn⁴⁺ red phosphor from crushed quartz schist by wet chemical etching,” Electrochem. Solid-State Lett. 12(2), J20–J23 (2009). [CrossRef]  

53. T. Arai and S. Adachi, “Excited states of 3d³ electrons in K₂SiF₆:Mn⁴⁺ red phosphor studied by photoluminescence excitation spectroscopy,” Jpn. J. Appl. Phys. 50(9), 092401 (2011). [CrossRef]  

54. M. Novita and K. Ogasawara, “Comparative study of multiplet structures of Mn⁴⁺ in K₂SiF₆, K₂GeF₆, and K₂TiF₆ based on first-principles configuration–interaction calculations,” Jpn. J. Appl. Phys. 51, 22604 (2012).

55. M. H. Du, “Chemical trends of Mn⁴⁺ emission in solids,” J. Mater. Chem. C Mater. Opt. Electron. Devices 2(14), 2475–2481 (2014). [CrossRef]  

56. Q. Shao, L. Wang, L. Song, Y. Dong, C. Liang, J. He, and J. Jiang, “Temperature dependence of photoluminescence spectra and dynamics of the red-emitting K₂SiF₆:Mn⁴⁺ phosphor,” J. Alloys Compd. 695, 221–226 (2017). [CrossRef]  

57. A. Lazarowska, S. Mahlik, M. Grinberg, C. C. Lin, and R.-S. Liu, “Pressure effect on the zero-phonon line emission of Mn⁴⁺ in K₂SiF_6.,” J. Chem. Phys. 143(13), 134704 (2015). [CrossRef]   [PubMed]  

58. S. Adachi, H. Abe, R. Kasa, and T. Arai, “Synthesis and properties of hetero-dialkaline hexafluorosilicate red phosphor KNaSiF₆:Mn⁴⁺,” J. Electrochem. Soc. 159(2), J34–J37 (2012). [CrossRef]  

59. Y. K. Xu and S. Adachi, “Photoluminescence and raman scattering spectra in (NH₄)₂XF₆:Mn⁴⁺ (X = Si, Ge, Sn, and Ti) red phosphors,” J. Electrochem. Soc. 159(1), E11–E17 (2012). [CrossRef]  

60. Y. K. Xu and S. Adachi, “Properties of Mn⁴⁺-activated hexafluorotitanate phosphors,” J. Electrochem. Soc. 158(3), J58–J65 (2011). [CrossRef]  

61. F. Tang, Z. Su, H. Ye, M. Wang, X. Lan, D. L. Phillips, Y. Cao, and S. Xu, “A set of manganese ion activated fluoride phosphors (A₂BF₆:Mn⁴⁺, A = K, Na, B = Si, Ge, Ti): synthesis below 0 °C and efficient room-temperature photoluminescence,” J. Mater. Chem. C Mater. Opt. Electron. Devices 4(40), 9561–9568 (2016). [CrossRef]  

62. T. Takahashi and S. Adachi, “Synthesis of K₂SiF₆:Mn⁴⁺ red phosphor from silica glasses by wet chemical etching in HF/KMnO₄ solution,” Electrochem. Solid-State Lett. 12(8), J69–J71 (2009). [CrossRef]  

63. H.-D. Nguyen, C. C. Lin, and R.-S. Liu, “Waterproof alkyl phosphate coated fluoride phosphors for optoelectronic materials,” Angew. Chem. Int. Ed. Engl. 54(37), 10862–10866 (2015). [CrossRef]   [PubMed]  

64. M. Novita, T. Honma, B. Hong, A. Ohishi, and K. Ogasawara, “Study of multiplet structures of Mn⁴⁺ activated in fluoride crystals,” J. Lumin. 169, 594–600 (2016). [CrossRef]  

65. K. Ogasawara, T. Ishii, I. Tanaka, and H. Adachi, “Calculation of multiplet structures of Cr³⁺ and V³⁺ in α-Al₂O₃ based on a hybrid method of density-functional theory and the configuration interaction,” Phys. Rev. B 61(1), 143–161 (2000). [CrossRef]  

66. F. Yuan and H. Ryu, “Ce-doped YAG phosphor powders prepared by co-precipitation and heterogeneous precipitation,” Mater. Sci. Eng. B Solid-State Mater. Adv. Technol. 107(1), 14–18 (2004).

