Highly Luminous N3--Substituted Li2MSiO4-δN2/3δ:Eu2+ (M = Ca, Sr, and Ba) for White NUV Light-Emitting Diodes.

The N3--substituted Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) phosphors were systematically prepared and analyzed. Secondary-ion mass spectroscopy measurements revealed that the average N3- contents are 0.003 for Ca, 0.009 for Sr, and 0.032 for Ba. Furthermore, the N3- incorporation in the host lattices was corroborated by infrared and X-ray photoelectron spectroscopies. From the photoluminescence spectra of Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) phosphors before and after N3- doping, it was verified that the enhanced emission intensity of the phosphors is most likely due to the N3- doping. In Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) phosphors, the maximum wavelengths of the emission band were red-shifted in the order Ca < Ba < Sr, which is not consistent with the trend of crystal field splitting: Ba < Sr < Ca. This discrepancy was clearly explained by electron-electron repulsions among polyhedra, LiO4-MO n , SiO4-MO n , and MO n -M'O n associated with structural difference in the host lattices. Therefore, the energy levels associated with the 4f65d energy levels of Eu2+ are definitely established in the following order: Li2CaSiO4:Eu2+ > Li2BaSiO4:Eu2+ > Li2SrSiO4:Eu2+. Furthermore, using the Williamson-Hall (W-H) method, the determined structural strains of Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) phosphors revealed that the increased compressive strain after N3- doping induces the enhanced emission intensity of these phosphors. White light-emitting diodes made by three N3--doped phosphors and a 365 nm emitting InGaN chip showed the (0.333, 0.373) color coordinate and high color-rendering index (R a = 83). These phosphor materials may provide a platform for development of new efficient phosphors in solid-state lighting field.

Introduction

Recently, many researchers have focused their attention on rare-earth-doped phosphor materials for solid-state lighting, especially in white light-emitting diodes (w-LEDs). Commercialized w-LEDs are made up of a blue-emitting LED chip and a yellow-emitting phosphor (Y3Al5O12:Ce3+ (YAG:Ce)). However, they exhibit a low color-rendering index (CRI) because they generate weak red emission.[1] To solve the problem, optimal phosphors should be newly developed. Therefore, new classes of phosphors based on metal silicates, metal sulfides, metal oxy-nitrides, and metal nitrides have been developed.[2−9] Liu et al. reported that Li2CaSiO4:Eu2+ (with a tetragonal phase, space group I4̅2m) has high absorption from the UV to the near-UV region and a strong emission at ∼480 nm with a narrow bandwidth.[9] Orange-yellow-emitting Li2SrSiO4:Eu2+ (with a trigonal crystal system and belonging to space group P3121) was prepared and compared to the commercialized YAG:Ce phosphor by Varadaraju et al.[10] Recently, the crystal structure and PL properties of Eu2+-doped Li2BaSiO4 have been elucidated and discussed by Kulshreshtha et al.[11] and Kim et al.,[12], i. e., the hexagonal crystal system (belonging to space group P63cm) and green emission at ∼508 nm, respectively. Based on the structures of the host materials in Li2MSiO4 (M = Ca, Sr, and Ba), it is presumed that the luminescent properties of these phosphor materials may be closely related to the chemical environment around the Eu2+-activator ion. It is well known that alkaline-earth silicon–oxynitride phosphors with the RE-ion activator (RE = Ce3+, Eu2+, Yb3+, etc.) have been intensively studied in recent years because they are potential candidate materials for improving the low luminous efficiency and low CRI for w-LEDs.[13−16] The N3– ions substituted partially for O2– ions in the host lattice could change the electronic structure of Eu2+ owing to their difference in electronegativity, ionic radius, and magnitude of crystal field splitting, as well as the nephelauxetic (cloud expanding) effect of the two ions,[3,17] which results in the change of the PL properties of the phosphors before and after N3– doping. Song et al. mentioned that the partial nitridation of Li2SrSiO4–3N:Eu2+ (x = 0.01) phosphors induced a highly enhanced PL intensity by as high as 190%.[18] Therefore, it is presumed that their results for the N3– doping effect on an increase of the luminescent intensity are meaningful even though the N3– contents doped into the crystal sites were not determined. However, systematic research for the N3– doping effect of Li2MSiO4 (M = Ca and Ba) has not been pursued to the best of our knowledge. Furthermore, the present study is also motivated by the desire to devise a white near-UV LED because Eu2+-activated Li2MSiO4 (M = Ca, Sr, and Ba) phosphors exhibit blue, orange-yellow, and green emission, respectively. Herein, we report on the highly enhanced luminescence of N3–-substituted Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) and a white near-UV LED using three phosphors.

