Structural evolutions and significantly reduced thermal degradation of red-emitting Sr2Si5N8:Eu2+ via carbon doping

Lihong Liuab, Le Wang*ac, Yuanqiang Lid, Chenning Zhange, Yujin Chof, Siu Wing Orc, Tianliang Zhoug, Takashi Takedab, Naoto Hirosakib and Rong-Jun Xie*bg
aCollege of Optical and Electrical Technology, China Jiliang University, Hangzhou, Zhejiang 310018, China. E-mail:
bSialon Group, National Institute for Materials Science, Namiki 1-1, Tsukuba, Ibaraki 305-0044, Japan. E-mail:
cDepartment of Electrical Engineering, The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong
dIntematix Corporation, Fremont, CA 94538, USA
eFine Particles Engineering Group, Materials Processing Unit, National Institute for Materials Science, Tsukuba, Ibaraki 305-0047, Japan
fNano Device Characterization Group, MANA, National Institute for Materials Science, Namiki 1-1, Tsukuba, Ibaraki 305-0044, Japan
gCollege of Materials, Xiamen University, Xiamen, Fujian 316005, China

Received 28th June 2017 , Accepted 17th July 2017

First published on 17th July 2017

A red-emitting nitridosilicate phosphor, Sr2Si5N8:Eu2+, shows very promising photoluminescence properties but exhibits serious thermal degradation, thus making it difficult to be used practically as a color converter in white light-emitting diodes (wLEDs). To alleviate this problem, we introduce carbon into the Sr2Si5N8 lattice to form thermally robust carbidonitride phosphors (Sr2Si5CxN8−4x/3:Eu2+). The carbon doping, evidenced by a variety of analytical techniques, leads to structural evolutions including lattice shrinkage, shortening of the average bond length of Eu–(C,N), and the removal of Eu3+ ions from the lattice. The photoluminescence intensity and quantum efficiency of phosphors are greatly improved by the carbon doping and reach the maximum at x = 0.5, dominantly owing to the enhanced absorption of Eu2+. Thanks to the increased oxidation resistance of Eu2+ due to the stronger covalency of Si–(C,N) and Sr(Eu)–(C,N) bonds, thermal degradation is significantly reduced from 16 to 0.8% when the carbon doping increases from x = 0 to 1.25. In addition, thermal quenching is also reduced by 10% at 300 °C and the quantum efficiency declines slowly with increasing temperature when carbon is substituted for nitrogen. At 300 °C, the internal quantum efficiencies are 55% and 62% for x = 0 and 0.5, respectively. The enhanced thermal stability of the carbon-doped sample is also confirmed by smaller variations in the luminous efficacy and color coordinates of monochromatic red LEDs.

1. Introduction

Phosphor-converted wLEDs, which are known as indispensable solid-state light sources for the next-generation green lighting technology and display backlighting, have attracted much attention from both academic and industrial communities. It is owing to their excellent features including, but not limited to, energy saving, environmental friendliness, compactness and durability.1–5 The major challenges in the field of wLEDs focus on how to achieve high luminous efficacy, stable color coordinates, and brilliant color-rendering properties. As one of the key materials in solid state lighting, phosphors play a crucial role in determining the optical quality, lifetime, reliability and cost of wLEDs.6–11 Accordingly, it is an endless task for chemists and materials scientists to develop highly efficient and thermally robust luminescent materials for realizing wLEDs with superior optical properties and longevity.

Traditionally, wLEDs are fabricated by coating a yellow phosphor on a blue LED chip.12,13 Although this class of LEDs shows high luminous efficacies, their color rendering index (Ra) is poor (<80), making them unacceptable for general illumination. One of the options to overcome this problem is to combine an additional red phosphor for compensating the red spectral part.14 Several promising luminescent nitride materials, such as M2Si5N8:Eu2+ (M = Ca, Sr, Ba), CaAlSiN3:Eu2+ and Sr[LiAl3N4]:Eu2+, have been discovered and applied in wLEDs to improve the color rendition.15–23 Among these red phosphors, Sr2Si5N8:Eu2+ shows an orange-red emission, and has a shorter emission wavelength than the other two, enabling us to fabricate wLEDs with high luminous efficacy.15,16 However, Sr2Si5N8:Eu2+ suffers from serious thermal degradation, which leads to low reliability and a large luminous loss of wLED devices.24 To apply it as a robust color converter in wLEDs practically, the thermal degradation of Sr2Si5N8:Eu2+ needs to be alleviated or solved.