67. Y. N. Wu, C. H. Cheng, and Z. X. Xiong, “Enhancement of Blue Emission via Upconversion in Solid-State Synthesized Hexagonal NaYF₄:Ln³⁺,” Key Eng. Mater. 368–372, 398–401 (2008). [CrossRef]  

68. W. Massa, “Fluorides: solid-state chemistry,” in Encyclopedia of Inorganic Chemistry (John Wiley & Sons, Ltd, 2006).

69. S. Adachi and T. Takahashi, “Direct synthesis and properties of K₂SiF₆:Mn⁴⁺ phosphor by wet chemical etching of Si wafer,” J. Appl. Phys. 104(2), 023512 (2008). [CrossRef]  

70. T. Nakamura, Z. Yuan, and S. Adachi, “Micronization of red-emitting K₂SiF₆:Mn⁴⁺ phosphor by pulsed laser irradiation in liquid,” Appl. Surf. Sci. 320(2014), 514–518 (2014). [CrossRef]  

71. L. Lv, X. Jiang, S. Huang, X. Chen, and Y. Pan, “The formation mechanism, improved photoluminescence and LED applications of red phosphor K₂SiF₆:Mn⁴⁺,” J. Mater. Chem. C Mater. Opt. Electron. Devices 2(20), 3879–3884 (2014). [CrossRef]  

72. G. M. Rao, “Electrowinning of silicon from K₂SiF₆-molten fluoride systems,” J. Electrochem. Soc. Electrochem. Sci. Technol. 127(9), 1940–1944 (1980). [CrossRef]  

73. H. Bode, H. Jenssen, and F. Bandte, “Über eine neue Darstellung des Kalium-hexafluoromanganats(IV),” Angew. Chem. 65(11), 304 (1953). [CrossRef]  

74. E. V. Radkov, L. S. Grigorov, A. A. Setlur, and A. M. Srivastava, “Red line emitting phosphors for use in LED applications,” U.S. patent 7,497,973 B2 (2009).

75. H. Zhu, C. C. Lin, W. Luo, S. Shu, Z. Liu, Y. Liu, J. Kong, E. Ma, Y. Cao, R.-S. Liu, and X. Chen, “Highly efficient non-rare-earth red emitting phosphor for warm white light-emitting diodes,” Nat. Commun. 5, 4312 (2014). [CrossRef]   [PubMed]  

76. Y. Jin, M.-H. Fang, M. Grinberg, S. Mahlik, T. Lesniewski, M. G. Brik, G.-Y. Luo, J. G. Lin, and R.-S. Liu, “Narrow red emission band fluoride phosphor KNaSiF₆:Mn⁴⁺ for warm white light-emitting diodes,” ACS Appl. Mater. Interfaces 8(18), 11194–11203 (2016). [CrossRef]   [PubMed]  

77. M. Kim, W. B. Park, C. H. Kim, K.-S. Sohn, and B. K. Bang, “Radiative and non-radiative decay rate of K₂SiF₆:Mn⁴⁺ phosphors,” J. Mater. Chem. C Mater. Opt. Electron. Devices 4(8), 1166–1169 (2015).

78. L. Wang, X. Wang, T. Kohsei, K. Yoshimura, M. Izumi, N. Hirosaki, and R.-J. Xie, “Highly efficient narrow-band green and red phosphors enabling wider color-gamut LED backlight for more brilliant displays,” Opt. Express 23(22), 28707–28717 (2015). [CrossRef]   [PubMed]  

79. R. J. Lyons, A. A. Setlur, A. R. Deshpande, and L. S. Grigorov, “Color stable manganese-doped phosphors,” U.S. patent 8,252,613 B1 (2012).

80. M. Kaneyoshi and F. Echizen-shi, “Red phosphor,” U.S. patent 3 015 529 A1 (2015).

81. M. Kaneyoshi and Y. Takai, “Preparation of complex fluoride and complex fluoride phosphor,” U.S. patent 8,974,696 B2 (2015).

