Structure and luminescence properties of novel rare-earthdoped silicon nitride based materialsLi, Y.

DOI:10.6100/IR594350

Published: 01/01/2005

Document VersionPublisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the author's version of the article upon submission and before peer-review. There can be important differencesbetween the submitted version and the official published version of record. People interested in the research are advised to contact theauthor for the final version of the publication, or visit the DOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and page numbers.

Link to publication

Citation for published version (APA):Li, Y. Q. (2005). Structure and luminescence properties of novel rare-earth doped silicon nitride based materialsEindhoven: Technische Universiteit Eindhoven DOI: 10.6100/IR594350

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal ?

Take down policyIf you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediatelyand investigate your claim.

Download date: 13. Feb. 2018

Structure and Luminescence Properties of Novel

Rare-Earth Doped Silicon Nitride Based Materials

Yuan Qiang Li

CIP-DATA LIBRARY TECHNISCHE UNIVERSITEIT EINDHOVEN Li, Yuan Qiang Structure and luminescence properties of novel rare-earth doped silicon

nitride based materials / by Yuan Qiang Li. – Eindhoven : Technische

Universiteit Eindhoven, 2005.

Proefschrift. – ISBN 90-386-2677-0

NUR 913

Trefwoorden: nitride / luminescentie / lanthanide / aard alkali / kristalstructuur / X-ray diffractie

Subject headings: nitride / luminescence / lanthanide / alkaline earth / crystal structure / X-ray

diffraction

Copyright © 2005, Y.Q. Li Printed by the University Press Facilities, Eindhoven, The Netherlands Cover design by Paul Verspaget, Grafische Vormgeving - Communicatie Cover picture: Luminescence of CaAl2-xSixO4-xNx:Eu2+, (Ca,Sr)2Si5N8:Eu2+, CaSi10Al2N16:Eu2+, BaSi2O2N2:Eu2+ and SrSi2O2-δN2+2/3δ:Eu2+ under the excitation wavelength of 365 nm (rotation from the top in a clockwise direction).

Structure and Luminescence Properties of Novel

Rare-Earth Doped Silicon Nitride Based Materials

PROEFSCHRIFT

ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven

op gezag van de Rector Magnificus, prof.dr.ir. C.J. van Duijn, voor een commissie aangewezen door het College voor Promoties

in het openbaar te verdedigen op dinsdag 30 augustus 2005 om 16.00 uur

door

Yuan Qiang Li

geboren te Shandong, China

Dit proefschrift is goedgekeurd door de promotoren: prof.dr. G. de With en prof.dr.ir. M.C.M. van de Sanden Copromotor dr. H.T. Hintzen

Table of Contents Introduction 1 1 General introduction 1 2 Conversion phosphors for white-light LEDs (Solid-state white lighting) 2 3 Silicon nitride-based materials 4 4 The rare-earth ions in the silicon nitride-based materials 6 5 Scope and outline of this thesis 7 References 9 Chapter 1. Luminescence properties of Ce3+-activated alkaline earth silicon nitride M2Si5N8 (M = Ca, Sr, Ba) materials 11 Abstract 11 1.1 Introduction 12 1.2 Experimental 13 1.3 Results and discussion 14 1.4 Conclusions 26 References 27 Chapter 2. Luminescence properties of red-emitting M2Si5N8:Eu2+ (M = Ca, Sr, Ba) LED conversion phosphors 28 Abstract 28 2.1 Introduction 29 2.2 Experimental 31 2.3 Results and discussion 33 2.3.1. Phase formation and the solubility of Eu2+ in M2Si5N8 (M = Ca, Sr, Ba) 33 2.3.2. Diffuse reflection spectra 35 2.3.3. Photoluminescence properties of Eu2+ in M2Si5N8 38 2.4 Conclusions 45 References 45 Chapter 3. The effect of replacement of Sr by Ca on the structural and luminescence properties of red-emitting Sr2Si5N8:Eu2+ phosphor 47 Abstract 47 3.1 Introduction 48 3.2 Experimental 49 3.3 Results and discussion 51 3.3.1. Effect of incorporation of Ca2+ on the structural characteristics of Sr2Si5N8:Eu2+ 51 3.3.2. Effect of Ca2+ substitution on the luminescence properties of Sr2Si5N8:Eu2+ 55 3.4 Conclusions 58 References 59

i

Chapter 4. Synthesis, structure and luminescence properties of Eu2+ and Ce3+ activated BaYSi4N7 60 Abstract 60 4.1 Introduction 61 4.2 Experimental 62 4.3 Results and discussion 63 4.3.1. Structure determination of undoped BaYSi4N7 63 4.3.2. Solubility of Eu and Ce ions in the BaYSi4N7 host lattice 66 4.3.3. Diffuse reflection of Eu and Ce doped BaYSi4N7 72 4.3.4. Luminescence of BaYSi4N7:Eu2+ 75 4.3.5. Luminescence of BaYSi4N7:Ce3+ 79 4.4 Conclusions 81 References 82 Chapter 5. Preparation, structure and photoluminescence properties of Eu2+ and Ce3+-doped SrYSi4N7 84 Abstract 84 5.1 Introduction 85 5.2 Experimental 86 5.3 Results and discussion 87

5.3.1. Preparation 87 5.3.2. Structure determination 88

5.3.3. Reflection spectra of the undoped and doped SrYSi4N7 compounds 93 5.3.4. Luminescence properties 96

5.5 Conclusions 102 References 102 Chapter 6. Structure and luminescence properties of YTbSi4N6C 104

Abstract 104 6.1 Introduction 105 6.2 Experimental 106 6.3 Results and discussion 107

6.3.1. Synthesis of YTbSi4N7 107 6.3.2. Structure determination 109 6.3.3. Luminescence properties of YTbSi4N6C and Tb2Si4N6C 111 6.3.4. Energy transfer from Ce3+ to Tb3+ in YTbSi4N6 114

6.4 Conclusions 118 References 119 Chapter 7. Structure and luminescence properties of Ce3+-doped Y2Si4N6C 121

Abstract 121 7.1 Introduction 122 7.2 Experimental 124

ii

7.3 Results and discussion 125 7.3.1. Synthesis and phase formation 125 7.3.2. X-ray powder diffraction data and structure of Y2Si4N6C 126 7.3.3. Incorporation of Ce3+ in Y2Si4N6C 130 7.3.4. Luminescence properties of Ce3+-doped Y2Si4N6C 131 7.4 Conclusions 135 References 135 Chapter 8. Luminescence properties of Eu2+-doped MAl2-xSixO4-xNx (M = Ca, Sr, Ba) conversion phosphors for white-LED applications 138 Abstract 138 8.1 Introduction 139 8.2 Experimental 140 8.3 Results and discussion 142

8.3.1. Effect of (SiN)+ substitution for (AlO)+ in MAl2O4 (M = Ca, Sr, Ba) on phase formation and structure 142 8.3.2. Luminescence properties of Eu-doped MAl2-xSixO4-xNx (M = Ca, Sr, Ba) 146 8.4 Conclusions 154 References 154 Chapter 9. Luminescence properties of Eu2+-activated alkaline earth silicon oxynitride MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba): a promising class of novel LED conversion phosphors 157 Abstract 157 9.1 Introduction 158 9.2 Experimental 159 9.3 Results and discussion 161 9.3.1. Phase identification 161 9.3.2. Luminescence of Eu2+-doped MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba) 166 9.3.3. Effect of the Eu2+ concentration on the luminescence of BaSi2O2N2:Eu2+ 169 9.4 Conclusions 172 References 173 Chapter 10. Luminescence of a new class UV- blue- emitting phosphors MSi2O2-δN2+2/3δ:Ce3+ (M = Ca, Sr, Ba) 175 Abstract 175 10.1 Introduction 176 10.2 Experimental 177 10.3 Results and discussion 179

10.3.1. X-ray powder diffraction data of CaSi2O2N2 and SrSi2ON8/3 179 10.3.2. Optical properties 179 10.4 Conclusions 189 References 190

iii

Chapter 11. Optical and magnetic properties of EuSi2O2N2 192 Abstract 192 11.1 Introduction 193 11.2 Experimental 194 11.3 Results and discussion 195 11.3.1. Phase formation 195 11.3.2. X-ray powder diffraction data of EuSi2O2N2 196 11.3.3. Optical properties 200 11.3.4. Magnetic properties 202 11.4 Conclusions 204 References 205 Summary 207 Samenvatting 212 Curriculum vitae 217 Acknowledgements 218 List of publications 221

iv

Introduction

1. General introduction

The rare earth elements have had and still have a unique and important impact on our

lives. The unfilled 4f electronic structure of the rare earth elements makes them have

special properties in luminescence, magnetism and electronics, which could be used to

develop many new materials for various applications such as phosphors, magnetic

materials, hydrogen storage materials and catalysts [1].

Rare-earth-doped luminescent materials (i.e. phosphors) are known to emit at distinct

and different wavelengths in the electromagnetic spectrum and have been widely used in

color cathode ray tubes (CRT), tri-phosphor fluorescent lamps, X-ray intensifying screens

and newly developed vacuum mercury-free lamps, as well as various types of displays

such as plasma display panels, field emission displays and projection TVs [2]. Recently,

breakthroughs in inorganic light emitting diodes (LEDs) technology [3, 4] are

significantly catalyzing the development of energy-efficient solid-state lighting (SSL)

with long lifetime. Solid-state lighting technology has now already penetrated in a variety

of specialty applications, in effect, LEDs have completely changed the “world of

luminance”, for example automobile brake lights, traffic signals, liquid crystal displays

and mobile backlights, flashlights and all manner of architectural spotlights [5]. In

particular, the invention of high-efficiency blue-emitting InGaN-based LEDs makes

realization of efficient, full-spectrum white-light LEDs for general illumination possible

by using conversion phosphors, and will bring about a revolution in lighting industry [6].

However, the excitation sources in LEDs are quite different from those traditional

sources, such as mercury gas-discharge fluorescent lighting. Thus only a very limited

number of present phosphors can meet the minimum requirements for white-light LED

applications. Therefore, to modify existing and explore new phosphor materials with

improved properties for use in white-light LEDs is extremely urgent.

A new class of inorganic phosphors, viz. rare-earth-doped silicon-nitride based

materials, has attracted much attention in recent years due to their high chemical and

thermal stability, as well as their unusual luminescence properties compared to oxide,

sulphide and halide based phosphors or their combinations. Y-Si-O-N:Ce3+ [7],

1

Introduction

LaSi3N5:Eu3+ [8] and M2Si5N8:Eu2+ (M = Ca, Sr, Ba) [9, 10] as well as Eu2+- or Ce3+-

doped α-Sialon [11, 12] are just a few representatives in these potential resources. It is

just the presence of nitrogen in the host lattice that makes the rare-earth ions exhibit

unique optical properties. In this thesis, the focus is not only on investigation of existing

nitride materials to better understand the nature and properties of the rare-earth ions (i.e.