Several methods have been proposed to enhance the thermal stability of Sr2Si5N8:Eu2+. Brinkley et al. applied a remote-phosphor configuration to separate phosphors from the blue LED chip, and significantly increase the thermal stability of Sr2Si5N8:Eu2+.25 Yeh et al. realized the reduction of thermal degradation in Sr2Si5N8:Eu2+ by partially substituting Sr by Ba, which is due to the increased covalence of Eu–N bonds with Ba substitution.26 Zhang et al. heat-treated Sr2Si5N8:Eu2+ in a N2 or N2–H2 atmosphere, and greatly improved its thermal stability.27,28 This improvement is due to the surface passivation of phosphors. Tian introduced carbon and oxygen into Sr2Si5N8:Eu2+ to form carbidonitride phosphors Sr2Si5N8−(4x/3+z)CxO3z/2:Eu2+, and observed a remarkable improvement in maintaining luminescence at 150 °C.29 The reduced thermal quenching can be interpreted by the contraction of the lattice along the [010] direction that reduces the coupling of phonon modes to the electronic transition of Eu2+ ions.30 Grieco et al. further analyzed the origin of enhanced aging stability by correlating it with structural changes and energy transfer rates of Sr2Si5N8−(4x/3+z)CxO3z/2:Eu2+, and addressed that the lattice distortion associated with the carbon/oxygen substation contributed to improved luminescence stability.30

Carbidonitride phosphors, such as Al1−xSixCxN1−x:Eu2+ and Y2Si4N6C:Ce3+, have recently emerged as a new class of thermally robust luminescent materials for wLEDs.31,32 The crystal structure of these phosphors is composed of highly condensed three-dimensional Si(C,N)4 or Al(C,N)4 tetrahedral building blocks. In these tetrahedral units, the center Si or Al atoms are bonded to C or C/N atoms at four corners, making the carbidonitride phosphors more thermally stable than their carbon-free counterparts. This is attributable to the increased hardness or rigidity of their crystal structures. By considering the changes in the local coordination environment of Eu2+ and the covalence of chemical bonds, it is also possible to reduce the thermal degradation of Sr2Si5N8 by introducing hard Si–C bonds into the host lattice to form the robust carbidonitride tetrahedral framework. Differing from ref. 30, this work aims at investigating the effect of carbon doping on the thermal degradation of Sr2Si5CxN8−4x/3:Eu2+, by understanding the changes in their structural lattice, the valence state of Eu, their thermal cycling behavior and their temperature-dependent quantum efficiency. The reliability of carbon-doped Sr2Si5N8:Eu2+ was also testified by aging the monochromatic red LEDs fabricated by combining a blue LED chip with the red phosphors.

2. Experimental section

2.1. Sample preparation

A series of carbon-doped Sr2Si5N8:Eu2+ (Sr1.96Si5CxN8−4x/3:Eu0.04, x = 0, 0.25, 0.5, 0.75, 1, 1.25) were prepared by firing the powder mixtures of Si3N4, Sr3N2, EuN, and SiC in a gas-pressure furnace (FVPHR-R-10, FRET-40, Fujidempa Kogyo Co. Ltd, Osaka, Japan) with a graphite heater. For each sample, an additional amount of Sr3N2 (equal to the carbon amount) was added as the flux material. The powder mixtures were packed into boron nitride crucibles and heated at a constant heating rate of 600 °C h−1 in a vacuum (<10−3 Pa) from room temperature to 800 °C. At 800 °C, nitrogen gas (99.999% purity) was introduced into the chamber. The powder samples were heated at 1750 °C for 2 h under a nitrogen gas pressure of 1.0 MPa. After firing, the electric power was shut off, and the samples were cooled down naturally with the furnace.

2.2. Phase identification

The phase identification of the synthesized powders was performed by X-ray powder diffraction (XRD, Smart Lab, Rigaku) operating at 40 kV and 40 mA and using Cu Kα1 radiation (λ = 1.5406 Å). The data were collected using a step scan model with a step size of 0.02° and a counting time of 10 seconds per step in the 2θ range of 10–120° for the Rietveld Refinement. The Rietveld refinement from the powder XRD patterns was analyzed using the GSAS package.33 The initial structural model of Sr2Si5N8 was adopted for the structural refinement of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0 and 0.5).34

2.3. Emission scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDS)

Microstructure observations and energy dispersive X-ray spectroscopy (EDS) measurements of the phosphor samples were performed at room temperature using a high-resolution field emission scanning electron microscope (Hitachi, S4800).

2.4. Electron spin resonance (ESR)

Electron spin resonance (ESR) measurements were conducted with an X-band spectrometer (JEOL, JES-FA200) using a cryostat (JEOL, ES-CT470). The absolute magnitudes of the g-value and ESR linewidth, and the number of spins were calibrated using a Mn2+ standard sample.35

2.5. Fourier transform infrared spectra

Fourier transform infrared (FT-IR) spectra were recorded on a FT-IR spectrometer (model 4200, JASCO, Tokyo, Japan) using the standard KBr method.

2.6. X-ray absorption fine structure (XAFS) spectra

The valence state of Eu ions in phosphors was measured using XAFS spectra, which was recorded at the BL37XU beamline of Spring-8 synchrotron radiation facility (Hyogo, Japan).

2.7. Diffuse reflectance spectra

The diffuse reflectance spectra of Sr1.96Si5CxN8−4x/3:Eu0.04 with varying carbon doping concentrations were recorded on a UV-VIS spectrophotometer with an integrating sphere (JASCO; Ubest V-560). The reflection spectrum of Spectralon diffusive white standards was used for calibration.

2.8. Photoluminescence spectra

Photoluminescence spectra were recorded at room temperature using a fluorescence spectrophotometer (F-4500, Hitachi Ltd, Tokyo, Japan) with a 200 W Xe lamp as an excitation source. The emission spectrum was corrected for the spectral response of a monochromator and a Hamamatsu R928P photomultiplier tube using a light diffuser and a tungsten lamp (Noma, 10 V, 4 A).