Results and Discussion

Characterization of Crystal Structure

The crystal structure of each compound was characterized by its X-ray diffraction (XRD) pattern using the Rietveld method[19] with the FullProf program.[20] From the ICSD database, the equivalent isotropic displacement parameters and atomic coordinates were used. Using a pseudo-Voigt function with an asymmetry correction at low angles, the shape of the peak was fitted. The obtained patterns from Rietveld refinement of the XRD data are presented in Figure . The final values of the equivalent isotropic displacement parameters and atomic coordinates are given in Table S1 (and Tables S3 and S5). The selected bond distances are presented in Table S2 (and Tables S4 and S6). Figure shows the change in lattice volume before and after the nitridation of each Li2MSiO4 compound.

Figure 1

Rietveld refinement profile of the powder XRD data for LCSO:Eu, LSSO:Eu, and LBSO:Eu before and after N3– doping. The measured, fitted data, expected reflection positions, and difference between the measured and fitted data are depicted as black circles, red lines, green lines, and blue lines, respectively.

Figure 2

Lattice volume changes before and after N3– doping of each Li2MSiO4:Eu (M = Ca, Sr, and Ba) compound.

Lattice volume changes before and after N3– doping of each Li2MSiO4:Eu (M = Ca, n class="Chemical">Sr, and Ba) compound.

Determination of N3– Contents by Secondary-Ion Mass Spectroscopy

To determine the N3– contents of Li2CaSiO4−δN2/3δ (LCSON), Li2SrSiO4−δN2/3δ (LSSON), and Li2BaSiO4−δN2/3δ (LBSON), secondary-ion mass spectrometry (SIMS) analysis was used. It is well known that SIMS gives the best detection limit for identifying elements with very low concentrations compared to other techniques in surface analysis area. The elemental composition using SIMS can be quantitatively determined with an ion-implanted standard material.[3,21,22]Figure shows the atomic intensity and concentration versus the sputter depth of the compounds. For Li, Si, and O atoms among the three phosphors, no difference in the secondary-ion intensities was observed, while the intensities of secondary-ion for Ca, Sr, and Ba atoms changed because of the different sensitivities of the elements. The concentrations of N atom are analyzed as a function of sputter depth from the surface to the inner region of the grains (∼5000 nm) for the three samples. The contents of N3– ion were calculated as shown in Figure d. The average contents of N3– ion for LCSON, LSSON, and LBSON were 0.003, 0.009, and 0.032, respectively. It is remarkable that the N3–-ion content increases with larger alkaline-earth metal-ion size (Ba2+ > Sr2+ > Ca2+), probably implying that Li2MSiO4−δN2/3δ:Eu2+ (M = Ca, Sr, and Ba) phosphors with the larger ion size can easily accommodate more N3– ions in the crystal sites. The framework of the Li2MSiO4 (M = Ca, Sr, and Ba) structure is connected by SiO4 and LiO4 tetrahedra with M2+ ions, where two SiO4 and LiO4 tetrahedra are linked through sharing of oxygen atoms. Therefore, the bond lengths of MO polyhedra may be affected by the N3– content introduced into the host lattice. The average bond distances for the three compounds (see Tables S2, S4, and S6) were determined as 2.535 Å for LCSON, 2.625 Å for LSSON, and 2.877 Å for LBSON.