Eu2+, Ce3+ and Tb3+) in these lattices, but also on the exploration of new silicon-nitride

based materials for white-light LED applications.

2. Conversion phosphors for white-light LEDs (Solid-state white lighting)

In the field of general lighting, it is well-known that first the incandescent and

subsequently fluorescent lamps have dominated the lighting market over 100 years. As a

new type of lighting for general illumination, the efficiency of white-light LEDs has

already surpassed that of the incandescent lamps, and will be competitive with the

fluorescent lamps just within ten years (Fig. 1).

Fig. 1. Development of luminous efficiency of traditional and LED lamps [13].

2

Introduction

A further increase in efficiency is expected in the next decades. White-light LEDs have

numerous advantages over those conventional lamps, such as long life-time, energy-

saving, compact and mercury-free which are all very important aspects for environmental

reasons. Without doubt, white-light LEDs have high promise to replace traditional

incandescent and fluorescent lamps as a next generation general lighting.

However, in order to achieve this aim and to fulfill the requirements of this

innovational technology, we have to face several technical challenges. As LEDs are

monochromatic by nature, generating white light from LEDs can be realized by two

general approaches [3, 4]: (1): mixing individual red-green-blue (RGB) LED

combinations to generate white light; and (2): a single InGaN-based blue (~ 465 nm) or

near-UV (NUV: 370 – 410 nm) LED chip coated with one (i.e. yellow emitting) or more

(i.e. green-red and blue-green-red emitting for blue- and NUV-LEDs, respectively)

phosphors that down-convert some of the emission to generate white light by mixing.

Although the RGB approach yield high efficiencies as there is no photon down-

conversion loss, for balancing the RGB output over temperature and over operational

lifetime it requires complex drive and control circuitry. In contrast, phosphor-converted

white LEDs are low-cost, compact devices that can be manufactured to specific color

requirements and that in this respect function more like traditional fluorescent lamps. Fig.

2 shows a schematic structure of a phosphor-conversion white LED lamp, as a typical

YAG:Ce Blue-LED

Yellow emitting

InGaN

phosphor (YAG:Ce)

Blue emitting LED

Fig. 2 Schematic structure of a phosphor-conversion white LED.

3

Introduction

example, based on a blue LED chip combined with a yellow-emitting broadband

phosphor (i.e. YAG:Ce3+ based phosphors) deposited on it. The yellow emitting phosphor

converts a major fraction of the blue excitation light from the chip into yellow light, and

when both are combined white light results [4, 14]. As an alternative, a combination of

green-emitting (e.g., SrGaS2:Eu2+ [2, 15] or (Ca, Sr, Ba)2SiO4:Eu2+ [16]) and red-emitting

(e.g., Ca1-xSrxS:Eu2+ [15]) phosphors instead of a single yellow-emitting phosphor, has

also been proposed which provides extended color range and improved color-rendering

index [15]. Clearly, phosphor materials play a key role to make high quality white light

LEDs. As conversion phosphors for white-light LEDs, these materials must have high

absorption in the near-UV to blue spectral region (370 – 470 nm), a high quantum

efficiency (≥ 90%), high thermal and chemical stability, low thermal quenching and

maintenance of high quantum efficiency in an encapsulating polymer matrix and

minimized degradation. Unfortunately, there are very few existing phosphor materials

that can efficiently convert the UV-blue emission from the LED into green, and in

particular, red light. In addition, red-emitting phosphors that can be efficiently pumped

by UV-blue LEDs are very scarce. With respect to the nowadays applied phosphors,

YAG:Ce3+ based phosphors exhibit reduced efficiency caused by thermal quenching and

suffer from reduced lifetime due to phosphor deterioration [3]. Sulfide-based phosphors

(like SrGaS2:Eu2+ and Ca1-xSrxS:Eu2+) are rather unstable and suffer from large thermal

quenching [15, 17]. Therefore, alternative phosphors with improved properties are in

great demand and have to be urgently prepared, characterized and tested under

application conditions. Consequently, as an explorative research, creating and designing

novel luminescent materials by various ways is a major motivation, strongly reflected in

most chapters in this thesis where we have focused on rare-earth doped silicon-nitride

based materials as a potential class of promising LED conversion phosphors.

3. Silicon-nitride based materials

There are three main groups of nitride materials according to the type of the

interatomic chemical bonding: ionic-like (i.e. alkali and alkaline earth metal nitride),

covalent (i.e. boron, aluminium and silicon nitride) and metallic-like (i.e. transition metal

4

Introduction

nitride) [18, 19]. Because of the presence of oxygen, oxynitride materials normally

exhibit ionic and covalent characteristics or in between them. In this thesis, the nitride

based materials are restricted to the compounds with covalent Si-N bonds (Si-O and Si-C

may be included), in particular these materials containing alkaline-earth (i.e. Ca, Sr, Ba)

and rare-earth ions, such as M-Si-N, M-Ln-Si-N, Ln-Si-C-N, M-Si-O-N and M-Si-Al-O-

N (M = alkaline earth ion, Ln = rare-earth ion).

As a father of nitride materials, α- and β-Si3N4 have to be addressed firstly [20]. Si3N4

consists of a three-dimensional network composed of corner-sharing SiN4 tetrahedra. All

nitrogen atoms are connected to three Si atoms (NSi3 unit). With a molar ratio Si:N = 3:4,

Si3N4 shows the highest degree of cross-linking network. Consequently, Si3N4 based

materials exhibit high chemical and thermal stability, combined with outstanding

mechanical properties [20].

Ternary and quaternary nitrides are composed of metal ions (i.e. alkaline-earth M and

rare-earth Ln) and tetrahedral SiN4 units to which the metal ions are directly coordinated

with nitrogen atoms. However, since the Si-N-Si bonds are broken by introduction of the

metal ions (e.g., Si-N-Si + LnN Si-N- - Ln3+ - -N-Si), the degree of cross-linking

between the SiN4 tetrahedra decreases with increasing M(Ln)/Si ratio in the sequence for

some known compounds: Si3N4 > MSi7N10 > M2Si5N8 > MLnSi4N7 > Ln3Si6N11 > MSiN2

> Ba5Si2N6 > M4SiN4 [21]. As a result, nitrogen exhibits a large range of coordinations

by silicon from a single bond (N[1] in M4SiN4 [22, 23] and Ba5Si2N6 [24]), twofold (N[2]

in MSiN2 [25]) and threefold coordination (N[3] in M2Si5N8 [26, 27] and BaSi7N10 [28] or

SrSi7N10 [29]) to unusual fourfold (N[4] in MYbSi4N7 (M = Sr, Ba) [30, 31] and cubic

Si3N4 [20]) bond with silicon. As expected, the distances Si-N[4] are significantly longer

than those of Si-N[3], Si-N[2] and Si-N[1] [21]. As the degree of cross-linking decreases the

stability of the nitride compounds becomes worse (e.g. Ba5Si2N6, BaSiN2 and Ba4SiN4

are water and air sensitive [22-25]), in this thesis we just focus our work on those

compounds with high degree of cross-linking of SiN4 tetrahedra.

In contrast to the alkaline-earth silicon nitrides, the number of alkaline-earth silicon

oxynitride compounds is very limited up till now; so many new ones still have to be

discovered. Moreover the structures of the known compounds are not well-defined.

CaSi2O2N2 has been found to be a layer structure silicon oxynitride composed of SiON3

5

Introduction

tetrahedra [32], which makes this compound structurally more close to the alkaline-earth

silicates [33].

4. The rare-earth ions in the silicon-nitride based materials

It is well-known that the rare-earth Eu2+ and Ce3+ ions show the 4f ↔ 5d transitions

resulting in a broad band emission in the UV to visible range. Because one electron

within the 5d orbit taking part in the formation of chemical bonding (in the excited state),

the position of the excitation and emission bands strongly depends on the host lattices, i.e.

crystal structure and composition [2]. Therefore, this behaviour allows us to tailor the

excitation and emission spectra by varying the host lattices and adjusting chemical

composition. In this thesis, the Eu2+ and Ce3+ ions are our firstly chosen activators. In

comparison with Eu2+ and Ce3+, the ground state configuration of the Tb3+ ion is 4f8 and

the excited state configuration is 4f75d1, in which the 4f shell is half-filled. As the 4f shell

is well shielded by the outer electrons within the 5s and 5p orbits, the 4f → 4f transitions

of Tb3+ are hardly influenced by the environments. Thus, Tb3+ shows 4f-4f sharp line

emission. Additionally, the 4f75d excitation band is normally located at higher energies

(< 254 nm), so in order to absorb the 254 nm radiation efficiently, the Ce3+ ion is used as

a sensitizer through the energy transfer Ce3+ Tb3+. For example in the case of the

commercial phosphor CeMgAl10O19:Tb3+ (CAT) for use in mercury gas-discharge lamps,

by this way Tb3+ yields efficient green emission from the 5D4 (to 7FJ, J = 6 – 0) level [2].

However, this kind of energy transfer is impossible to be applied for white LEDs because

the transfer occurs at about the 5D3 level (i.e., ~ 370 nm) of Tb3+. A challenge is direct

sensitization of the green Tb3+ emission (5D4 7FJ) for UV-blue light by energy transfer

to the 5D4 level itself (i.e., ~ 490 nm) of Tb3+, which is of great interest for use in white

LEDs. Therefore, the feasibility of using Ce3+ as a sensitizer of Tb3+ is also investigated

in this thesis.

Rare-earth activator ions (like Eu2+, Ce3+ and Tb3+) can be incorporated into the

appropriate alkaline-earth or rare-earth (i.e. Y, La) sites in silicon-nitride or oxynitride

based lattices. By coordination with nitrogen a larger nephelauxetic effect is expected

because nitrogen is more covalent compared to oxygen (i.e. electronegativity 3.07 vs 3.61

6

Introduction

for N and O, respectively). In addition, because nitrogen has a higher formal charge (-3)

than oxygen (-2), a larger crystal-field splitting can be realized. The combination of

above-mentioned two effects is anticipated to eventually result in the lowest 5d excitation

band of Eu2+ or Ce3+ shifting to lower energy (i.e. to longer wavelength more to the

visible region).

5. Scope and outline of this thesis

This thesis reports explorative research concerning the luminescence properties of rare-

earth-doped silicon-nitride based materials for white-light LED applications. The focus is

on searching for improved or unconventional properties in existing materials and in

addition design, synthesis of novel silicon-nitride based materials through special crystal

chemical substitutions (e.g. BaN by YC, or AlO by SiN, etc.). Finally, the invention of

new silicon-nitride based materials and understanding of the relationship between

luminescence properties and structure/composition are also challenging parts of the work

described in this thesis from an application and scientific point of view, respectively.

Here, it is noted that no attempt has been made to give a detailed spectroscopic study,

which is an interesting subject of subsequent studies.