82. J. E. Murphy, W. W. Beers, R. J. Lyons, and F. Du, “Processes for preparing color stable manganese-doped complex fluoride phosphors,” U.S. patent 2015/0166887 A1 (2015).

83. J. E. Murphy, “Processes for preparing color stable manganese-doped phosphors,” U.S. patent 2015/0054400 A1 (2015).

84. T. Yoshida, “Fluoride phosphor and light emitting device using the same and method of manufacturing the fluoride phosphor,” U.S. patent 9,120,972 B2 (2015).

85. J. W. Park, I. H. Lee, T. H. Kim, J. H. You, C. W. Lee, and C. S. Yoon, “Fluoride phosphor and light emitting device, and methods of manufacturing the same,” U.S. patent 2016/133799 A1 (2016).

86. A. Setlur, M. Brewster, F. Garcia, M. C. Hill, R. Lyons, J. Murphy, T. Stecher, S. Stoklosa, S. Weaver, U. Happek, D. Aesram, and A. Deshpande, Optimized Phosphors for Warm White LED Light Engines (2012). [CrossRef]  

87. C. Liao, R. Cao, Z. Ma, Y. Li, G. Dong, K. N. Sharafudeen, and J. Qiu, “Synthesis of K₂SiF₆:Mn⁴⁺ phosphor from SiO₂ powders via redox reaction in HF/KMnO₄ solution and their application in warm-white LED,” J. Am. Ceram. Soc. 96(11), 3552–3556 (2013). [CrossRef]  

88. H. D. Nguyen, C. C. Lin, M. Fang, and R. S. Liu, “Synthesis of Na₂SiF₆:Mn⁴⁺ red phosphors for white LED applications by co-precipitation,” J. Mater. Chem. C Mater. Opt. Electron. Devices 2(48), 10268–10272 (2014). [CrossRef]  

89. M. J. Lee, Y. H. Song, Y. L. Song, G. S. Han, H. S. Jung, and D. H. Yoon, “Enhanced luminous efficiency of deep red emitting K₂SiF₆:Mn⁴⁺ phosphor dependent on KF ratio for warm-white LED,” J. Mater. Chem. C Mater. Lett. 141(2015), 27–30 (2015). [CrossRef]  

90. J. H. Oh, H. Kang, M. Ko, and Y. R. Do, “Analysis of wide color gamut of green/red bilayered freestanding phosphor film-capped white LEDs for LCD backlight,” Opt. Express 23(15), A791–A804 (2015). [CrossRef]   [PubMed]  

91. Q. Zhou, Y. Zhou, Z. Wang, Y. Liu, G. Chen, J. Peng, J. Yan, and M. Wu, “Fabrication and application of non-rare earth red phosphors for warm white-light-emitting diodes,” RSC Advances 5(103), 84821–84826 (2015). [CrossRef]  

92. B.-E. Yeo, Y.-S. Cho, and Y.-D. Huh, “Synthesis and photoluminescence properties of a red-emitting phosphor, K₂SiF₆:Mn⁴⁺, for use in three-band white LED applications,” Opt. Mater. 51(2016), 50–55 (2016). [CrossRef]  

93. R. A. Talewar, C. P. Joshi, and S. V. Moharil, “Enhancing the photovoltaic performance of CdTe/CdS solar cell via luminescent downshifting using K₂SiF₆:Mn⁴⁺ phosphors,” in AIP Conference Proceedings (2016), p. 140019.

94. X. Li, X. Su, P. Liu, J. Liu, Z. Yao, J. Chen, H. Yao, X. Yu, and M. Zhan, “Shape-controlled synthesis of phosphor K₂SiF₆:Mn⁴⁺ nanorods and their luminescence properties,” CrystEngComm 17(4), 930–936 (2015). [CrossRef]  

95. Z. Wang, Y. Zhou, Z. Yang, Y. Liu, H. Yang, H. Tan, Q. Zhang, and Q. Zhou, “Synthesis of K₂XF₆:Mn⁴⁺ (X=Ti, Si and Ge) red phosphors for white LED applications with low-concentration of HF,” Opt. Mater. 49(2015), 235–240 (2015). [CrossRef]  