Figure 3

Atomic intensity and concentration versus sputter depth of LCSON (a), LSSON (b), LBSON (c), and N3– contents (d) by SIMS measurements.

Atomic intensity and concentration versus sputter depth of LCSON (a), LSSON (b), LBSON (c), and N3– contents (d) by n class="Chemical">SIMS measurements.

Evidence of N3– Incorporated in the Host Lattice from Infrared and X-ray Photoelectron Spectroscopy Measurements

The infrared (IR) spectra of Li2MSiO4 (M = Ca, Sr, and Ba) compounds before and after N3– doping are shown in Figure . As the chemical bond of Si–O is stronger than that of M–O (M = Li, Ca, Sr, and Ba), the internal vibrations of SiO4 tetrahedra are nearly independent of the lattice vibrations and exclusively observed in the 400–1000 cm–1 range. As shown in Figure , the [SiO4] internal modes are assigned to the Si–O stretching modes[23−25] between 770 and 950 cm–1 and to the O–Si–O bending modes[26,27] between 400 and 560 cm–1. Furthermore, based on the far-IR modes of MCO3 (M = Ca, Sr, and Ba) (Figure S1), the force constant (K) of M–O,[28] the reduced mass (μ) of M–O (μ = 11.43 for Ca–O, μ = 13.53 for Sr–O, μ = 14.33 for Ba–O), and the M–O stretching modes are estimated to be 314 cm–1 for Ca–O stretching, 257 cm–1 for Sr–O stretching, and 239 cm–1 for Ba–O stretching. Remarkably, the Si–O stretching modes between 770 and 950 cm–1 are somewhat decreased after N3– doping, which implies that the N3– ions are partially substituted into the Li2MSiO4 (M = Ca, Sr, and Ba) host lattice. The Gaussian-fitted IR modes of LSSO and LSSON between 1100 and 770 cm–1 (Figure ) corroborate the presence of N3– ions substituted into the host lattice, i.e., after N3– doping, the sub-band centered at 825 cm–1 shifts to the band with a maximum at 817 cm–1. This chemical shift can be easily explained by a comparison of the chemical bond distance (1.62 Å for SiO2 and 1.73 Å for α-Si3N4)[29,30] and the chemical bond energy (454 kJ/mol for SiO2 and 426 kJ/mol for α-Si3N4)[31] between SiO2 and α-Si3N4. X-ray photoelectron spectroscopy (XPS) results warrant the evidence of N3– doping, closely related to the red shift of the Si–O stretching modes (from IR measurement) in these compounds. Figure shows Si 2p (ref SiO2, ref α-Si3N4, LSSO, and LSSON) and N 1s (LSSO and LSSON) XPS binding energies. All XPS patterns were fitted after a Shirley background correction. As presented in Figure , the Si 2p binding energy (ref SiO2) is 102.8 eV with a single Gaussian band (Full width at half-maximum = 1.9 eV), whereas the Si 2p binding energies (ref α-Si3N4) is composed of two bands (101.4 and 102.8 eV). In the Si 2p binding energy (ref α-Si3N4), the sub-band at 102.8 eV may be due to superficial oxidation. The Gaussian-fitted LSSO and LSSON XPS images (middle) reveal that the Si 2p binding energy of LSSON (100.7 eV) is lower than that of LSSO (101.5 eV), which means that the N3– ions are partially substituted into the LSSO crystal lattice. Furthermore, the N 1s binding energy (393.4 eV) of LSSON corroborates the N3– introduction in the LSSO host lattice.

Figure 4

Fourier-transform infrared (FT-IR) spectra of the Li2MSiO4:Eu (M = Ca, Sr, and Ba) compound.

Figure 5

Gaussian-fitted IR bands of LSSO (a) and LSSON (b) between 1100 and 750 cm–1.

Figure 6

Si 2p and N 1s binding energies of SiO2, α-Si3N4, LSSO, and LSSON using XPS.

Fourier-transform infrared (FT-IR) spectra of the Li2MSiO4:Eu (M = Ca, n class="Chemical">Sr, and Ba) compound.