2.9. Thermal quenching and thermal degradation

The temperature-dependent photoluminescence was measured using a Hamamatsu MPCD-7000 multichannel photodetector with a 200 W Xe-lamp as an excitation source. The phosphor powder was loaded in a hot plate connected to an MPCD-7000 and then heated to the desired temperature at a heating rate of 100 °C min−1. The sample was held at a certain temperature for 5 min to reach thermal equilibrium, which will guarantee a uniform temperature distribution both in the surface and in the interior of samples. The thermal cycling test (heating–cooling–heating) was run five times to evaluate the thermal degradation of phosphor samples.

2.10. High-temperature quantum efficiency

The temperature-dependent quantum efficiency was measured by using a QE-1100 phosphor quantum yield spectrophotometer (Otsuka Electronics, Japan) in the temperature regime of 25–300 °C under 450 nm excitation. External (η0) and internal (ηi) quantum efficiencies (QEs) were calculated by using the equations proposed in ref. 36.

2.11. Fabrication of red-emitting LEDs

Red LEDs were fabricated by combining a blue LED chip (λem = 450 nm) with red Sr1.96Si5CxN8−4x/3:0.04Eu2+ (x = 0 and 0.5) phosphors. They were driven by a forward-bias current of 20 mA at room temperature. The luminous efficacy and color coordinates of aged LEDs (0–1000 h) were measured by using an integrating sphere spectroradiometer system (LHS-1000, Everfine Co., Hangzhou, China).

3. Results and discussion

3.1. Phase identification and crystal structure

The XRD patterns of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0, 0.25, 0.5, 0.75, 1, 1.25) indicate a phase-pure Sr2Si5N8 even when the carbon doping amount is as large as 16 mol% (x = 1.25), as seen in Fig. 1. No SiC diffraction peaks are identified, evidencing the complete dissolution of SiC in the lattice of Sr2Si5N8.
image file: c7tc02908j-f1.tif
Fig. 1 XRD patterns of Sr1.96Si5CxN8−4x/3:Eu0.04 with varying carbon doping concentrations, indicating the pure phase of Sr2Si5N8.

The crystal structure of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0 and 0.5) was further explored in detail to illustrate the change in the structure and coordination environments around Eu2+ ions. Both compositions are crystallized in an orthorhombic system and the space group of Pmn21.37 The Rietveld refinement of the XRD powder pattern (x = 0.5) is presented in Fig. 2. The refined unit cell parameters are a = 5.7112(23) Å, b = 6.8194(29) Å, c = 9.3353(4) Å and V = 363.5769(27) Å3 for the carbon-free sample, whereas these are a = 5.7102(20) Å, b = 6.8138(24) Å, c = 9.3415(33) Å and V = 363.4539(22) Å3 for the carbon-doped sample (x = 0.5) (Table 1). This indicates that the lattice shrinks along the a- and b-axes and expands slightly along the c-axis, which finally causes the decrease of the lattice volume with carbon doping.

image file: c7tc02908j-f2.tif
Fig. 2 Rietveld refinement XRD pattern of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0.5).
Table 1 Crystallographic data of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0 and 0.5)
Formula Sr1.96Eu0.05Si5N8 Sr1.96Eu0.04Si5C0.5N7.333 (x = 0.5)
Crystal system Orthorhombic Orthorhombic
Space group Pmn21 Pmn21
a (Å) 5.7112(23) 5.7102 (20)
b (Å) 6.8194(29) 6.8138(24)
c (Å) 9.3353(4) 9.3415(33)
V3) 363.5769(27) 363.4539(22)
Rwp 7.5% 8.8%
Rp 5.5% 4.6%
χ2 3.6% 4.9%

Table 2 lists the atomic coordinates of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0.5), obtained from the Rietveld refinement. The structure consists of Si(C,N)4 tetrahedra and Sr/Eu atoms that are resided in the voids formed by the tetrahedral framework (Fig. 3a). The same with the carbon-free sample, the Eu2+ ions in the carbon-doped sample are randomly distributed on two individual Sr crystallographic sites (Fig. 3b and c). With carbon doping, the bond length of (Sr,Eu)1–(C,N) decreases, and that of (Sr,Eu)2–(C,N) increases slightly (Table 3). The bond lengths of (Sr,Eu)1–(C,N) and (Sr,Eu)2–(C,N) are 2.7440 and 2.8201 Å for x = 0.5, respectively. Those are 2.7687 and 2.8133 Å for x = 0. The bond length of Si–(C,N) becomes a little bit longer associated with carbon doping, which is attributable to the larger bond length of Si–C (1.90 Å) in comparison to that of Si–N (1.74 Å). They are 1.733 and 1.736 Å in average for x = 0 and 0.5, respectively. The changes in the local structure of Eu2+ will no doubt result in variations of the photoluminescence of Eu2+. Moreover, the change in the bond length provides a direct evidence of the dissolution of carbon in the lattice of Sr2Si5N8.