This thesis can be subdivided into the following three main parts:

1. Nitrides

2. Nitride-carbides

3. Oxy-nitrides

The first part (Chapters 1 – 3) studies the luminescence properties of Ce- and Eu-

doped M2Si5N8 (M = Ca, Sr, Ba). The luminescence properties of the Ce ions are

presented in the first chapter. As a new class of Eu2+ red-emitting phosphor materials,

M2Si5N8:Eu2+ (M = Ca, Sr, Ba) shows excellent luminescence properties and has been

used for white LED lighting. The effect of the Eu concentration and the type of M on the

luminescence properties has been investigated in detail (Chapter 2). In addition,

investigation of the influence of partial substitution of Ca for Sr on the luminescence and

7

Introduction

structure is necessary for improvement of its performance, these results are given in

Chapter 3.

The second part (Chapters 4 – 7) describes structural and luminescence properties of

Eu-, Ce-, Tb-doped (M, Ln)LnSi4N6(N, C) (M = Ca, Sr, Ba, Ln = Y). These materials can

be divided into two groups according to the composition: e.g., rare-earth-silicon-nitride

and rare-earth-silicon-nitride-carbide. The latter can be deduced from the BaYSi4N7

lattice by replacement of BaN with YC. In addition, the crystal structure changes from a

hexagonal (P63 mc) to a monoclinic (P21/c) unit cell as the size difference becomes

smaller between both metal ions. Subsequently, these lattices were doped with Ce3+ and

Tb3+ on the Y site. In particular, Ce-doped Y2Si4N6C shows promising for white-light

LED applications (Chapter 6). In Ce-doped YTbSi4N6C, an unusual energy-transfer

manner is found by which green line emission of Tb3+ can be realized by exciting the

Ce3+ ion in the UV-blue range (390 - 480 nm), this is shown in Chapter 7.

The third part (Chapters 8 – 11) deals with silicon (and/or aluminium) oxynitride

materials (viz. Eu- and Ce-doped M2Al2-xSixO4-xNx, M = Ca, Sr, Ba), their structural and

luminescence properties. M2Al2-xSixO4-xNx:Eu2+ (M = Ca, Sr, Ba) (x < 0.6) is deduced

from the well-known phosphor MAl2O4:Eu2+ with stuffed tridymite structure [34, 35]

through (SiN)+ substitution for (AlO)+, contrary to the formation of Sialons which is

obtained from Si3N4 by replacement of (SiN)+ by (AlO)+ [36] (Chapter 8). BaAl2-xSixO4-

xNx:Eu2+ can be efficiently excited in the range of 390 - 440 nm radiation, which makes

this material attractive as a green-emitting phosphor for white-LED lighting applications.

For complete replacement of Al by Si (x = 2), a group of materials with general

composition MSi2O2-δN2+2/3δ results. Two new silicon oxynitride compounds are found

and characterized, i.e. BaSi2O2N2 (Chapter 9), EuSi2O2N2 (Chapter 11). The

luminescence properties demonstrate that Eu-doped MSi2O2-δN2+2/3δ is a new class of

conversion phosphors (i.e. yellow, green-yellow and blue-green emission for M = Ca, Sr,

Ba, respectively) with high promise for white-light LED applications (Chapter 9).

Surprisingly, within a single material together with just only Ce3+, BaSi2O2N2:Ce,Na

yields white light (Chapter 10).

To fully understand the structural characteristics (e.g., the ordering of O/N and Al/Si)

and the nature of rare-earth ions in silicon-nitride based materials, neutron diffraction and

8

Introduction

spectroscopic studies are necessary for subsequent further work. Furthermore,

development of new approaches with low-cost and controllable to produce silicon-nitride

based phosphors should be extended from an industry point of view.

References:

1. B.J. Beaudry and K.A. Gschneidner, Handbook on the Physics and Chemistry of Rare

Earth, vol. 1, ed. By K.A. Gschneidner, Jr and L.R. Eyring, North-Holland,

Amsterdam, 1978.

2. G. Blasse and B.C. Grabmaier, Luminescent Materials, Springer-Verlag, Berlin, 1994.

3. J.Y. Taso, Ed., in Light Emitting Diodes (LEDs) for General Illumination

Update2002, Optoelectronics Industry Development Association, Washington,

DC (2002).

4. A. Zukauskas, M.S. Shur and R. Gaska, Introduction to Solid-State Lighting, John

Wiley & Sons, Inc., New York, 2002.

5. Y. Narukawa, Optic & Photonics News, 2004, 4, 25.

6. S. Nakamura, Appl. Phys. Lett., 1994, 64, 1687.

7. J.W.H. van Krevel, H.T. Hintzen, R. Metselaar, and A. Meijerink,

J. Alloys Comp., 1998, 268, 272

8. K. Uheda, H. Takizawa, T. Endo, H.Yamane, M. Shimada, C.M. Wang and

M. Mitomo, J. Lumin., 2000, 87-89, 967.

9. H.T. Hintzen, J.W.H. van Krevel and G. Botty, EP-1104 799 A1, 1999.

10. H.A. Höppe, H. Lutz, P. Morys, W. Schnick and A. Seilmeier, J. Phys. Chem.

Solids, 2000, 61, 2001.

11. J.W.H. van Krevel, J.W.T. van Rutten, H. Mandal, H.T. Hintzen, and R. Metselaar,

J. Solid State Chem., 2002, 165, 19.

12. R.J. Xie, N. Hirosaki, M. Mitomo, Y. Yamamoto, T. Suehiro and N. Ohashi, J. Am.

Ceram. Soc., 2004, 87, 1368.

13. M.G. Craford, LEDs for Solid State Lighting: Technology, Application, and the

Remaining Challenges, Hong Kong International Technology Conference, 2004.

9

Introduction

14. P. Schlotter, J. Baur, Ch. Hielscher, M. Kunzer, H. Obloh, R. Schmidt and J.

Schneider, Mater. Sci. Eng., 1999, B59, 390.

15. R. Mueller-Mach, G.O. Mueller, M.R. Krames, J. IEEE, 2002, 8, 339.

16. J.S. Kim, P.E. Jeon, J.C. Choi and H.L. Park, Solid State Sci., 2005, 133, 187.

17. M. Yamada, T. Naitou, K. Izuno, H. Tamaki, Y. Murazaki, M. Kameshima and

T. Mukai, Jpn. J. Appl. Phys., 2003, 42, L20.

18. R. Marchand, F. Tessier, A Le Sauze, N Diot, Inter. J. Inorg. Mater., 2001, 3, 1143.

19. B.V. Beznosikov, J. Struct. Chem., 2003, 44, 885.

20. G. Petzow, M. Herrmann, Structure and Bonding, 2002, 102, 47.

21. W. Schnick and H. Huppertz, Chem. Eur. J., 1997, 3, 679.

22. Y. Laurent, Rev. Chim. Miner., 1968, 5, 1019.

23. G. Lang, C.R. Acad. Sci., Ser. C., 1969, 268, 1785.

24. H. Yamane, F.J. DiSalvo, J. Alloys Comp., 1996, 240, 33.

25. Z.A. Gal, P.M. Mallinson, H.J. Orchard and S.J. Clarke, Inorg. Chem., 2004,

43, 3998.

26. T. Schlieper, W. Milius and W. Schnick, Z. Anorg. Allg. Chem, 1995, 621, 1380.

27. T. Schlieper and W. Schnick, Z. Anorg. Allg. Chem., 1995, 621, 1037.

28. H. Huppertz and W. Schnick, Chem. Eur. J., 1997, 2, 249.

29. G. Pilet, H.A. Höppe, W. Schnick and S. Esmaeilzadeh, Solid State Sci., 2005,

7, 391.

30. H. Huppertz and W. Schnick, Angew. Chem. Int. Ed. Engl. 1996, 108, 2115.

31. H. Huppertz and W. Schnick, Z. Anorg. Allg. Chem., 1997, 623, 212.

32. H.A. Höppe, F. Stadler, O. Oeckler and W. Schnick, Angew. Chem. Int. Ed., 2004,

43, 5540.

33. F. Liebau, Structrual Chemistry of Silicates, Springer-Verlag, Berlin, 1985.

34. A.R. Schulze and H.K. Müller-Buschbaum, Z. Anorg. Allg. Chem., 1981, 475, 205.

35. Horkner and H.K. Müller-Buschbaum, Z. Anorg. Allg. Chem., 1979, 451, 40.

36. W K.H. Jack, J. Mater. Sci., 1976, 11, 1135.

10

Chapter 1

Luminescence properties of Ce3+-activated alkaline earth silicon nitride

M2Si5N8 (M = Ca, Sr, Ba) materials ABSTRACT

The luminescence properties of Ce3+, Li+ or Na+ co-doped alkaline-earth silicon nitride

M2Si5N8 (M = Ca, Sr, Ba) are reported. The solubility of Ce3+ and optical properties of

M2-2xCexLixSi5N8 (x ≤ 0.1) materials have been investigated as function of the cerium

concentration by X-ray powder diffraction analysis and fluorescence spectroscopy. X-ray

diffraction results show that the maximum solubility of Ce3+ in M2Si5N8 is about 2.5

mol% (x ≈ 0.05) for both Ca2Si5N8 and Sr2Si5N8, and at 1.0 mol% (x ≤ 0.02) for Ba2Si5N8.

The Ce3+-activated M2Si5N8 phosphors exhibit broad emission bands with maxima at 470,

553 and 451 nm for M = Ca, Sr, Ba, respectively, due to the 4f 5d transition of Ce3+.

In addition, M2Si5N8: Ce3+, Li+ (M = Sr, Ba) obviously shows two Ce3+ emission centers

due to the fact that the Ce3+ ions occupy two M sites. With increasing Ce3+ concentration

both absorption and emission intensity increase and the position of the emission bands

show a slight red-shift (<10 nm). The influence of using Na+ instead of Li+ ion as charge

compensator on emission and excitation properties is small but Na+ enhances the

emission intensity because of larger solubility of Ce3+ in M2Si5N8 (M = Ca, Sr). With

increasing the ionic radius going from Ca to Sr and Ba, the ratio of the emission intensity

to the absorption intensity of Ce3+ decreases related to a decreasing Ce3+ solubility. An

intense absorption and excitation band in the blue range (370 – 450 nm) in M2Si5N8:Ce,

Li(Na) (M = Ca, Sr) points out that these materials are promising conversion phosphors

for white-light LEDs.

Keywords: alkaline earth silicon nitride; luminescence; phosphors; cerium; X-ray

diffraction; white-light LEDs.

11

Chapter 1

1.1. Introduction

Optical properties of Ce-doped compounds have been widely investigated. There is

great interest in Ce3+-doped compounds or crystals for applications as phosphors,

scintillators and tunable lasers [1-3]. It is well established that most Ce3+ activated

phosphors show broad band emission in the UV and visible ranges due to the 4f 05d1 – 4f 1 transition of Ce3+ [1,4] and that their luminescence decays are fast. Usually, the trivalent

cerium ion Ce3+ with the electronic configuration 4f 1 has 2F7/2 and 2F5/2 manifolds as the

ground states separated by ~2000 cm-1 due to spin-orbit coupling. The lower manifold 2F5/2 is populated and the manifold 2F7/2 is almost empty at room temperature. The excited

configuration is 5d which is split by the crystal field in 2 to 5 components. As the

positions of the energy levels of 5d excited states of Ce3+ are not only affected by the

symmetry and strength of the crystal field but also by the degree of covalent bonding, it

causes variations in the absorption and emission from UV to long-wavelength by varying

the host lattice [1].