Gaussian-fitted IR n class="Chemical">bands of LSSO (a) and LSSON (b) between 1100 and 750 cm–1.

Si 2p and N 1s binding energies of n class="Chemical">SiO2, α-Si3N4, LSSO, and LSSON using XPS.

Photoluminescence Monitored under UV Light

Figure presents photoluminescence (PL) spectra of Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) before and after N3– doping. The excitation spectra of the three phosphors are somewhat different in shape, probably because of the distinct crystal system of host lattices: tetragonal for Li2CaSiO4, trigonal for Li2SrSiO4, and hexagonal for Li2BaSiO4. For Ca and Sr, the emission intensities are increased after N3– doping, by factors of 1.8 and 1.5, respectively. For Ba, there is no considerable increase even after N3– doping. The emission spectra exhibit broad bands with band maxima between 450 and 650 nm: 480 nm for Ca, 569 nm for Sr, and 509 nm for Ba. The 4f–5d transition energy of Eu2+ is greatly dependent on the local environment because the local structure around Eu2+-activator ion has a great influence on the centroid shift energy (CS) and crystal field splitting (CFS) of the 5d levels as well as the Stokes shift (ΔS) of the emission. Among the factors determining the 4f–5d transition energy of Eu2+, the center of gravity (barycenter) of the 5d levels is lowered relative to the free Eu2+ in the vacuum state, mainly ascribed to the covalent bonding character between Eu2+ and coordinating anions (O2–). Separately from a centroid shift, the neighboring anions have an additional influence on the 4f65d1 energy level of Eu2+, known as crystal field splitting (CFS). The magnitude of the CFS is dependent on the geometrical interaction between 5d orbitals and anion ligands: the energy of the 5d electron is raised because of the larger repulsion between the 5d electron in an orbital oriented toward an anion ligand and the electrons in the neighboring ligands, whereas that of the 5d electron in an orbital oriented away from an anion ligand is lowered. It should be mentioned that Pauling’s rule 3 based on the hard-spheres electrostatic model[32] states that the sharing of edge and face by two polyhedrons decreases the stability of an ionic structure because of the increase in cation–cation repulsions as the cations get close together. For a certain coordination geometry, the degree of cation–cation repulsions also increases in the following order: corner-shared < edge-shared < face-shared. Thus, Pauling’s rule 3 means that most ionic solids will prefer to be corner-shared rather than edge- or face-shared because of the bond stability. Notably, Morrison verified that the shift of the central energy of 4f5d configurations (for activator ion) is directly proportional to the summation of ligand polarizabilities over all nearest coordinating anion ligands.[33] Shi et al. reported that the barycenter energy of 4f5d configuration on Eu2+ ions is strongly dependent on the environmental factor (he)where N is the number of ligands, fc is the average fractional covalence, Q is the charge of the nearest anion, and α is the average bond volume polarizability.[34]

Figure 7

PL spectra of Li2MSiO4:Eu (M = Ca, Sr, and Ba) compound before and after N3– doping; M = Ca (a), M = Sr (b), M = Ba (c), relative emission intensity (d).

PL spectra of n class="Chemical">Li2MSiO4:Eu (M = Ca, Sr, and Ba) compound before and after N3– doping; M = Ca (a), M = Sr (b), M = Ba (c), relative emission intensity (d).