Table 2 Atomic coordinates and isotropic displacement parameters of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0.5)
Atom Fraction X Y Z Ui/Ue × 100
(Sr,Eu)1 (0.98, 0.02) 0 0.8691(21) −0.5895 0.72
(Sr,Eu)2 (0.98, 0.02) 0 0.8807(21) −0.2210(12) 0.88
Si1 1.00 0.2524(27) 0.6662(21) 0.0923(7) 0.19
Si2 1.00 0 0.0561(30) 0.0861(7) 0.34
Si3 1.00 0 0.4210(65) −0.1301(6) 0.21
Si4 1.00 0 0.4020(67) 0.3110(6) 0.16
(N,C)1 (0.9165, 0.0625) 0 0.1900(17) −0.0631(20) 0.08
(N,C)2 (0.9165, 0.0625) 0.2473(7) 0.9098(6) 0.0849(11) 0.05
(N,C)3 (0.9165, 0.0625) 0.2488(10) 0.4479(7) −0.5776(7) 0.07
(N,C)4 (0.9165, 0.0625) 0 0.5830(11) 0.1851(8) 0.10
(N,C)5 (0.9165, 0.0625) 0 0.1692(18) 0.2551(20) 0.09
(N,C)6 (0.9165, 0.0625) 0 0.4266(11) −0.3187(8) 0.10

image file: c7tc02908j-f3.tif
Fig. 3 (a) Crystal structure of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0.5) viewed from [100], Sr1 (red) and Sr2 (pink), and (b) bond lengths of Sr1–(C,N) and Sr2–(C,N) in carbon-doped Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0.5).
Table 3 (Sr,Eu)–N and (Sr,Eu)–(C,N) bond distances of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0 and 0.5)
Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0) Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0.5)
(Sr,Eu)–N Bond length (Å) (Sr,Eu)–(C,N) Bond length (Å)
(Sr,Eu)1–N1 2.893(3) × 2 (Sr,Eu)1–(C,N)1 2.894(2)
(Sr,Eu)1–N2 2.689(11) × 2 (Sr,Eu)1–(C,N)1 2.899(3)
(Sr,Eu)1–N4 2.883(7) (Sr,Eu)1–(C,N)2 2.647(7) × 2
(Sr,Eu)1–N5 2.565(16) (Sr,Eu)1–(C,N)4 2.869(7)
    (Sr,Eu)1–(C,N)5 2.508(14)
Average 2.7687 Average 2.7440
(Sr,Eu)2–N1 2.555(17) (Sr,Eu)2–(C,N)1 2.573(13)
(Sr,Eu)2–N2 2.675(12) × 2 (Sr,Eu)2–(C,N)2 2.722(8) × 2
(Sr,Eu)2–N3 3.001(6) × 2 (Sr,Eu)2–(C,N)3 2.978(6) × 2
(Sr,Eu)2–N5 2.893(3) × 2 (Sr,Eu)2–(C,N)5 2.884(2) × 2
Average 2.8133 Average 2.8201

3.2. EDS mapping

Fig. 4 presents the elemental mapping of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0.5). It is seen that the phosphor particle contains elements of Sr, Si, Eu, N and C, indicating that carbon is accommodated in the lattice of Sr2Si5N8.
image file: c7tc02908j-f4.tif
Fig. 4 EDS elemental mapping of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0.5), showing the presence of carbon in the lattice.

3.3. ESR measurement

The ESR method is a spectroscopic analysis to observe the transition between levels occurring when the unpaired electron is placed in a magnetic field.38 The number and state information of the unpaired electron can be described by the g value obtained from ESR spectra. The g value is calculated using the following formula:39
= BB0 (1)
where , μB and B0 are the emitting photon of energy, the Bohr magneton and the strength of the external magnetic field, respectively. The ESR spectra can thus be used to probe directly the local environments of rare earth activator ions in phosphor through their unpaired electrons by g values. As the Eu2+ ion has an uncoupled electron (4f7, S = 7/2), while Eu3+ does not have one (4f6, S = 3), the signal observed in Fig. 5 is therefore thought to be originated from Eu2+ ions.40,41 The ESR spectra of the carbon-doped sample (Sr1.96Si5CxN8−4x/3:Eu0.04, x = 0.5) differ in the signal intensity and the line shape from the carbon-free sample (Sr1.96Si5N8:Eu0.04). The enhanced intensity in the carbon-doped sample can be ascribed to (i) the reduced concentration of Eu3+ and (ii) the increased number of spins caused by the tetravalent C4− substitution for the trivalent N3− in the lattice.42 The appearance of new peaks at g = 4.712 and 7.8239 could be attributed to Eu2+ in slightly distorted lattice sites, as the doping of carbon leads to structural evolutions. These new peaks imply that there are multiple local environments around Eu2+ associated with carbon doping.

image file: c7tc02908j-f5.tif
Fig. 5 ESR spectra of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0 and 0.5), showing additional signals after carbon doping.

3.4. FT-IR spectra

FT-IR spectra of Sr1.96Si5CxN8−4x/3:Eu0.04 with varying x values are given in Fig. 6. For the carbon-free sample, the absorption peaks at ∼926, ∼954, and 1000 cm−1 are assigned to Si–N and Si–N–Si stretching vibrations.43 With the carbon substitution, a new absorption peak appears at 1375 cm−1, which is assigned to the Si–C bending vibrations.44 On the other hand, the Si–N and Si–N–Si absorption peaks disappear with increasing carbon substitution, owing to their overlap with the Si–C peaks as well as the reduction in their intensity. All these results confirm the successful substitution of Si–N by Si–C in the structure of Sr2Si5N8:Eu2+.
image file: c7tc02908j-f6.tif
Fig. 6 FT-IR spectra of Sr1.96Si5CxN8−4x/3:Eu0.04 with varying carbon doping concentrations, indicating the presence of Si–C bonds.