Although the luminescence properties of Ce3+ activated compounds are rather well

known, most of these investigations are focused on oxides, sulfides and halides [1-4]. To

our knowledge, only a few studies on luminescence of Ce3+-doped nitrides or oxynitrides

have been reported [5-8]. Recently, the luminescence properties of a series of Ce3+ -

doped oxynitride compounds in the Y-Si-O-N system (Y5(SiO4)3N, Y4Si2O7N2 YSiO2N

and Y2Si3O3N4) and a modified Ce3+-doped Y2Si3-xAlxO3+xN4-x melilite compound have

been presented [6, 7]. Those investigations have shown that long wavelength emission of

Ce3+ can be observed, for example, Y4Si2O7N2:Ce exhibits a maximum emission band up

to 504 nm. This long-wavelength emission is ascribed to the large crystal-field and high

covalency of the lattice site due to nitrogen incorporation. Ce-doped Ca-α-sialon also

shows long-wavelength emission peaking at 515 – 540 nm, with high quantum efficiency

[8, 9]. In addition, Ce-doped lanthanide-oxynitride glasses show very interesting

luminescence properties and the Ce3+ emission can be varied over a large spectral interval

(380 – 500 nm) with change of chemical composition and Ce concentration [10].

With respect to pure nitride phosphors, so far only the Ce3+ luminescence in MYSi4N7

(M = Sr, Ba) has been reported [11, 12]. Another interesting nitride host lattice is

12

Chapter 1

M2Si5N8 (M = Ca, Sr, Ba) for which Eu2+ luminescence has been reported [5, 13, 14].

Unusual red Eu2+ emission (600-660 nm) with intense absorption bands in the visible

range was found. These red-emitting nitride materials have already demonstrated to be

high potential conversion phosphors in white-light LEDs based on blue-emitting (In,

Ga)N chips. As far as we know, no reports have been given with regard to Ce3+

luminescence in M2Si5N8 (M = Ca, Sr, Ba).

In the present paper, we focus on the influence of M cation type on the solubility limit

of the Ce3+ ion in M2Si5N8 (M = Ca, Sr, Ba) and the luminescence properties of Ce3+-

doped M2Si5N8 (M = Ca, Sr, Ba) at room temperature. Effect of the charge compensator

on the luminescence properties is also discussed.

1.2. Experimental

1.2.1. Sample preparation

Powder samples with composition M2-2xCexLixSi5N8 (0 ≤ x ≤ 0.1) and M2-

2xCexNaxSi5N8 (x = 0.1) were prepared using Ce (Alfa, > 99%, lumps), Li (Merck, > 99%,

lumps), Na (Merck, > 99%, pieces), Ba (Aldrich, > 99%, pieces) and Sr (Aldrich, 99%,

pieces) metals, and the nitrides Ca3N2 (Alfa, 98%, powder) as well as Si3N4 (Permascand,

P95H, α content 93.2%; Oxygen content: ~1.5%) as starting materials. Firstly, the binary

alkaline-earth nitrides MNx (M = Sr, Ba; x ≈ 0.6 – 0.67) were synthesized by nitridation

of Ba and Sr metals at 550 and 800 °C, respectively, for 5-10 hours in a horizontal tube

furnace under nitrogen atmosphere. Subsequently, appropriate amounts of the metal (Ce,

Li or Na), alkaline-earth nitrides Ca3N2 and MNx (M = Sr, Ba) and Si3N4 were weighed

out, mixed and ground in an agate mortar. The powder mixtures were fired in

molybdenum crucibles at 1300-1400 °C for 12 h in a horizontal tube furnace under N2-H2

(10%) atmosphere. After firing, the samples were gradually cooled down in the furnace.

Subsequently, the samples were ground and refired for 32 h under the same conditions.

All processes were handled in a dry glovebox flushed with dry nitrogen because of air

and water sensitivity of some starting materials.

13

Chapter 1

1.2.2. X-ray diffraction analysis

All measurements were performed on finely ground samples, which were analyzed by

X-ray powder diffraction (Rigaku, D/MAX-B) using Cu-Kα radiation at 40 kV and 30

mA with a graphite monochromator.

Lattice parameters determination was carried out by the least-squares method from X-

ray diffraction data collected between 10 and 90° 2θ in step scan with a step size of 0.01 o

in 2θ and a count time of 5 seconds using 10 wt% silicon powders as an internal standard.

1.2.3. Optical measurements

The diffuse reflectance, emission and excitation spectra of the samples were measured

at room temperature by a Perkin Elmer LS 50B spectrophotometer equipped with a Xe

flash lamp. The reflection spectra were calibrated with the reflection of black felt

(reflection 3%) and white barium sulfate (BaSO4, reflection ~100%) in the wavelength

region of 230-700 nm. The excitation and emission slits were set at 5 nm. The emission

spectra were corrected by dividing the measured emission intensity by the ratio of the

observed spectrum of a calibrated W-lamp and its known spectrum from 300 to 900 nm.

Excitation spectra were automatically corrected for the variation in the lamp intensity by

a second photomultiplier and a beam-splitter. All the spectra were measured with a scan

speed of 100 nm/min.

1.3. Results and discussion

1.3.1. The solubility of Ce3+ in M2Si5N8

According to the powder X-ray diffraction patterns, the samples with the composition

M2-2xCexLixSi5N8 (0 ≤ x ≤ 0.1) and M2-2xCexNaxSi5N8 (x = 0.1) were obtained as nearly

single phase with only a trace of unknown secondary phase.

The degree of substitution of Ce3+ for M2+ together with Li+ as charge compensator

was determined in Ce, Li co-doped M2Si5N8. The lattice parameters varying with the Ce3+

concentration are given in Fig.1.1 The values obtained for the undoped compounds are

very close to those previously published [15, 16]. The unit cell volume shows a very

slight decrease by only 0.03% and 0.04% with x up to 0.05 for Ca2-2xCexLixSi5N8 and Sr2-

14

Chapter 1

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11721.40

721.45

721.50

721.55

721.60

721.65

721.70

721.75

721.80

721.85

0.05

0.04

0.03

0.02

0.01

0.00

Uni

t cel

l vol

ume

(Å3 )

x

Ca2Si5N8:Ce,Li

∆V/V

O (%

)

(a)

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11363.12

363.14

363.16

363.18

363.20

363.22

363.24

363.26

363.28

363.30

363.32

363.34

363.36

0.06

0.05

0.04

0.03

0.02

0.01

0.00

Uni

t cel

l vol

ume

(Å3 )

x

Sr2Si5N8:Ce,Li

∆V/

V O (%

)

(b)

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11376.56

376.62

376.68

376.74

376.80

376.86

376.92

376.98

377.04

377.10

377.16

377.22

0.08

0.06

0.04

0.02

0.00

-0.02

-0.04

-0.06

-0.08

Uni

t cel

l vol

ume

(Å3 )

x

Ba2Si5N8:Ce,Li

∆V/V

0 (%)

(c)

Fig.1.1. Ce3+ concentration dependence of unit cell volume and relative volume change

for (a) Ca2-2xCexLixSi5N8, (b) Sr2-2xCexLixSi5N8, and (c) Ba2-2xCexLixSi5N8.

15

Chapter 1

2xCexLixSi5N8, respectively. This decrease can be understood from the fact that the Ce3+-

Li+ ion pair has a smaller size than that of Ca2+ and Sr2+ [17]. On the other hand, the unit

cell volume for Ba2-2xCexLixSi5N8 samples does not show a significant decrease compared

to the undoped host lattice also at high x value (x > 1.0, not shown in Figure 1). Thus the

solubility limit of Ce3+ in Ca2Si5N8 and Sr2Si5N8 co-doped with Li+ is approximately x =

0.05, while only a very small amount of Ce3+ can be incorporated in Ba2Si5N8 and the

substitution degree of Ce, Li for Ba is about x = 0.02 or less. Considering the similar

structure with Sr2Si5N8, a larger difference in the ionic radius between Ce3+ and Ba2+ (20

– 26% as compared to 9-13% for the difference between Ce3+ and Sr2+) is considered to

be the main reason for the lower solubility of Ce3+ in Ba2Si5N8. This very low solubility

of Ce3+ can also be found for heterovalent Ce-substitution in Ba3(PO4)2 and Ba-silicate

compounds [18, 19].

1.3.2. Diffuse reflectance spectra

Fig. 1.2 shows the diffuse reflectance spectra of M2-2xCexLixSi5N8 (M = Ca, Sr, Ba)(0 ≤ x ≤ 0.1). The undoped M2Si5N8 sample is grey-white powder and shows strong

absorption in the UV range. The absorption edges are at about 243, 244 and 247 nm for

Ca2Si5N8, Sr2Si5N8 and Ba2Si5N8, respectively, essentially due to the valence-to-

conduction band transitions of the M2Si5N8 host. The Ce3+, Li+ co-doped M2Si5N8 (M =

Ca, Sr, Ba) powders have daylight color varying from light to deep yellow-green due to

absorption bands superimposed in the blue range (400-450 nm) in the reflectance spectra

(Fig. 1.2). These bands are ascribed to 4f 5d transitions of Ce3+ because with increasing

Ce concentration the absorption bands become stronger.

Ca2Si5N8: Ce, Li possesses a light yellow-green color and has an intense absorption

band with three sub-bands identified at about 395, 367, and 327 nm (Fig. 1.2(a)).

Sr2Si5N8: Ce, Li has a deep yellow-green color and displays a strong, broad absorption

band with two overlapping maxima peaking at about 375 and 420 nm (Fig. 1.2(b)).

Ba2Si5N8: Ce, Li shows a light dull yellow-green color with a very weak absorption band

at about 410 nm (Fig. 1.2(c)). Clearly, from the absorption intensity it can be concluded

that Ca2Si5N8 and Sr2Si5N8 are more suitable for Ce3+ incorporation than Ba2Si5N8, in

16

Chapter 1

agreement with the XRD results on the Ce3+ solubility limit. It is important to note that

the Ce3+ ion shows absorption in the UV-blue range in all cases (Fig. 1.2).

200 300 400 500 600 7000

10

20

30

40

50

60

70

80

90

100

Ca2Si5N8:Ce,Li

Ref

lect

ion

(%)

Wavelength (nm)

x 0 0.02 0.05

(a)

200 300 400 500 600 700

0

10

20

30

40

50

60

70

80

90

100

Ref

lect

ion

(%)

Wavelength (nm)

x 0 0.02 0.05

Sr2Si5N8:Ce,Li (b)

200 300 400 500 600 700

0

10

20

30

40

50

60

70

80

90

100

Ba2Si5N8:Ce,Li

Ref

lect

ion

(%)

Wavelength (nm)

x 0 0.02 0.05

(c)

Fig.1.2. The diffuse reflectance spectra of (a) Ca2-2xCexLixSi5N8 (b) Sr2-2xCexLixSi5N8,

and (c) Ba2-2xCexLixSi5N8 (x = 0, 0.02, 0.05).