Therefore, the clear explanation for the wavelength shift of the emission band maxima in Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) requires reconsideration of the structural aspects of the compounds because the energy levels of Eu2+ are perturbed by neighbor cations and anions, such as O2–, Si4+, Li+, and M2+ (M = Ca, Sr, and Ba). Figure shows the bonding types between a MO polyhedron and SiO4 (or LiO4) tetrahedra in Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba). For example, if one assumes that the Eu2+ activator occupies the Ca1 site in a CaO8 polyhedron, the repulsion between Eu2+ and Si4+ (or Li+) is more prominent than that from the neighbor CaO8 polyhedra (×8) with corner-sharing through O atoms because of the difference in the bond distance between Ca–O and Si–O (or Li–O). As shown in Figure , a MO polyhedron is connected by SiO4 tetrahedra, LiO4 tetrahedra, and M’O polyhedra with shared corners, edges, and faces depending on the crystal structure in Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba). According to Pauling’s rule 3, in the only corner-shared polyhedra, the cations (or anions) are away from each other, while the distances between the cations (or anions) are getting shorter from edge-shared to face-shared, which finally results in the more enhanced electron–electron repulsions. Consequently, the energy levels of the Eu2+ activator stabilized in a MO polyhedron are perturbed and changed depending on the electron–electron repulsions between neighbor anions and Eu2+. Table presents the bonding type between an AO4 tetrahedron (A = Li and Si) and a MO polyhedron (M = Ca, Sr, and Ba). From the qualitative estimation of Table , the energy order of the 5d energy levels (Eu2+) can be determined as followsThe fact that the emission spectra of Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) (see Figure ) show variation of the band maxima (at 480 nm for Ca, 569 nm for Sr, and 509 nm for Ba) warrants that our interpretation of the emission band maxima from Ca to Ba in these compounds is reasonable. Figure presents the diffuse reflectance spectra (DRS) of Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) before and after N3– doping. It is evident that the absorption intensities between 250 and 500 nm of Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) are greatly increased compared to those of the host lattices of Li2MSiO4. Furthermore, for M = Ca and Sr, the absorption bands of the N3–-doped compounds are greatly intensified compared to those of Li2MSiO4:Eu2+, probably as a result of the N3– doping, which is in agreement with the PL results (see Figure ). From the DRS, the band gap energies (Eg) were determined using the Kubelka–Munk transformation, and these are presented in Figure . It should be noted that Eg of the Li2CaSiO4:Eu2+ compound was previously reported as 5.21 eV, using the Kohn–Sham density functional theory method.[35] In this study, we determined the following Eg values of Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba): 5.44 eV for Ca, 5.16 eV for Sr, and 5.34 eV for Ba. From band structure calculations,[36,37] the valence band of Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) is mainly formed by the O 2p states and the conduction band is predominantly made up of M 5d and Si 3p states. For Li2MSiO4 host lattices, the band gap energies progressively decrease along with Ca → Ba → Sr because the order of energy levels of M 5d orbitals is Li2CaSiO4 > Li2BaSiO4 > Li2SrSiO4, as mentioned in the energy order of the 5d energy levels (2).

Figure 8

Local structures of Li2MSiO4:Eu (M = Ca, Sr, and Ba) with a fixed M1O-polyhedron connected by neighbor polyhedra (SiO4-tetrahedra, LiO4-tetrahedra, and M’O-polyhedra); LCSO (a), LSSO (b), and LBSO (c). Blue, green, and yellow colors represent SiO4-tetrahedron, LiO4-tetrahedron, and MO-polyhedron, respectively.

Table 1

Bonding Characteristic of Li2MSiO4:Eu (M = Ca, Sr, and Ba) between a Fixed M1O-Polyhedron and Neighbor Polyhedra (SiO4-Tetrahedra, LiO4-Tetrahedra, and M’O-Polyhedra)a

	SiO4-tetrahedron		LiO4-tetrahedron		MOn-polyhedron
	bonding type	bond length (Å)	bonding type	bond length (Å)	bonding type	bond length (Å)
Ca1O8 in Li2CaSiO4	C-sharing (×4)	Ca1–Si = 3.566	E-sharing (×8)	Ca1–Li = 2.997	C-sharing (×8)	Ca1–Ca = 4.817
	E-sharing (×2)	Ca1–Si = 3.239
Sr1O8 in Li2SrSiO4	C-sharing (×4)	Sr1–Si = 3.731	C-sharing (×4)	Sr1–Li = 3.593	C-sharing (×8)	Sr1–Sr = 4.962
	E-sharing (×2)	Sr1–Si = 3.275	E-sharing (×2)	Sr1–Li = 3.192
			F-sharing (×2)	Sr1–Li = 2.532
Ba1O9 in Li2BaSiO4	C-sharing (×3)	Ba1–Si = 3.962	C-sharing (×5)	Ba1–Li = 4.189	C-sharing (×2)	Ba1–Ba = 6.011
	E-sharing (×3)	Ba1–Si = 3.585	E-sharing (×3)	Ba1–Li = 3.387	F-sharing (×4)	Ba1–Ba = 4.174
			F-sharing (×1)	Ba1–Li = 3.485