96. L. Huang, Y. Zhu, X. Zhang, R. Zou, F. Pan, J. Wang, and M. Wu, “HF-free hydrothermal route for synthesis of highly efficient narrow-band red emitting phosphor K₂Si_1–xF₆:xMn⁴⁺ for warm white light-emitting diodes,” Chem. Mater. 28(5), 1495–1502 (2016). [CrossRef]  

97. J. E. Murphy, F. Garcia-Santamaria, A. A. Setlur, and S. Sista, “PFS, K₂SiF₆:Mn⁴⁺: the red-line emitting LED phosphor behind GE’s TriGain Technology^TM platform,” SID Symp. Dig. Tech. Pap. 46(1), 927–930 (2015). [CrossRef]  

98. J. W. Moon, B. G. Min, J. S. Kim, M. S. Jang, K. M. Ok, K.-Y. Han, and J. S. Yoo, “Optical characteristics and longevity of the line-emitting K₂SiF₆:Mn⁴⁺ phosphor for LED application,” Opt. Mater. Express 6(3), 782–792 (2016). [CrossRef]  

99. A. A. Setlur, R. J. Lyons, J. E. Murphy, N. Prasanth Kumar, and M. Satya Kishore, “Blue light-emitting diode phosphors based upon oxide, oxyhalide and halide hosts,” ECS J. Solid State Sci. Technol. 2(2), R3059–R3070 (2012). [CrossRef]  

100. P. Arunkumar, Y. H. Kim, H. J. Kim, S. Unithrattil, and W. B. Im, “Hydrophobic organic skin as a protective shield for moisture-sensitive phosphor-based optoelectronic devices,” ACS Appl. Mater. Interfaces 9(8), 7232–7240 (2017). [CrossRef]   [PubMed]  

101. J. Olchowka, M. Suta, and C. Wickleder, “Green synthesis of small nanoparticles of A₂SiF₆ (A = Li - Cs) using Ionic Liquids as Solvent and Fluorine Source: A novel simple approach without HF,” Chemistry6, (2017).

102. J. H. You, I. H. Lee, T. H. Kim, J. W. Park, C. S. Yoon, and C. W. Lee, “Fluoride phosphor composite method of manufacturing fluoride phosphor composite, white light emitting apparatus, display apparatus, lighting device, and electronic device,” U.S. patent 2016/0160122 A1 (2016).

103. A. A. Setlur, R. J. Lyons, P. K. Nammalwar, R. Hanumanlha, J. E. Murphy, and F. Garcia, “Moisture-resistant phosphor compositions and associate methods,” U.S. patent 2015/1025876 A1 (2015).

104. M. Kaneyoshi and M. Ishii, “Phosphor surface treatment method,” U.S. patent 2015/0329770 A1 (2015).

105. J. E. Murphy, A. A. Setlur, F. Garcia, R. J. Lyons, A. I. Chowdhury, N. Karkada, and P. K. Nammalwar, “Color stable red-emitting phosphors,” U.S. patent 8,906,724 B2 (2014).

106. M. Kaneyoshi and K. Wataya, “Method for treating Mn-activated complex fluoride phosphor,” U.S. patent 2016/0122633 A1 (2016).

107. R. D. Shannon, “Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides,” Acta Crystallogr. A 32(5), 751–767 (1976). [CrossRef]  

108. J. H. Loehlin, “Redetermination of the structure of potassium hexafluorosilicate, K₂SiF₆,” Acta Crystallogr. Sect. C Cryst. Struct. Commun. 40(3), 570 (1984). [CrossRef]  

109. P. Bukovec and R. Hoppe, “Zur kenntnis von hexagonalem K₂[MnF₆],” J. Fluor. Chem. 23(6), 579–587 (1983). [CrossRef]  

110. R. Hoshino and S. Adachi, “Optical spectroscopy and degradation behavior of ZnGeF₆·6H₂O:Mn⁴⁺ red-emitting phosphor,” J. Lumin. 162, 63–71 (2015). [CrossRef]  

111. J. Ha, S. Song, G. Kim, J. Sim, and S. Choi, “Light-emitting device,” U.S. patent 2016/0141457 A1 (2015).

112. Y. Ooyabu, H. Fujii, T. Nakamura, and H. Ito, “Component for light-emitting device, light-emitting device and producing method thereof,” U.S. patent 2012/0012875 A1 (2012).