17

Chapter 1

Table 1.1. Composition, phase characteristics and luminescence properties of Ce3+- doped M2Si5N8 :Ce, A (M = Ca, Sr, Ba; A = Li, Na).

Composition Phase Body color Absorption bands (nm)

Ce 5d excitation bands (nm)

Emission bands (nm)

Crystal field splitting (cm-1)

Stokes shift (cm-1)

Ca2Si5N8 : Ce, Li Ca2Si5N8Light yellow-green

250, 327, 367, 395

261,288, 329, 365, 397 470 ~ 13100 ~ 3900

Ca2Si5N8 : Ce, Na Ca2Si5N8Light yellow-green

251, 329, 367, 395

260, 286, 329, 373, 396 471 ~ 13100 ~ 3900

Sr2Si5N8 : Ce, Li Sr2Si5N8 Yellow-green 240, 260, 327, 375, 420

260, 276, 330, 387, 431 (CeSr1) 259, 272, 327, 395 (CeSr2)

495 (CeSr1) 553 (CeSr2)

~15300(CeSr1) ~13300(CeSr2)

~ 3000 ~ 7200

Sr2Si5N8 : Ce, Na Sr2Si5N8 Yellow-green 242, 261, 327, 377, 422

260, 279, 328, 396, 434 (CeSr1) 261, 280, 326, 395 (CeSr2)

520 (CeSr1) 556 (CeSr2)

~15400(CeSr1) ~13000(CeSr2)

~ 3800 ~ 7300

Ba2Si5N8 : Ce, Li Ba2Si5N8 Light yellow-green 250, 370, 410

260, 284, 384, 415(CeBa1) 257, 285, 380, 405(CeBa2)

451, 497 (CeBa1) 561 (CeBa2)

~14400(CeBa1) ~14200(CeBa2)

~ 2000 ~ 6900

Ba2Si5N8 : Ce, Na Ba2Si5N8Light yellow-green 253, 373, 412

258, 285, 384, 416 (CeBa1) 259, 286, 384, 406 (CeBa2)

457, 495 (CeBa1) 560 (CeBa2)

~14700(CeBa1) ~14000(CeBa2)

~ 2200 ~ 6800

18

Chapter 1

1.3.3. Luminescence of M2Si5N8:Ce, Li

The results obtained for the Ce3+-activated M2Si5N8 (M = Ca, Sr, Ba) compounds are

summarized in Table 1.1. Besides Ce3+, Li+ co-doped samples, also Ce3+, Na+ co-doped

samples are included in this overview table.

1.3.3.1. Ca2Si5N8: Ce, Li

Fig. 1.3 shows the excitation and emission spectra of Ca2-2xCexLixSi5N8 for x = 0.02,

0.05 and 0.1. Three distinct excitation bands are detected around 250, 329 and 397 nm,

plus a weak band and two shoulders at 288, 261 and 370 nm, respectively. These bands

can also be found in the corresponding reflection spectra (Fig. 1.2(a)). Definitely, the

shortest excitation band around 250 nm originates from host lattice excitation as can be

concluded from the reflection spectrum (Fig. 1.2(a)). The remaining excitation bands are

assigned to Ce3+ 4f 5d transitions separated by crystal field splitting of the 5d state.

The emission spectra display a narrow (FWHM ~ 95 nm for x = 0.05) symmetric

band extending from 400 to 640 nm with a maximum at about 470 nm irrespective of the

excitation wavelength (Fig. 1.3). Although there are two crystallographic Ca sites in

Ca2Si5N8 [16], only a highly symmetric emission band implies a similar environment

around the two CeCa sites which probably makes the emission bands largely overlap for

the two centers. Also the doublet bands due to the transitions from the 5d excited state to

the two ground state levels of Ce3+ (2F7/2 and 2F5/2) cannot be distinguished directly which

means a relatively strong crystal field at the Ce3+ ion in Ca2Si5N8 resulting in extensive

splitting of the 4f ground state [20]. However the emission band can be decomposed into

two Gaussian bands centered at about 465 and 510 nm with an energy difference of about

1900 cm-1 corresponding favourably to the splitting of the 4f1 ground state configuration

of the Ce3+ ion (the energy difference between 2F7/2 and 2F5/2 levels normally is about

2000 cm-1 [1]).

For an excitation wavelength of 395 nm, the ratio of the emission intensity to the

absorption intensity (at 400 nm) increases for higher Ce3+ concentrations. A very limited

red-shift (<10 nm) of the emission band can be observed for Ca2Si5N8: Ce, Li for higher

Ce concentrations in agreement with a negligible change of the lattice parameters (~

0.03% for the unit cell volume). The crystal field splitting of the Ce3+ 5d level is about

19

Chapter 1

13100 cm-1 and the estimated Stokes shift is about 3900 cm-1 for Ca2Si5N8: Ce, Li (Table

1.1).

200 250 300 350 400 450 500 550 600 650 7000

50

100

150

200

250

Inte

nsity

(a. u

.)

Wavelength (nm)

x 0.02 0.05 0.10

λem = 470 nm λexc = 395 nm

Ca2Si5N8:Ce,Li

Fig. 1.3. Excitation and emission spectra of Ca2-2xCexLixSi5N8 (x = 0.02, 0.05, 0.1).

1.3.3.2. Sr2Si5N8: Ce, Li

The excitation spectra of Ce3+, Li+ co-doped Sr2Si5N8 show approximately three

intense maxima at about 260, 325 and 397 nm, and include a weak band at 266 nm and a

shoulder around 435 nm (Fig. 1.4(a)). This observation fairly agrees with the reflection

spectra (Fig. 1.2(b)). In all cases, the emission spectra show very broad bands from 420

to 700 nm centered around 553 nm.

Excitation and emission spectra of Sr2Si5N8: Ce, Li (x = 0.02) with two different

monitoring wavelengths (490, 560 nm) are shown in Fig. 1.4(b). Evidently, with the

dominant excitation band changing from short to long wavelength, the corresponding

emission band shifts in the opposite direction, i.e. from long wavelength to short

wavelength. This result indicates that Ce3+ really occupies two different Sr sites in

Sr2Si5N8, and moreover exhibits significant differences in the crystal field splitting and in

particular the Stokes shift (Table 1.1). The fact that the structural difference between the

20

Chapter 1

200 250 300 350 400 450 500 550 600 650 700 750 800

0

100

200

300

400

500

600

700

λem = 560 nm λexc = 397 nm

Inte

nsity

(a. u

.)

Wavelength (nm)

x 0.02 0.05 0.10

Sr2Si5N8:Ce,Li

(a)

200 250 300 350 400 450 500 550 600 650 700 7500.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0Sr2Si5N8:Ce,Li

Inte

nsity

(a.u

.)

Wavelength (nm)

λem= 560 nm λexc= 397 nm

λem= 490 nm λexc= 430 nm

(b)

Fig.1.4.(a) Excitation and emission spectra of Sr2-2xCexLixSi5N8 (x = 0.02, 0.05, 0.1).

(b) Excitation and emission spectra of Sr2-2xCexLixSi5N8 (x = 0.02) with different

monitoring and excitation wavelengths. For solid line λmon = 560 nm, λexc = 397 nm;

for the dotted line λmon = 490 nm and λexc = 430 nm.

21

Chapter 1

two cation sites is larger in Sr2Si5N8 as compared to Ca2Si5N8 can well explain the above

observation [15, 16]. Because the average CeSr-N distance is smaller for CeSr(1) than that

for CeSr(2) together with a smaller coordination number (CN = 6 for CeSr(1) vs. CN = 7

for CeSr(2)), obviously, the crystal field at CeSr(1) is larger than that at CeSr(2). Therefore,

the splitting of the 5d levels of Ce3+ is larger for CeSr(1) than that for CeSr(2). Comparing

the estimated crystal field splitting (CFS) from the excitation spectra (λmon = 560 nm,

CFS ~ 13300 cm-1; λmon = 490 nm, CFS ~ 15300 cm-1), we can reasonably assign the

excitation band with a maximum at about 431 nm to CeSr(1), and the excitation band at

about 395 nm to CeSr(2) (Table 1.1).

With increasing the Ce3+ concentration, the ratio of the emission intensity to the

absorption intensity increases. Similar to Ca2Si5N8:Ce, Li, no significant emission band

shift (~ 6 nm red-shift) can be found in Sr2Si5N8:Ce, Li due to a very small shrinkage of

the unit cell volume (Fig. 1.1(b)). In Sr2Si5N8:Ce, Li, the ratio of the emission intensity of

CeSr(2) (~ 560 nm) to CeSr(1) (~ 495 nm) increases from 1.13 to 1.31 with the Ce3+

concentration increasing from 1 to 2.5 mol% (Fig. 1.4(a)), while still a further increase is

observed at 5.0 mol% Ce (~1.35). Therefore, the higher emission intensity of the CeSr(2)

center is ascribed to the energy transfer from CeSr(1) to CeSr(2) as the distance between

the Ce3+ ions decreases with the Ce concentration increasing [1].

1.3.3.3. Ba2Si5N8: Ce, Li

Also the excitation spectra of Ba2Si5N8:Ce,Li (Fig. 1.5) show a strong resemblance to

the reflection spectra (Fig. 1.2(c)). Two distinct broad bands in the excitation spectra can

be identified (Fig. 1.5), peaking at about 250 nm (consisting of several peaks starting

from 230 extending to 290 nm) and 405 - 415 nm, respectively. One shoulder around 384

nm is also present in the excitation spectra (Table 1.1).

The emission spectrum of Ba2-2xCexLixSi5N8 (x = 0.02) consists of a triplicate-peaked

band between 425 and 700 nm (Fig. 1.5). The maxima are located at about 451, 497, and

560 nm. The first two emission maxima are separated by 2050 cm-1 in a energy scale, and

are therefore ascribed to transitions from the 5d lowest crystal-field component to the 2F7/2 and 2F5/2 ground levels of Ce3+ incorporated on one of the Ba sites in Ba2Si5N8.

Furthermore, the emission band with a longer wavelength of about 560 nm is attributed to

22

Chapter 1

200 250 300 350 400 450 500 550 600 650 700 7500.0

0.2

0.4

0.6

0.8

1.0 λem= 455 nm λexc= 415 nm

415

Inte

nsity

(a.u

.)

Wavelength (nm)

0.2

0.4

0.6

0.8

1.0

200 250 300 350 400 450 500 550 600 650 700 750

Ba2Si5N8:Ce,Li λem= 565 nm λexc= 405 nm

405

Fig. 1.5. Excitation (solid line) and emission (dashed line) spectra of Ba2-2xCexLixSi5N8

(x = 0.02).