3.5. Valence state of Eu

Eu2+ usually co-exists with Eu3+ in the same nitride phosphor lattice, although no characteristic sharp emission lines of Eu3+ can be identified.6,26,45,46 In most cases, Eu3+ is detrimental to the luminescence of Eu2+ because it often acts as a killer.26,27,47 To have an idea of the valence of Eu in Sr2Si5N8, the XAFS analysis of the Eu L3 edge was conducted in the fluorescence mode. As seen in Fig. 7, the X-ray absorption near edge structure (XANES) spectra clearly show that carbon-free Sr2Si5N8 contains both Eu2+ (at 6973 eV) and Eu3+ (at 6982 eV), whereas the carbon-doped one consists of only Eu2+. It indicates that the carbon doping promotes the reduction of Eu3+ to Eu2+, which possibly occurs via the substitution of [Eu3+–N3−] by [2Eu2+–C4−]. In addition, SiC tends to react with SiO2 (present on the surface of Si3N4 starting powders), creating gaseous SiO and CO. These in situ generated reducing gases provide critical driving forces for the reduction of Eu3+ into Eu2+.48 The obvious change in the Eu2+/Eu3+ ratio also provides an additional evidence for the incorporation of carbon into the phosphor lattice. The removal of Eu3+ in Sr2Si5N8:Eu greatly enhances the photoluminescence of Eu2+, as will be presented later.
image file: c7tc02908j-f7.tif
Fig. 7 Eu L3 XANES spectra of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0 and 0.5). Eu2O3 was included as a reference sample to indicate the position of Eu3+.

3.6. Diffuse reflectance and photoluminescence

The diffuse reflection spectra of samples with varying carbon doping concentrations display a broad absorption band ranging from 350 to 580 nm, which is ascribed to the 4f7 → 4f6 5d electronic transitions of Eu2+ (Fig. 8). The carbon doping does not change the optical absorption edge of phosphor samples, but increases the absorption of Eu2+. We contribute this enhancement to the reduction of Eu3+ via carbon doping. In fact, the absorption efficiency of phosphors measured under 450 nm excitation at room temperature is improved from 64 (x = 0) to 85% (x = 0.5).
image file: c7tc02908j-f8.tif
Fig. 8 Diffuse reflectance spectra of Sr1.96Si5CxN8−4x/3:Eu0.04 with varying carbon concentrations, showing the enhanced absorption associated with carbon doping.

The photoluminescence spectra of Sr1.96Si5CxN8−4x/3:Eu0.04 are given in Fig. 9a. A broad emission band centered at ∼620 nm is seen for all samples, owing to the dipole-allowed 4f6 5d → 4f7 electronic transitions of Eu2+. The carbon doping leads to a small blueshift of the emission peak (∼2 nm), which is consistent with the report by Tian.29 As seen in Fig. 9b, the photoluminescence intensity becomes higher as the carbon doping concentration increases, and reaches the maximum at x = 0.5. The enhancement of the emission intensity is dominantly due to the increment of the Eu2+/Eu3+ ratio that significantly reduces the energy migration to the Eu3+ killer sites. Upon further increasing the carbon concentration, the lattice distortion occurs remarkably and the nitrogen vacancies form easily, both of which result in luminescence quenching.

image file: c7tc02908j-f9.tif
Fig. 9 Effects of carbon doping on the (a) emission spectra, (b) emission intensity, and (c) absorption and external quantum efficiencies of Sr1.96Si5CxN8−4x/3:Eu0.04 samples.

The absorption and external quantum efficiency of Sr1.96Si5CxN8−4x/3:Eu0.04 with varying x values measured under 450 nm excitation at room temperature are shown in Fig. 9c. The absorption and external quantum efficiency exhibit a similar tendency to the diffuse reflection spectra (Fig. 8) and the luminescence intensity (Fig. 9a), respectively. The highest external quantum efficiency of 77% is found for the sample with x = 0.5. This sample has an absorption efficiency of 85%. These efficiencies are about 18% and 14% higher than those of the sample without C doping.

3.7. Reduced thermal degradation

One of the main problems encountered in Sr2Si5N8:Eu2+ is the serious thermal degradation of photoluminescence, which leads to the low reliability of wLEDs and thus hinders its practical applications.24,26–28 We evaluated the thermal degradation of the title phosphors by using the thermal cycling protocol. As seen in Fig. 10a, the carbon-free sample degrades in luminescence by 16% in one thermal cycling, clearly showing the large thermal degradation. On the other hand, the luminescence degrades by only 6.5, 3.2 and 0.8% for samples with x = 0.25, 0.5 and 1.25, respectively. This means the luminescence loss is greatly recovered by the carbon doping. Fig. 10b presents the thermal cycling curves of five runs. It is clearly seen that after the first run the luminescence of the carbon-free sample gradually degrades in the next four runs, whereas that of the carbon-doped sample remains nearly unchanged (Fig. 10c). These experimental results confirm that the thermal degradation of Sr2Si5N8 can be significantly minimized or eliminated by the substitution of nitrogen by carbon.
image file: c7tc02908j-f10.tif
Fig. 10 (a) Thermal cycling curves (25 °C → 300 °C → 25 °C at a step of 50 °C) of one run for Sr1.96Si5CxN8−4x/3:Eu0.04 with varying carbon doping concentrations, (b) thermal cycling curves for five runs and (c) thermal degradation after five cycles of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0 and 0.5).