113. Y. Ooyabu, H. Ito, and T. Naito, “Reflecting resin sheet, light emitting diode device and producing method thereof,” U.S. patent 2012/0262054 A1 (2012).

114. Y. Ooyabu, H. Katayama, S. Umetani, H. Ito, and S. Wakiya, “Light-emitting device assembly and lighting device,” U.S. patent 2013/0229805 A1 (2013).

115. H. Sakuta, Y. Satou, N. Kijima, and B. Hong, “White light-emitting device and illumination device,” U.S. patent 8,890,403 B2 (2014).

116. Y. Ooyabu, T. Nakamura, H. Fujii, and H. Ito, “Light-emitting device,” U.S. patent 8,916,893 B2 (2014).

117. Y. Ooyabu, T. Nakamura, H. Fujii, and H. Ito, “Phosphor ceramic and light-emitting device,” U.S. patent 8,664,678 B2 (2014).

118. H. Sakuta, Y. Kohara, and Y. Satou, “Semiconductor light-emitting device, semiconductor light-emitting system and illumination fixture,” U.S. patent 8,779,455 B2 (2014).

119. J. Meyer, V. Weiler, and P. J. Schmidt, “Lighting device with light source and wavelength converting element,” U.S. patent 9,175,214 B2 (2015).

120. Y. S. Park and H. J. Hahm, “Light emitting device, backlight unit and display apparatus,” U.S. patent 2015/0301259 A1 (2015).

121. V. Weiler, P. J. Schmidt, and H.-H. Bechtel, “LED-based device with wide color gamut,” U.S. patent 2014/068440 A1 (2014).

122. M. Mitani and H. Katayama, “Silicone resin sheet, producing method thereof, encapsulating sheet, and light emitting diode device,” U.S. patent 8,569,792 B2 (2013).

123. Z. Zhong, X. Wang, J. Zhang, H. Zhong, and J.-B. Han, “Optical detection of magnetic field with Mn⁴⁺:K₂SiF₆ phosphor from room to liquid helium temperatures,” Appl. Phys. Lett. 110(21), 212405 (2017). [CrossRef]  

124. X. Zhang, H.-C. Wang, A.-C. Tang, S.-Y. Lin, H.-C. Tong, C.-Y. Chen, Y.-C. Lee, T.-L. Tsai, and R.-S. Liu, “Robust and stable narrow-band green emitter: an option for advanced wide-color-gamut backlight display,” Chem. Mater. 28(23), 8493–8497 (2016). [CrossRef]  

125. M. D. Tessier, D. Dupont, K. De Nolf, J. De Roo, and Z. Hens, “Economic and Size-Tunable Synthesis of InP/ZnE (E = S, Se) Colloidal Quantum Dots,” Chem. Mater. 27(13), 4893–4898 (2015). [CrossRef]  

126. Y. Yang, Y. Zheng, W. Cao, A. Titov, J. Hyvonen, J. R. Manders, J. Xue, P. H. Holloway, and L. Qian, “High-efficiency light-emitting devices based on quantum dots with tailored nanostructures,” Nat. Photonics 9(March), 1–9 (2015).

127. D. Dupont, M. D. Tessier, P. F. Smet, and Z. Hens, “Indium Phosphide-Based Quantum Dots with Shell-Enhanced Absorption for Luminescent Down-Conversion,” Adv. Mater. 29(29), 1700686 (2017). [CrossRef]   [PubMed]  

128. W. Zhang, W. Yang, P. Zhong, S. Mei, G. Zhang, G. Chen, G. He, and R. Guo, “Spectral optimization of color temperature tunable white LEDs based on perovskite quantum dots for ultrahigh color rendition,” Opt. Mater. Express 7(9), 3065 (2017). [CrossRef]  