Ce3+ incorporated on the second Ba site. Two different excitation maxima centered at

about 415 and 405 nm can be observed by monitoring at the emission wavelengths of 455

and 560 nm, respectively (Fig. 1.5). As also found in Sr2Si5N8:Ce,Li, the position of the

emission maxima shifts to shorter wavelength with an increase in excitation wavelength.

Considering the crystal field splitting (Table 1.1), the emission band at 451 nm is

ascribed to the CeBa1 emission center (corresponding to excitation wavelength at about

415 nm) while the emission band centered at 560 nm (corresponding to the excitation

wavelength at about 405 nm) can be assigned to the CeBa2 emission center.

Similar to Ce3+, Li+-doped Ca2Si5N8 and Sr2Si5N8, with increasing Ce3+ concentration

the ratio of the emission intensity to the absorption intensity increases and no significant

band shift occurs because of very limited Ce3+ incorporation as mentioned earlier.

23

Chapter 1

Compared with Sr2Si5N8:Ce,Li with the same overall Ce3+ concentration, the emission

intensity of the second emission center (at 560 nm) is significantly lower than that of the

first emission center (at about 451 nm). This suggests that the Ce3+ ions preferentially

occupy the smaller Ba(1) site and only a very small amount of Ce3+ is incorporated on the

larger Ba(2) site in agreement with expectations from the smaller size of Ce3+ as

compared to Ba2+. In addition, the energy transfer between Ce3+ ions is less efficient due

to a larger distance between the nearest Ce3+ ions in Ba2Si5N8:Ce,Li than that in

Sr2Si5N8:Ce,Li.

1.3.4. Influence of the M ions (M = Ca, Sr, Ba) and the charge compensator

(Na vs. Li)

From luminescence spectra it is clear that Ce3+ is incorporated on two sites in

Sr2Si5N8 and Ba2Si5N8 with the same crystal structure. For both Ce sites the Stokes shift

is higher in Sr2Si5N8 as compared to Ba2Si5N8, which is as expected because a larger

relaxation of an excited Ce3+ ion can take place when a Ce3+ ion substitutes on a smaller

lattice site based on the fact that the ionic radius of Sr2+ is smaller than that of Ba2+ [17].

The Stokes shift of the smaller M(1) site being lower than that of the larger M(2) site

indicates a less rigid coordination of Ce3+ on the M(2) site. As already mentioned this

large M(2) site is only occupied by a very low fraction of the Ce3+ ions in Ba2Si5N8. So

not only the total amount of incorporated Ce3+ is limited (as concluded from XRD

measurements) due to size differences between Ba2+ and Ce3+, but moreover preferential

occupation of Ce3+ on the smaller Ba(1) site takes place. For Ce-doped Ca2Si5N8 a

random distribution of the Ce3+ ions over both Ca sites is expected from similar sizes of

Ce3+ and Ca2+ [17]. However, in this case the excitation bands of the two Ce3+ centers

cannot be discriminated, indicating similar coordinations on both Ca sites [16]. Also for

Ce-doped Ca2Si5N8, the value of the crystal field splitting is smaller than that of

M2Si5N8:Ce3+ (M = Sr, Ba), while actually larger values would be expected due to the

lower coordination number (CN = 5) as well as a shorter CeCa-N distance as compared to

the Sr and Ba compounds. Evidently the crystal structure of Ca2Si5N8 induces more

covalent and rigid binding of Ce3+ than in the case of the M2Si5N8 structure (M = Sr, Ba).

24

Chapter 1

Accordingly, Ca2Si5N8:Ce3+, Li+ exhibits the highest emission intensity (with a maximum

at about 470 nm) in M2Si5N8:Ce3+, Li+.

5 10 15 20 25 30 35 40 45 50

200 250 300 350 400 450 500 550 600 650 700 750 800

0.0

0.2

0.4

0.6

0.8

1.0 λem = 570 nm λexc = 397 nm λexc = 430 nm

Inte

nsity

(a. u

.)

Wavelength (nm)

520 556

Sr1.8Ce0.1Li0.1Si5N8

(222

)(213

)(032

)(1

31)

(130

)(1

23)

(220

)(0

31)

(030

)(2

12)

(023

)(0

04)

(202

)(1

22)

(113

)

(210

)

(103

)

(013

)(2

00)

(120

)

(112

)(0

21)

(020

)

(012

)(1

11)

(110

)

(002

)

(101

)

(011

)Inte

nsity

(cou

nts)

2θ (deg.)

(010

)

*

Sr1.8Ce0.1Na0.1Si5N8

Fig. 1.6. XRD pattern of Sr1.8Ce0.1Li0.1Si5N8 and Sr1.8Ce0.1Na0.1Si5N8. (*) indicates the

second phase. Inset shows the excitation and emission spectra corresponding to

Sr1.8Ce0.1Na0.1Si5N8.

By comparing the co-doped pairs of Ce - Li or Ce - Na and Ce alone in M2Si5N8, it is

found that the type of charge compensator has a small influence on the luminescence

behaviors (Table 1.1). This indicates that the CeM surroundings are rather similar with

and without Li+ or Na+ co-doping in M2Si5N8 host lattices. In addition, X-ray diffraction

analysis shows that the solubility of Ce3+ ions in M2Si5N8 (M = Ca, Sr) is increased using

Na+ instead of Li+ ions. In the case of Sr2Si5N8:Ce, Na, at least 5 mol% Ce3+ can be

incorporated into Sr2Si5N8 lattice (Fig. 1.6). Consequently, the ratio of the emission

intensity to the absorption intensity (at 400 nm) is increased significantly by using Na+ as

charge compensator in Ce-doped M2Si5N8 (M = Ca, Sr). As a result of more Ce3+

incorporation into the lattice, the observed emission bands shift slightly towards long

25

Chapter 1

wavelength (see the inset in Fig. 1.6). This observation is understood by the fact that the

Na+ ion (1.02 Å for CN = 6) is closer in size to Ca2+ and Sr2+ than the Li+ ion (0.74 Å for

CN = 6) [17], i.e. a (Ce-Na) pair matches better with Ca or Sr than a (Ce-Li) pair, which

can enhance the solubility of Ce3+. With the radius of the M cations increasing, the ratio

of the emission intensity to the absorption intensity (at 400 nm) of Ce3+ decreases

obviously because a larger size difference between M2+ and Ce3+ ions makes the

solubility of Ce3+ in the host lattice decrease from Ca, Sr to Ba.

Finally, it is worthwhile to mention that the absorption or the excitation bands of

Ca2Si5N8:Ce, Li and Sr2Si5N8:Ce, Li perfectly match with the blue light source of

(In,Ga)N-based LEDs in the range of 370 – 450 nm, so in combination with other

phosphors these materials are capable of generating white-light.

1.4. Conclusions

The luminescence properties of M2-2xCexAxSi5N8 (0≤ x ≤0.1) (M = Ca, Sr, Ba; A = Li,

Na) were studied. With increasing the size differences between Ce3+ and M2+, the

solubility limit of Ce3+ decreases from x ≈ 0.05 for Ca2Si5N8 and Sr2Si5N8 to x ≤ 0.02 for

Ba2Si5N8. In all the Ce3+-doped alkaline-earth silicon nitride compounds, broad bands

with different amount of sub-bands are observed corresponding to 4f1 4f05d 1

transitions of Ce3+. Two distinct emission centers are found for Ce3+ ions in M2Si5N8: Ce,

A (M = Sr, Ba; A = Li, Na), of which the emission maxima are located at about 495 and

553 nm (A = Li); 520 and 556 nm (A = Na) for Sr2Si5N8, and 451 and 560 nm for

Ba2Si5N8. In contrast, only a narrow symmetric emission band (~ 470 nm) can be found

for Ca2Si5N8 due to similar environment around Ce3+ substituted on the two Ca2+ sites.

With the ionic radius of M2+ increasing going from Ca to Ba, the ratio of the emission

intensity to the absorption intensity (at 400 nm) of Ce3+ decreases. Except for Ba2Si5N8,

the emission intensity can be enhanced by using Na+ instead of Li+ as charge

compensator resulting from a higher solubility of Ce3+ in M2Si5N8 (M = Ca, Sr). A strong

absorption and excitation band in the UV-blue range of 370 – 450 nm in Ca2Si5N8:Ce3+

and Sr2Si5N8:Ce3+ demonstrates these materials to be of high potential for white LED

lighting applications.

26

Chapter 1

References

1. G. Blasse and B.C. Grabmaier, Luminescent materials, Spring-Verlag, Berlin, 1994

2. P. Dorenbos, Phys. Rev., 2001, B64, 125117.

3. N. Yamashita, Y. Michitsuji and S. Asano, J. Electrochem. Soc., 1987, 134, 2932.

4. C. Feldmann, T. Justel, C. R. Ronda and P. J. Schmidt, Adv. Funct. Mater., 2003, 13, 511.

5. J.W.H. van Krevel, Ph.D. thesis, Eindhoven University of Technology, 2000

6. J.W.H. van Krevel, H.T. Hintzen, R. Metselaar and A. Meijerink, J. Alloys and

Comp., 1998, 268, 272

7. J.W.H. van Krevel, H.T. Hintzen and R. Metselaar, Mater. Res. Bull., 2000, 35, 747.

8. J.W.H. van Krevel, J.W.T. van Rutten, H. Mandal, H.T. Hintzen and R. Metselaar,

J. Solid State Chem., 2002, 165, 19.

9. R.J. Xie, N. Hirosaki, M. Mitomo, Y. Yamamoto, T. Suehiro and N. Ohashi, J. Am.

Ceram. Soc., 2004, 87, 1368.

10. D. de Graaf, H.T. Hintzen and G. de With, J. Lumin., 2003, 104, 131.

11. Y.Q.Li, G. de With and H.T. Hintzen, J. Alloys and Comp., 2004, 385, 1.

12. Y.Q.Li, C.M. Fang, G. de With and H.T. Hintzen, J. Solid State Chem., 2004, 177,

4687.

13. H.T. Hintzen, J.W.H. van Krevel and G. Botty, EP 1104 799 A1, 1999.

14. H.A. Hoppe, H. Lutz, P. Morys, W. Schnick and A. Seilmeier, J. Phys. Chem. Solids,

2000, 61, 2001

15. T. Schlieper, W. Milius and W. Schnick, Z. Anorg. Allg. Chem,, 1995, 621, 1380.

16. T. Schlieper and W. Schnick, Z. Anorg. Allg. Chem., 1995, 621, 1037.

17. R.D. Shannon, Acta Cryst., 1976, A32, 751.

18. M.J.J. Lammers, H.C.G. Verhaar and G. Blasse, Mater. Chem. Phys., 1986, 16, 63.

19. P. V. Kelsey, Jr. and Jesse J. Brown, Jr, J. Electrochem. Soc., 1976, 123, 1384.

20. T.R.N. Kutty, Mater. Res. Bull., 1990, 25, 343.

27

Chapter 2

Luminescence properties of red-emitting M2Si5N8:Eu2+

(M = Ca, Sr, Ba) LED conversion phosphors ABSTRACT The influence of the type of the alkaline-earth ion and the Eu2+ concentration on the

luminescence properties of Eu2+-doped M2Si5N8 (M = Ca, Sr, Ba) has been investigated.