C, E, and F mean corner, edge, and face, respectively.

Figure 9

Diffuse reflectance spectra of LCSO:Eu (a), LSSO:Eu (b), and LBSO:Eu (c) before and after N3– doping.

Figure 10

Band gap energies determined using the Kubelka–Munk transformation from DRS of LCSON:Eu (a), LSSON:Eu (b), and LBSON:Eu (c).

Local structures of Li2MSiO4:Eu (M = Ca, n class="Chemical">Sr, and Ba) with a fixed M1O-polyhedron connected by neighbor polyhedra (SiO4-tetrahedra, LiO4-tetrahedra, and M’O-polyhedra); LCSO (a), LSSO (b), and LBSO (c). Blue, green, and yellow colors represent SiO4-tetrahedron, LiO4-tetrahedron, and MO-polyhedron, respectively.

Band gap energies determined un class="Chemical">sing the Kubelka–Munk transformation from DRS of LCSON:Eu (a), LSSON:Eu (b), and LBSON:Eu (c).

N3– Doping Effect on the PL Intensity: Williamson–Hall Plot

To precisely examine the N3– doping effect, the structural strain induced by the N3–-doping was estimated because the ionic radius of N3– (1.46 Å at CN = 4) is larger than that of O2– ion (1.38 Å at CN = 4). The strain can be determined using the Williamson–Hall (W–H) method from XRD profile analysis.[38] The total peak width at half-maximum intensity (β) is determined from the summation of the size broadening (βD) and the strain broadening (βs)The size broadening is correlated with the Scherrer equation: βD = kλ/(D cos θ), where D is the crystallite size, λ is the wavelength of the X-ray, and k is the shape factor. The strain broadening is expressed by ε = βs/4 tan θ, where ε is maximum strain (tensile or compressive). Thus, we get the W–H equation from eq The plot of β cos θ/λ versus 4 sin θ/λ can give crystallite size from the y-intercept and strain due to lattice deformation from the slope. The W–H plots present the stain (ε) of Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) before and after N3– doping (Figure ): for Li2CaSiO4:Eu2+, from −0.012% (before N3– doping, LCSO) to −0.026% (after N3– doping, LCSON); for Li2SrSiO4:Eu2+, from 0.0056% (before N3– doping, LSSO) to −0.0095% (after N3– doping, LSSON); and for Li2BaSiO4:Eu2+, from −0.035% (before N3– doping, LBSO) to −0.032% (after N3– doping, LBSON). The positive and negative values correspond to the tensile and compressive strains, respectively. As the temperature increases, generally, a crystal lattice undergoes tensile stress and expands. It is generally accepted that under the tensile stress, the lattice vibrations and/or the formation of defects force the phosphors more easily to promote nonradiative relaxation.[39−41] Thus, it is presumed that the enhancement of PL intensity is ascribed to the compensation of the thermally induced tensile stress by the compressive stress due to the N3– doping in Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) phosphors. The fact that the PL intensities of the two phosphors (for Ca and Sr) remarkably enhanced after N3– doping, whereas that of Li2BaSiO4:Eu2+ is not changed, corroborates the structural strain effect on the PL intensity (see Figure ). Furthermore, the structural strain effect by the N3– doping prominently appears in UV–visible absorbance spectra (see Figure ). Evidently, for Ca and Sr, the stronger absorption bands between 300 and 500 nm are present after N3– doping, while for Ba, the absorption bands are nearly same before and after N3– doping. Presumedly, the compressive strain plays an important role in the effective transfer of the absorbed energy from the host lattice to the activator ion even though the difference of strain values before and after N3– doping is very small, resulting in the enhanced absorption band and PL intensity.