129. K. Yoshimura, H. Fukunaga, M. Izumi, M. Masuda, T. Uemura, K. Takahashi, R.-J. Xie, and N. Hirosaki, “White LEDs using the sharp β-sialon:Eu phosphor and Mn-doped red phosphor for wide-color gamut display applications,” J. Soc. Inf. Disp. 24(7), 449–453 (2016). [CrossRef]  

130. X. Zhang, M.-H. Fang, Y.-T. Tsai, A. Lazarowska, S. Mahlik, T. Lesniewski, M. Grinberg, W. K. Pang, F. Pan, C. Liang, W. Zhou, J. Wang, J.-F. Lee, B.-M. Cheng, T.-L. Hung, Y.-Y. Chen, and R.-S. Liu, “Controlling of Structural Ordering and Rigidity of β-SiAlON:Eu through Chemical Cosubstitution to Approach Narrow-Band-Emission for Light-Emitting Diodes Application,” Chem. Mater. Just Accepted Manuscript (2017).

131. F. Zhang, H. Zhong, C. Chen, X. G. Wu, X. Hu, H. Huang, J. Han, B. Zou, and Y. Dong, “Brightly luminescent and color-tunable colloidal CH₃NH₃PbX₃ (X = Br, I, Cl) quantum dots: potential alternatives for display technology,” ACS Nano 9(4), 4533–4542 (2015). [CrossRef]   [PubMed]  

132. G. Yang, Q. Fan, B. Chen, Q. Zhou, and H. Zhong, “Reprecipitation synthesis of luminescent CH₃NH₃PbBr₃/NaNO₃ nanocomposites with enhanced stability,” J. Mater. Chem. C Mater. Opt. Electron. Devices 4(48), 11387–11391 (2016). [CrossRef]  

133. H. Huang, Q. Xue, B. Chen, Y. Xiong, J. Schneider, C. Zhi, H. Zhong, and A. L. Rogach, “Top-Down Fabrication of Stable Methylammonium Lead Halide Perovskite Nanocrystals by Employing a Mixture of Ligands as Coordinating Solvents,” Angew. Chem. Int. Ed. Engl. 56(32), 9571–9576 (2017). [CrossRef]   [PubMed]  

134. S. Lou, T. Xuan, C. Yu, M. Cao, C. Xia, J. Wang, and H. Li, “Nanocomposites of CsPbBr₃ perovskite nanocrystals in an ammonium bromide framework with enhanced stability,” J. Mater. Chem. C Advance Article (2017). [CrossRef]  

135. F. Baur, F. Glocker, and T. Jüstel, “Photoluminescence and energy transfer rates and efficiencies in Eu³⁺ activated Tb₂Mo₃O₁₂,” J. Mater. Chem. C Mater. Opt. Electron. Devices 3(9), 2054–2064 (2015). [CrossRef]  

136. K. V. Ivanovskikh, J. M. Ogieglo, A. Zych, C. R. Ronda, and A. Meijerink, “Luminescence temperature quenching for Ce³⁺ and Pr³⁺ d-f emission in YAG and LuAG,” ECS J. Solid State Sci. Technol. 2(2), R3148–R3152 (2012). [CrossRef]  

137. D. Luo, L. Wang, S. W. Or, H. Zhang, and R.-J. Xie, “Realizing superior white LEDs with both high R9 and luminous efficacy by using dual red phosphors,” RSC Advances 7(42), 25964–25968 (2017). [CrossRef]  

138. P. F. Smet and J. J. Joos, “White light-emitting diodes: Stabilizing colour and intensity,” Nat. Mater. 16(5), 500–501 (2017). [CrossRef]   [PubMed]  

139. C. Hoelen, H. Borel, J. de Graaf, M. Keuper, M. Lankhorst, C. Mutter, L. Waumans, and R. Wegh, “Remote phosphor LED modules for general illumination – towards 200 lm/W general lighting LED light sources,” Proc. SPIE 7058, 70580M (2008). [CrossRef]  

140. P. Dorenbos, “Thermal quenching of Eu²⁺ 5d–4f luminescence in inorganic compounds,” J. Phys. Condens. Matter 17(50), 8103–8111 (2005). [CrossRef]