XRD analysis shows that Eu2+-doped Ca2Si5N8 forms a limited solid-solution with a

maximum solubility about 7 mol% having a monoclinic lattice. The Eu2+ ion can be

totally incorporated into Sr2Si5N8 and Ba2Si5N8 forming complete solid-solutions with

orthorhombic lattices. M2Si5N8:Eu2+ (M = Ca, Sr) shows typical broad band emission in

orange to red spectral range (600 – 680 nm) depending on the type of M and the

europium concentration. Ba2Si5N8:Eu2+ shows yellow to red emission with maxima from

580 to 680 nm with increasing the Eu2+ contents. The long-wavelength excitation and

emission is attributed to the effect of a high covalency and a large crystal field splitting

on the 5d band of Eu2+ in the nitrogen environment. Both the luminescence intensity and

the conversion efficiency increase going from Ca to Ba and Sr under excitation at 465 nm.

With increasing the europium concentration, the emission band shows a red-shift in all

M2Si5N8:Eu2+ compounds due to the Stokes shift and the reabsorption process by Eu2+.

M2Si5N8:Eu2+ demonstrates to be highly promising red-emitting phosphors for LED

applications.

Keywords: luminescence, alkaline-earth-silicon-nitride, phosphor, europium, X-ray

powder diffraction, white-light LEDs.

28

Chapter 2

2.1. Introduction

A lighting revolution is sweeping all the over world and is stealthily coming in and

improving our everyday life. In comparison with incandescent and fluorescent lamps, the

InGaN-based white light-emitting diode (LEDs) has many advantages in

energy-efficiency, long-lifetime, compact, environment friendly and designable [1-3].

Excitingly, the efficiency of white LED lighting has already exceeded the incandescent

lamps and now is competitive with fluorescent lamps [1-5]. Without doubt, the white

LED lighting is setting foot in the lighting industry and greatly challenges the

conventional lighting.

In this solid-state lighting innovation, the wavelength conversion phosphor materials

play a crucial role as they once did in fluorescent lamps [6-9]. White LED lighting within

a phosphor–LED system can be realized by several approaches: First, a combination of

an InGaN based blue-LED chip (emitting blue light at 465 nm) with a yellow phosphor

(i.e. YAG:Ce3+ based materials) [10-11]. Second, a blue-LED chip combined with a

green- (~ 530 nm) and a red-emitting (> 600 nm) phosphor [12] instead of single yellow-

emitting phosphor. The two phosphors absorb the blue light from the InGaN chip and

convert it into green and red light and then by color mixing the white light is generated.

In addition, using an UV-LED plus blue-, green- and red- emitting phosphors also can

reach this purpose [12]. In comparison with the former, the latter two ways provide

improved colour rendering and a wide range of color temperatures. As conversion

phosphors they must strongly absorb UV-blue light (370 - 450 nm) and efficiently re-emit

in the red, green or blue part of the spectrum. However, the well-developed phosphors

used for current mercury gas-discharge fluorescent lamps can not be directly applied to

white LED lighting because of a very low absorption in the UV-blue range (370 – 450

nm). So far, only a limited number of phosphors is available, for example yellow-emitting

YAG:Ce3+ [1, 13], green-emitting SrGaS4:Eu2+ and red-emitting Sr1-xCaxS:Eu2+ based

phosphors [12]. With respect to the above mentioned phosphor materials, either they have

29

Chapter 2

low luminous efficacy (i.e. YAG:Ce3+) [3] or low stability against humid, thermal and

radiative environments (i.e. Sr1-xCaxS:Eu2+) [3, 12] which significantly reduces the

quality of the output light and the service lifetime of LEDs [14]. Therefore, it is urgent to

develop novel conversion phosphors with improved properties to be used for white LED

lighting for general illumination.

Among the green-, yellow- and red-emitting phosphors, the red-emitting phosphors

are the most urgent ones to be improved. Apart from Sr1-xCaxS:Eu2+ based phosphors,

traditionally powerful red-emitting phosphors with sharp line-emission, i.e. Y2O3:Eu3+,

can not serve this function in white LED lighting [1, 15]. Recently, the invention of a new

class of red-emitting phosphors M2Si5N8:Eu2+ (M = Ca, Sr, Ba) [16, 17] has triggered a

renewal in the field of luminescent materials. Subsequently, a number of new

nitride-based phosphors have been found and investigated in most recently years [3, 16,

18-24]. Van Krevel et al. [17] reported unusual long-wavelength Eu2+ emission (620 –

660 nm) with absorption bands in the visible range in M2Si5N8:Eu2+ (M = Ca, Sr, Ba).

The long-wavelength emission is attributed to a large covalency and crystal-field splitting

effect on the Eu2+ 5d band due to the presence of nitrogen [17]. Later, Höppe et al. [24]

studied the luminescence properties in a series of Ba2-xEuxSi5N8 compounds which

revealed two emission maxima peaking at about 600 nm, corresponding to two

crystallographic Ba(Eu) sites in the Ba2Si5N8 host lattice. The emission maxima shift to

longer wavelength with increasing Eu-content due to reabsorption processes of Eu2+ [24].

With respect to the above mentioned investigations, there still remain a number of

questions, for example, in order to further improve the luminescence properties, the

solubility of Eu2+ and the influence of impurities in M2Si5N8 should be clarified.

Moreover, the influence of the type of M ion and the Eu concentration on the

luminescence properties are important issues.

In the present study, we focus on the investigation of the solubility of Eu2+ in the

M2Si5N8 host-lattice and the influence of the type of the M ions and the Eu2+

30

Chapter 2

concentration on the luminescence and efficiency of M2Si5N8:Eu2+.

2.2. Experimental

2.2.1. Starting materials

The binary nitride precursors SrNx (x ≈ 0.6 – 0.66), BaNx (x ≈ 0.6 – 0.66) and EuNx

(x ≈ 0.94) were pre-prepared by the reaction of the pure strontium metal (Aldrich, 99.9%,

pieces), barium metal (Aldrich, 99.9%, pieces) and Eu metal (Csre, 99.9%, lumps) under

flowing dried nitrogen at 800, 550, and 800 °C, respectively, for 8 – 16 h in horizontal

tube furnaces. In addition, calcium nitride powder Ca3N2 (Alfa, 98%) and α-Si3N4

powder (Permascand, P95H, α content 93.2%; Oxygen content: ~1.5%) are used as the

as-received raw materials.

2.2.2. Synthesis of undoped and Eu2+-doped M2Si5N8

Polycrystalline M2-xEuxSi5N8 (0 ≤ x ≤ 0.2 for M = Ca, 0 ≤ x ≤ 2.0 for M = Sr, Ba)

powders were prepared by a solid state reaction method at high temperature. The Ca3N2,

SrNx, BaNx and EuNx as well as α-Si3N4 powders were weighed out in the appropriate

amounts and subsequently mixed and ground together in an agate mortar. The powder

mixtures were then transferred into molybdenum crucibles. All processes were carried out

in a purified-nitrogen-filled glove-box. Subsequently those powder mixtures were fired

twice (with a medium grinding in between) in a horizontal tube furnace at 1300 – 1400 oC for 12 and 16 h, respectively, under flowing 90%N2-10%H2 atmosphere. After firing,

the samples were gradually cooled down in the furnace. There was no apparent reaction

of the prepared nitrides with the Mo crucibles.

2.2.3. X-ray diffraction data collection and structure refinement

All measurements were performed on finely ground samples, which were analyzed

31

Chapter 2

by X-ray powder diffraction (Rigaku, D/MAX-B) using Cu-Kα radiation at 40 kV and 30

mA with a graphite monochromator.

The phase formation of undoped and Eu2+-doped M2Si5N8 materials were examined

using a routine scan (2 °/min). Lattice parameters determination was carried out by the

least-squares method from X-ray diffraction data collected in step scan with a step size of

0.01o in 2θ and a count time of 10 seconds between 10 and 90° 2θ using silicon powder

as an internal standard. Structure refinement was carried out by the Rietveld method [25],

using the program GSAS [26, 27]. The structural parameters of M2Si5N8 (M = Ca, Sr, Ba)

[28, 29] were used as the initial parameters for structural refinement of Eu2+-doped

M2Si5N8 (M = Ca, Sr, Ba).

2.2.4. Optical measurements

The diffuse reflectance, emission and excitation spectra of the samples were obtained

at room temperature by a Perkin Elmer LS 50B spectrophotometer equipped with a Xe

flash lamp. The reflection spectra were calibrated with the reflection of black felt

(reflection 3%) and white barium sulfate (BaSO4, reflection ~100%) in the wavelength

region of 230-700 nm. The excitation and emission slits were set at 5 nm. The emission

spectra were corrected by dividing the measured emission intensity by the ratio of the

observed spectrum of a calibrated W-lamp and its known spectrum from 300 to 900 nm.

Excitation spectra were automatically corrected for the variation in the lamp intensity by

a second photomultiplier and a beam-splitter. All the spectra were measured with a scan

speed of 100 nm/min.

The relative luminescence intensity (Φem) was the quotient of the area under the

emission curve of the sample vs. the related area of a standard material. The conversion

efficiency of the luminescence was determined by the following equation (1):

4

em emc

abs BaSiO ref

η Φ Φ= =Φ Φ −Φ

(1)

32

Chapter 2

Where the absorbed radiation (Φabs) is the difference between incident radiation (Φ0) and

reflected radiation (Φref). Φ0 is determined by the reflection spectrum of a white reference

standard e.g. BaSO4 with λexc = 465 nm. In addition, a Sbcose yellow-orange-emitting

phosphor (Sr1-x-yBaxCaySiO4:Eu2+) and a red-emitting Sarnoff (Ca1-xSrxS:Eu2+) phosphor

is used as the standard materials for the measurement of the orange- and red-emitting

phosphors.

2.3. Results and discussion

2.3.1. Phase formation and the solubility of Eu2+ in M2Si5N8 (M = Ca, Sr, Ba)

Like in the case of M2Si5N8:Ce3+, Li+ [23], a small amount of CaSi2O2N2-like phase

in Ca2-xEuxSi5N8 and an unknown second phase in Ba2-xEuxSi5N8 were detected. While

Sr2-xEuxSi5N8 is a nearly single-phase compound. Fig. 2.1 shows the unit cell volume

dependence on the fraction of x in M2-xEuxSi5N8. For Ca2-xEuxSi5N8, as expected, the unit

cell volume expands with increasing the Eu2+ concentration in the range of x = 0 to 0.14

(i.e. 0 – 7 mol% with respect to Ca), which is consistent with the fact that Eu2+ is larger in

size than Ca2+ [30]. When the x value surpasses 0.14, the unit cell volume keeps constant

corresponding to the appearance of impurity phases. This implies that the maximum

solubility of Eu2+ in Ca2Si5N8 is around x = 0.14 (i.e. 7 mol%) (Fig. 2.1(a)). In contrast,

the unit cell volume of M2-xEuxSi5N8 (M = Sr, Ba) nearly linearly decreases with an

increase of x due to the replacement of Sr2+ or Ba2+ with the smaller Eu2+ ion [30],

perfectly following Vegard’s law in the two systems (Fig. 2.2 (b)-2.2 (c)). In addition, the

lattice shrinkage for Ba is larger than that for Sr when Eu2+ is incorporated. These

observations can be well explained by their different ionic radii and structural

characteristics. In the case of Ca2-xEuxSi5N8, the ionic radius of Eu2+ (1.17 Å, CN = 6 [30])

is larger than that of Ca2+ (1.00 Å, CN = 6 [30]) by about 15% and the two end-members

Ca2Si5N8 and Eu2Si5N8 compounds have different crystal structures (i.e. monoclinic vs.