Figure 11

Williamson–Hall plots of LCSO:Eu, LSSO:Eu, and LBSO:Eu before and after N3– doping.

Photoluminescence of LEDs

To examine the potential of the as-synthesized phosphors for near-UV LED application, phosphor-converted LEDs were made using a phosphor powder and a InGaN LED with 365 nm emission. Figure shows the emission spectra of Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) phosphors before and after N3– doping using an InGaN LED (λmax = 365 nm) under forward-bias currents from 10 to 50 mA. Figure clearly shows that the 365 nm UV light emitted from the InGaN chip is absorbed by the phosphors and simultaneously downconverted into intensive, wide-band-emitting light. Upon increasing the forward-bias current from 10 to 50 mA, the emission intensities of the phosphors progressively increase. At the same time, the luminous output increases and the shape and position of the LED emission bands exhibit similarity, thus confirming that the phosphors show a stable emission property. As presented in Figure , the maximum LED emission intensity for Ca and Sr increases after N3– doping, which is consistent with the PL spectra (see Figure ). For a white LED, the N3–-doped Li2MSiO4−δN2/3δ:Eu2+ (M = Ca, Sr, and Ba) phosphors were mixed and monitored using a color coordinate meter to determine the appropriate mixing proportions. The appropriate mixing proportions were determined as 46.7 wt % Ca, 46.6 wt % Sr, and 6.7 wt % Ba. Figure presents the CIE chromaticity of LEDs monitored under a forward-bias current of 50 mA. The N3–-doped Li2CaSiO4−δN2/3δ:Eu2+ emits blue light with the (0.090, 0.253) chromaticity coordinate. The N3–-doped Li2SrSiO4−δN2/3δ:Eu2+ emits yellow light with the (0.496, 0.495) chromaticity coordinate. The N3–-doped Li2BaSiO4−δN2/3δ:Eu2+ emits green light with the (0.241, 0.584) chromaticity coordinate. Finally, the white zone formed by mixing the light from the three phosphors corresponds to the (0.333, 0.373) chromaticity coordinate, a color-rendering index of Ra = 83, and correlated color temperature of 5480 K.

Figure 12

Photoluminescence of LEDs in the Li2MSiO4:Eu (M = Ca, Sr, and Ba) compound before and after N3– doping; LCSO:Eu (a), LSSO:Eu (b), LBSO:Eu (c), LCSON:Eu (d), LSSON:Eu (e), and LBSON:Eu (f).

Figure 13

CIE chromaticities of LCSON:Eu, LSSON:Eu, LBSON:Eu, and the mixed powder with three phosphors monitored under 365 nm UV light.

Photoluminescence of LEDs in the n class="Chemical">Li2MSiO4:Eu (M = Ca, Sr, and Ba) compound before and after N3– doping; LCSO:Eu (a), LSSO:Eu (b), LBSO:Eu (c), LCSON:Eu (d), LSSON:Eu (e), and LBSON:Eu (f).

CIE chromaticities of LCSON:Eu, LSSOn class="Chemical">N:Eu, LBSON:Eu, and the mixed powder with three phosphors monitored under 365 nm UV light.