As aforementioned, thermal degradation is a big problem for Sr2Si5N8:Eu2+, which is mainly caused by the oxidation of Eu2+. Yeh et al. also investigated the thermal degradation of Sr2Si5N8:Eu2+, and found that the partial substitution of Sr by Ba could reduce the thermal degradation. This is ascribed to the inhibition of oxidation of Eu2+ due to the increased covalence of Ba–N bonds (compared to Sr–N ones). In this work, as carbon has a smaller Pauling's electronegativity than nitrogen (2.55 vs. 3.04 eV), the Si–C and Sr(Eu)–C chemical bonds are more covalent than the corresponding Si–N and Sr(Eu)–N ones, which thus prevents Eu2+ from oxidation under the thermal cycling test. In addition, the concentration of Eu3+ in the lattice is significantly reduced and finally turns out to be zero as the carbon substitution proceeds. This validates the key role of carbon in improving the oxidation resistance of Eu2+, which enhances the thermal stability and reduces the thermal degradation of the phosphor.

3.8. Enhanced high-temperature luminescence efficiency

The thermal quenching of Sr2Si5N8:Eu2+ is also alleviated by carbon doping, as shown in Fig. 10(a) and (b). At 150 °C, the luminescence declines by ∼12% (with respect to the intensity measured at room temperature) for all samples. At 300 °C, it decreases by 52% for the carbon-free sample, whereas it reduces by 42–39% for the carbon-doped sample. The decrease of thermal quenching is again attributable to the improved covalence of Sr(Eu)–(C,N) bonds as well as the rigidity of the structure built up on robust Si(C,N)4 tetrahedra.

The temperature-dependent quantum efficiencies of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0 and 0.5) are illustrated in Fig. 11. As mentioned in Part 3.6, the absorption efficiency is 64 and 85% for samples with x = 0 and 0.5, respectively. This enhancement is ascribed to the reduced concentration of Eu3+ associated with the carbon substitution which in turn increases the Eu2+ concentration. The internal quantum efficiency almost remains unchanged (92%) for both samples, but it decreases faster for the carbon-free sample at temperatures above 150 °C. At 300 °C, the internal quantum efficiency is, respectively, 55 and 62% for x = 0 and 0.5, clearly indicating the improved thermal stability by carbon substitution. As seen in the inset in Fig. 11, under 450 nm excitation the external quantum efficiency at room temperature is 58 and 74%, which decreases to 32.8 and 48.6% at 300 °C for samples with x = 0 and 0.5, respectively.

image file: c7tc02908j-f11.tif
Fig. 11 Temperature-dependent internal quantum efficiency of Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0 and 0.5). The inset shows the temperature-dependent external quantum efficiency. The measurements were taken by exciting the samples at 450 nm.

3.9. Improved stability in the color point

For practical applications, phosphors and white LEDs should exhibit high reliability against aging at ambient temperature. The variations in luminous efficacy and color coordinates can be used to evaluate the reliability of both phosphors and LED devices. In this work, we fabricated monochromatic red LEDs by combining a blue LED chip with the red Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0 and 0.5) phosphor, and conducted the aging test, as shown in Fig. 12. After 1000 h, the luminous efficacy of the red LED by using Sr1.96Eu0.04Si5N8 reduces by 14.9%, whereas it only declines by 9.7% for the LED using carbon-doped Sr1.96Eu0.04Si5C0.5N7.333. During aging the variations in color coordinates (Δx and Δy) definitely occur, but they are mitigated via carbon doping. Upon carbon doping (x = 0.5), after aging for 1000 h Δx is reduced from 1.26 to 0.92%, and Δy from 0.8 to 0.68%, respectively. Although the reduction is not so great, the reliability of the LEDs is indeed improved to some degree. Both results again evidence the enhanced thermal stability of Sr2Si5N8:Eu2+ by carbon doping.
image file: c7tc02908j-f12.tif
Fig. 12 (a) Electroluminescence spectrum and the color point of the red LED prepared by using carbon-doped Sr1.96Si5CxN8−4x/3:Eu0.04 (x = 0.5); the variations of Δx (b) and Δy (c) as a function of aging time for Sr1.96Si5CxN8−4x/3:Eu0.04 (blue for x = 0 and red for x = 0.5).