Conclusions

In this work, Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) phosphors before and after N3– doping were systematically prepared and analyzed. SIMS measurement revealed that the average contents of N3– ion for LCSON, LSSON, and LBSON were 0.003, 0.009, and 0.032, respectively, probably implying that Li2MSiO4−δN2/3δ:Eu2+ (M = Ca, Sr, and Ba) phosphors with larger ion size can easily accommodate more N3– ions in the crystal lattice sites. Furthermore, the N3– sited in the host lattices was corroborated by IR and XPS analyses, i.e., the chemical shift of Si–O stretching mode from 825 to 817 cm–1 after N3– doping (for LSSON) and the chemical shift of Si 2p binding energy from 101.5 to 100.7 eV and the presence of the N 1s binding energy at 393.4 eV after N3– doping (for LSSON). From the PL spectra in Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) phosphors before and after N3– doping, it was verified that the enhanced emission intensity of the phosphors is most likely due to N3– doping. The change of the emission band maxima was also investigated in these phosphors. In Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) phosphors, the maximum wavelengths of the emission bands were red-shifted in the order Ca < Ba < Sr, which is not consistent with the trend of crystal field splitting: Ba < Sr < Ca. This discrepancy between the shift of emission band maxima and trend based on the crystal field splitting was clearly explained by electron–electron repulsions among polyhedra, LiO4–MO, SiO4–MO, and MO–M’O associated with structural difference in host lattices. Therefore, the energy levels associated with the 4f65d → 4f7 transition of Eu2+ are definitely established in the following order: Li2CaSiO4:Eu2+ > Li2BaSiO4:Eu2+ > Li2SrSiO4:Eu2+. Furthermore, using Williamson–Hall (W–H) method, the determined structural strains of Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) phosphors revealed that the increased compressive strain after N3– doping plays an important role in the enhanced PL intensity of these phosphors. White LEDs fabricated by a combination of three N3–-doped phosphors and a 365 nm emitting InGaN chip exhibited the (0.333, 0.373) color coordinate and a high color-rendering index (Ra = 83). These phosphor materials may provide a platform for development of new efficient phosphors in solid-state lighting field.

Experimental Section

Li2MSiO4:Eu2+ (M = Ca, Sr, and Ba) phosphors were prepared from a stoichiometric mixture of MCO3 (M = Ca, Sr, and Ba), SiO2, Li2CO3, and Eu2O3 under a reducing atmosphere (4% H2/Ar) at 900 °C for 12 h. N3–-doped Li2MSiO4−δN2/3δ:Eu2+ (M = Ca, Sr, and Ba) phosphors were prepared from a stoichiometric mixture of MCO3 (M = Ca, Sr, and Ba), α-Si3N4, Li2CO3, and Eu2O3 under NH3 atmosphere at 900 °C for 12 h. The Eu2+ ion concentration was fixed at 1 mol %. X-ray diffraction analysis was carried out using a graphite monochromator (DMAX-2200PC, Rigaku). A step scan mode was selected in a 2θ range (10–110°) with a step size of 0.02° and counting time of 5 s for each step. The refinements of crystal structure were carried out by the Rietveld method using the FullProf program. The diffraction profiles were fitted using a pseudo-Voigt peak function and manually selected background points. FT-IR analyses were carried out using a Bruker VERTEX70 FT-IR spectrometer with an extension in the far-IR region. The extension consists of a multilayer mylar beam splitter, a room-temperature DLATGS detector with preamplifier, and polyethylene windows for the internal optical path. The binding energies of the elements were determined by an X-ray photoelectron spectrophotometer (ESCALAB 250) with a monochromatic Al Kα X-ray source (hν = 1486.6 eV) at Busan Center of Korea Basic Science Institute (KBSI). The determined binding energies were calibrated using the internal standard of adventitious carbon (C 1s at 284.6 eV). SIMS (CAMECA IMS-6f, France) was carried out to determine the elemental composition of the N3–-substituted phosphors. The SIMS standard was used with 14N isotope implanted into SiO2 film as N+ at 100 keV (a dose of 5 × 1014 ions/cm2). To determine the relative sensitivity factor (RSF) of each element, the intensity of secondary ion was corrected by the SIMS results from the standard material. For a careful examination, the Cs+ primary-ion beam was focused using an electron neutralizer for charge compensation (net impact energy = 15 keV; beam current = 20 nA). The PL spectra were obtained using a fluorometer (FS-2 model, Scinco) with a xenon lamp (150 W) under an operating voltage of 350 V. The diffuse reflectance spectra were obtained using a UV–visible spectrophotometer (UV-2600, Shimadzu) with a BaSO4 reference.