4. Conclusions

In summary, we synthesized carbon-doped Sr2Si5N8:Eu2+ phosphors (Sr1.96Si5CxN8−4x/3:Eu0.04, x = 0, 0.25, 0.5, 0.75, 1, 1.25) and investigated the effect of carbon doping on their structural evolution, photoluminescence properties and reliability. With carbon being substituted for nitrogen, the lattice shrinks along the a- and b-axes, and expands along the c-axis, finally causing the reduction of the lattice volume of Sr2Si5N8:Eu2+. The ESR, FT-IR and EDS measurements confirm the incorporation of carbon into the lattice of Sr2Si5N8. The EXAFS analysis indicates that Eu2+/Eu3+ coexists in the carbon-free sample, and Eu3+ is reduced to Eu2+ by the carbon doping. Under 450 nm excitation, the photoluminescence increases with the carbon doping, and reaches the maximum at x = 0.5. The thermal degradation is significantly reduced from 16 to 0.8% (first run of the thermal cycling test) by the carbon doping, which is owing to an increase in both the covalence of (Sr,Eu)–(C,N) bonds and the oxidation resistance of Eu2+. At 300 °C, the loss of internal quantum efficiency is reduced by 7% via the carbon doping (x = 0.5). The enhanced reliability of the carbon-doped Sr2Si5N8:Eu2+ is also confirmed by the smaller variations in both luminous efficacy and color coordinates of the aged red LEDs fabricated by using the title phosphors. These results indicate that carbon-doping enables the improvement of both the luminescence efficiency and thermal stability of Sr2Si5N8:Eu2+ phosphors, and making their practical application in white LEDs possible.


This work was supported in part by the National Natural Science Foundation of China (No. 61575182, 5157223, 51561135015, and 61405183), the Research Grants Council of the HKSAR Government (No. PolyU 5228/13E), the Natural Science Foundation of Zhejiang Province (No. LY16F050004), the Applied Research Program of Commonweal Technology of Zhejiang Province (No. 2015C31102) and JSPS KAKENHI (No. 15K06448).

Notes and references

  1. E. F. Schubert and J. K. Kim, Science, 2005, 308, 1274 CrossRef CAS PubMed.
  2. S. Pimputkar, J. S. Speck, S. P. DenBaars and S. Nakamura, Nat. Photonics, 2009, 3, 180–182 CrossRef CAS.
  3. R.-J. Xie and N. Hirosaki, Sci. Technol. Adv. Mater., 2007, 8, 588 CrossRef CAS.
  4. S. Ye, F. Xiao, X. Y. Pan, Y. Y. Ma and Q. Y. Zhang, Mater. Sci. Eng., R, 2010, 71, 1–34 CrossRef.
  5. S. Nakamura, T. Mukai and M. Senoh, Appl. Phys. Lett., 1994, 64, 1687 CrossRef CAS.
  6. L. Wang, R.-J. Xie, Y. Q. Li, X. J. Wang, C. G. Ma, D. Luo, T. Takeda, Y. T. Tsai, R. S. Liu and N. Hirosaki, Light: Sci. Appl., 2016, 5, e16155 CrossRef CAS.
  7. D. Luo, L. Wang, S. W. Or, H. Zhang and R.-J. Xie, RSC Adv., 2017, 7, 25964 RSC.
  8. C. C. Lin and R. S. Liu, J. Phys. Chem. Lett., 2011, 2, 1268 CrossRef CAS PubMed.
  9. Z. G. Xia and M. Andries, Chem. Soc. Rev., 2017, 46.1, 275 RSC.
  10. Z. G. Xia and Q. L. Liu, Prog. Mater. Sci., 2016, 84, 59 CrossRef CAS.
  11. Z. G. Xia, Z. H. Xu, M. Y. Chen and Q. L. Liu, Dalton Trans., 2016, 11214 RSC.
  12. S. Nakamura and G. Fasol, The Blue Laser Diode: GaN Based Light Emitters and Lasers, Springer-Verlag, Berlin, 1997 Search PubMed.
  13. R.-J. Xie, N. Hirosaki, K. Sakuma, Y. Yamamoto and M. Mitomo, Appl. Phys. Lett., 2004, 84, 5404 CrossRef CAS.
  14. L. Chen, C. C. Lin, C. W. Yeh and R. S. Liu, Materials, 2010, 3, 2172 CrossRef CAS.
  15. 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, J. Alloys Compd., 2006, 417, 273 CrossRef CAS.
  16. R.-J. Xie, N. Hirosaki, T. Suehiro, F. F. Xu and M. Mitomo, Chem. Mater., 2006, 18, 5578 CrossRef CAS.
  17. X. Q. Piao, T. Horikawa, H. Hanzawa and K.-i. Machida, Appl. Phys. Lett., 2006, 88, 161908 CrossRef.
  18. R.-J. Xie, N. Hirosaki, N. Kimura, K. Sakuma and M. Mitomo, Appl. Phys. Lett., 2007, 90, 191101 CrossRef.
  19. R. Mueller-Mach, G. Mueller, M. R. Krames, H. A. Hoppe and F. Stadler, Phys. Status Solidi A, 2005, 202, 1727 CrossRef CAS.
  20. K. Uheda, N. Hirosaki, Y. Yamamoto, A. Naoto, T. Nakajima and H. Yamamoto, Electrochem. Solid-State Lett., 2006, 9, H22 CrossRef CAS.
  21. N. Kimura, K. Sakuma, S. Hirafune, K. Asano, N. Hirosaki and R.-J. Xie, Appl. Phys. Lett., 2007, 90, 051109 CrossRef.
  22. H. Watanabe, H. Wada, K. Seki, M. Itou and N. Kijima, J. Electrochem. Soc., 2008, 155, F31 CrossRef CAS.
  23. P. Pust, V. Weiler, C. Hecht, A. Tucks, A. S. Wochnik, A. K. Henß, D. Wiechert, C. Scheu, P. J. Schmidt and W. Schnick, Nat. Mater., 2014, 13, 891 CrossRef CAS PubMed.
  24. R.-J. Xie, N. Hirosaki, T. Takeda and T. Suehiro, ECS J. Solid State Sci. Technol., 2013, 2, R3031 CrossRef CAS.
  25. S. E. Brinkley, N. Pfaff, K. A. Denault, Z. Zhang, H. T. Hintzen, R. Seshadri, S. Nakamura and S. P. DenBaars, Appl. Phys. Lett., 2011, 99, 241106 CrossRef.
  26. C. W. Yeh, W. T. Chen, R. S. Liu, S. F. Hu, H. S. Sheu, J. M. Chen and H. T. Hintzen, J. Am. Chem. Soc., 2012, 134, 14108 CrossRef CAS PubMed.
  27. C. N. Zhang, T. Uchikoshi, R.-J. Xie, L. H. Liu, Y. J. Cho, Y. Sakka, N. Hirosaki and T. Sekiguchi, J. Mater. Chem. C, 2015, 3, 7642 RSC.
  28. C. N. Zhang, T. Uchikoshi, R.-J. Xie, L. H. Liu, Y. J. Cho, Y. Sakka, N. Hirosaki and T. Sekiguchi, Phys. Chem. Chem. Phys., 2016, 18, 12494 RSC.
  29. Y. Tian, J. Solid State Light., 2014, 1, 11 CrossRef.
  30. C. Grieco, K. F. Hireskorn, A. T. Heitsch, A. C. Thomas, M. H. McAdon, B. A. Vanchura, M. M. Romanelli, L. L. Brehm, A. Leugers, A. N. Sokolv and J. B. Asbury, ACS Appl. Mater. Interfaces, 2017, 9, 12547 CAS.
  31. L. Wang, X. J. Wang, T. Takeda, N. Hirosaki, Y. T. Tsai, R. S. Liu and R.-J. Xie, Chem. Mater., 2015, 27, 8457 CrossRef CAS.
  32. C. H. Hsu and C. T. Lu, J. Am. Ceram. Soc., 2011, 94, 1691 CrossRef CAS.
  33. A. C. Larson and R. B. Von Dreele, Report LAUR 86-748. Los Alamos National Laboratory, Los Alamos, NM, 2000 Search PubMed.
  34. T. Schlieper, W. Milius and W. Schnick, Z. Anorg. Allg. Chem., 1995, 621, 1380 CrossRef CAS.
  35. Y. Cho, B. Dierre, N. Fukata, N. Hirosaki, K. Marumoto, D. Son, K. Takahashi, T. Takeda and T. Sekiguchi, Scr. Mater., 2016, 110, 109 CrossRef CAS.
  36. K. Ohkubo and T. Shigeta, J. Illum. Eng. Inst., 1999, 83, 87 CrossRef.
  37. T. Schlieper, W. Milius and W. Schnick, Z. Anorg. Chem., 1995, 621, 1380 CrossRef CAS.
  38. N. C. George, A. J. Pell, G. Dantelle, K. Page, A. Llobet, M. Balasubramanian and R. Seshadri, Chem. Mater., 2013, 25, 3979 CrossRef CAS.
  39. G. J. Hennig and R. Grün, Quat. Sci. Rev., 1983, 2, 157 CrossRef.
  40. H. D. Xie, J. Lu, Y. Guan, Y. L. Huang, D. L. Wei and H. J. Seo, Inorg. Chem., 2013, 53, 827 CrossRef PubMed.
  41. N. Yamashita, J. Electrochem. Soc., 1993, 140, 840 CrossRef CAS.
  42. S. Oshio, T. Matsuoka, S. Tanaka and H. Kobayahsi, J. Electrochem. Soc., 1998, 145, 3903 CrossRef CAS.
  43. M. I. Baraton, W. Chang and B. H. Kear, J. Phys. Chem., 1996, 100, 16647 CrossRef CAS.
  44. M. Bendikov, Y. Apeloig, S. Bokalov, I. Garbuzova and L. Leites, J. Phys. Chem. A, 2002, 106, 4880 CrossRef CAS.
  45. T. Takeda, N. Hirosaki, S. Funahashi and R.-J. Xie, Chem. Mater., 2015, 27, 5892 CrossRef CAS.
  46. K. Shioi, N. Hirosaki, R.-J. Xie, T. Takeda and Y. Q. Li, J. Alloys Compd., 2010, 504, 579–584 CrossRef CAS.
  47. J. Zhu, L. Wang, T. L. Zhou, Y. J. Cho, T. Suehiro, T. Takeda, M. Lu, T. Sekiguchi, N. Hirosaki and R. J. Xie, J. Mater. Chem. C, 2015, 3, 3181 RSC.
  48. J. Q. Wan, Q. Liu, G. Liu, Z. Zhou, J. Ni and R.-J. Xie, J. Mater. Chem. C, 2017, 5, 1614 RSC.

This journal is © The Royal Society of Chemistry 2017