33

Chapter 2

0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20720

721

722

723

724

725

Uni

t cel

l vol

ume

(Å3 )

x

(a)

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2361.0

361.5

362.0

362.5

363.0

363.5

364.0

364.5

365.0

Uni

t cel

l vol

ume

(Å3 )

x

(b)

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2

360

362

364

366

368

370

372

374

376

378

Uni

t cel

l vol

ume

(Å3 )

x

(c)

Fig. 2.1. The relationship between the unit cell volume of M2-xEuxSi5N8 and the x value,

(a) M = Ca, (b) M = Sr, (c) M = Ba.

34

Chapter 2

orthorhombic crystal system), therefore, the formation of a limited solid-solution is

expected. The M2Si5N8 (M = Sr, Ba) and Eu2Si5N8 compounds are isostructural with

orthorhombic crystal system [31]. Thus the formation of a complete solid-solution

between Sr2Si5N8 - Eu2Si5N8 and Ba2Si5N8 – Eu2Si5N8 can be understood. Furthermore,

the above mentioned conclusions are confirmed by their luminescence properties (see

sections 2.3.2 – 2.3.3).

No site preference of Eu2+ can be found by the Rietveld structure refinement based

on the powder XRD data for M2Si5N8:Eu2+ (M = Sr, Ba). It can be explained by the fact

that M2Si5N8 (M = Sr, Ba, Eu) are isostructural and M occupies two nonequivalent

crystallographic sites in Eu2Si5N8 [28, 29, 31]. For Ca2Si5N8:Eu2+, the Eu2+ ion shows a

slight site preference for the larger Ca site (i.e. Ca2) with the site occupancy of 0.6 for the

Eu concentration of 5 mol%. Therefore, the Eu2+ ions can be regarded as equally

distributed over the two M crystallographic sites in M2Si5N8, in agreement with their

luminescence properties (see section 2.3.3).

2.3.2. Diffuse reflection spectra

Fig. 2.2 shows the diffuse reflection spectra of undoped and Eu2+-doped M2Si5N8 (M

= Ca, Sr, Ba). All undoped M2Si5N8 compounds show a remarkable drop in reflection in

the UV range around 300 nm with an estimated band gap at about 250 nm for M = Ca,

265 nm for M = Sr and 270 nm for M = Ba, corresponding to the valence-to-conduction

band transitions of the M2Si5N8 host lattice. The intense reflection in the visible spectral

range is in agreement with the observed grey-white daylight color of undoped M2Si5N8.

Obviously, two broad absorption bands can be seen from the reflection spectra of low Eu

concentration samples (Fig. 2.2). The very broad absorption band (370 – 490 nm) is

attributed to the absorption of the Eu2+ ions due to the absence of such band in undoped

M2Si5N8. The intensity of this absorption increases for higher Eu concentration (Fig. 2.2).

Apart from this main absorption band in the visible range, a short-wavelength absorption

35

Chapter 2

200 300 400 500 600 700

0

20

40

60

80

Ref

lect

ion

(%)

Wavelength (nm)

x 0.00 0.02 0.10 0.14 0.20

(a)

200 300 400 500 600 700

0

20

40

60

80

x 0.00 0.02 0.10 0.20 0.40 0.80 1.20 1.60 2.00

Ref

lect

ion

(%)

Wavelength (nm)

(b)

36

Chapter 2

200 300 400 500 600 700

0

20

40

60

80

100

x 0.00 0.02 0.04 0.10 0.20 0.40 2.00

Ref

lect

ion

(%)

Wavelength (nm)

(c)

Fig. 2.2. Diffuse reflection spectra of M2-xEuxSi5N8, (a) M = Ca, (b) M = Sr, (c) M = Ba.

band centered at about 300 nm is also observable in M2Si5N8:Eu2+ which is also ascribed

to the absorption of Eu2+ because with increasing Eu2+ concentration the intensity of this

absorption band also increases. In contrast to the undoped samples, the daylight color of

Eu2+-doped M2Si5N8 shows orange to red for M = Ca, Sr and yellow to red for M = Ba

varying with the Eu concentration as a result of a strong absorption in the visible range

around 370 – 490 nm. Additionally, the onset of the reflection drop significantly shifts to

a longer-wavelength as the Eu concentration increases indicating that the absorption

range can be tailored by the Eu content (Fig. 2.2). The absorption edge continuously

shifts over the whole range 0 < x ≤ 2 to longer wavelength from 490 – 608 nm for

Sr2Si5N8:Eu2+ (Fig. 2.2b) and 460 – 608 nm for Ba2Si5N8:Eu2+ (Fig. 2.2c) in agreement

with the conclusion that Eu is totally incorporated into the Sr2Si5N8 and Ba2Si5N8 lattices.

For Ca2Si5N8:Eu2+, however, the red-shift of the absorption edge is found to be negligible

37

Chapter 2

for x = 0.2 as compared to x = 0.14 (Fig. 2.2a), in agreement with the solubility limit at x

≈ 0.14.

2.3.3. Photoluminescence properties of Eu2+ in M2Si5N8

Roughly speaking, there are five broad bands in the excitation spectra of

M2Si5N8:Eu2+ (M = Ca, Sr, Ba) (Table 2.1 and insets in Fig. 2.3). The position of these

excitation bands is almost independent of the type of the M ions, the Eu concentration

and the crystal structure, peaking at about 250, 300, 340, 395 and 460 nm (Fig. 2.3). Only

a small variation for various M ions can be observed, in agreement with the observed

diffuse reflection spectra. The first excitation band at ~ 250 nm is readily assigned to the

host lattice excitation (e.g. transition from the valence to conduction band for the

M2Si5N8 host lattices). The remaining excitation bands clearly originate from the 4f7

4f65d1 transition of Eu2+. Similarly, the most intense 5d excitation band of Eu2+ is located

at about 395 nm in M2Si5N8:Eu2+. The lowest energy levels of the 5d excitation band

(very broad at about 420 – 520 nm) seem to be further decomposed into two or three

subbands, especially for M = Sr, Ba at higher Eu concentrations. The 5d excitation band

of the Eu2+ ions at lower energy (> 400 nm) is attributed to the influence of highly

covalent bonding of MEu-N and a large crystal-field splitting due to the presence of

nitrogen [17, 18]. In addition, with increasing Eu2+ concentration these subbands at

longer wavelength become more intense corresponding to a degradation of the dominant

excitation band at about 395 nm, which suggests that Eu2+ self-absorption possibly occurs

for higher Eu concentration. Moreover, the absence of significant changes in the position

and shape of the excitation band suggests that the covalency of the Eu-N bonds and the

crystal field strength around the Eu2+ ions are very similar in the M2Si5N8 series (Table

2.1). For Eu2+ ions occupying two M sites more than four 5d excitation bands should be

observed. However, the appearance of only four obvious bands implies that the 5d

excitation bands of Eu2+ do serious overlap at room temperature and possibly some of

38

Chapter 2

them are hidden in the conduction band of the M2Si5N8 host lattice, similar to the case of

Ce3+-doped M2Si5N8 [23].

400 450 500 550 600 650 700 750 800 850 900

0.0

0.2

0.4

0.6

0.8

1.0

200 250 300 350 400 450 500 5500

100

200

300

400

500

600

Emis

sion

inte

nsity

(a.u

.)

Wavelength (nm)

x 0.02 0.10 0.14

x 0.02 0.10 0.14

Exci

tatio

n in

tens

ity (a

.u.)

Wavelength (nm)

(a)

500 550 600 650 700 750 800 850 900 950

0.0

0.2

0.4

0.6

0.8

1.0

200 250 300 350 400 450 500 550 600

50

100

150

200

Emis

sion

inte

nsity

(a.u

.)

Wavelength (nm)

x 0.02 0.10 0.20 0.40 0.80 1.20 1.60 2.00

x 0.02 0.10 0.20 0.40 0.80 1.20 1.60 2.00

Exci

tatio

n in

tens

ity (a

.u.)

Wavelength (nm)

(b)

39

Chapter 2

450 500 550 600 650 700 750 800 850 900

0.0

0.2

0.4

0.6

0.8

1.0

200 250 300 350 400 450 500 550

100

200

300

400

500

Inte

nsity

(a. u

.)

Wavelength (nm)

x 0.02 0.04 0.10 0.20 0.80 2.00

Inte

nsity

(a.u

.)

Wavelength (nm)

x 0.02 0.04 0.10 0.20 0.80 2.00

(c)

Fig. 2.3. Excitation (inset) and emission spectra of M2-xEuxSi5N8: (a) M = Ca, (b) M = Sr,

(c) M = Ba.

Here, it is worth noting that M2Si5N8:Eu2+ has not only high absorption but also

efficient excitation in the same spectral region of 400 – 470 nm (see Fig. 2.3), perfectly

matching with the radiative blue-light from the InGaN-based LEDs (~ 465 nm).

Eu is present as the divalent ion in all Eu-doped M2Si5N8 due to the absence of sharp

f-f transition line characteristic for Eu3+ in the excitation spectrum with the broad-band

emission characteristic (Fig. 2.3). As a result, this red emission is essentially assigned to

the 4f65d1 4f75d0 transition of the Eu2+ ion [6]. The position of the Eu2+ emission band

is strongly dependent on the type of M ion and the Eu concentration (see color point in

Fig. 2.4). Among M2Si5N8:Eu2+, Sr2Si5N8:Eu2+ is more close to the Sarnoff-red reference

sample (Sr1-xCaxS:Eu2+) (Fig. 2.4). For Ca2Si5N8:Eu2+, the emission maximum shifts to

long wavelength from 605 to 615 nm with increasing the Eu concentration (Fig. 2.3(a)).

This red-shift terminates at about x = 0.14 (i.e. 7 mol % Eu2+) in agreement with the

40

Chapter 2

maximum solubility of Eu2+ (~ 7 mol %) in Ca2Si5N8 (Fig. 2.1 (a)). A very small red-shift

(~ 10 nm) of the emission band is fairly consistent with a small amount of Ca2+

replacement by Eu2+. For M2Si5N8:Eu2+ (M = Sr, Ba), the emission band of Eu2+

successively shifts from orange for M = Sr and yellow for M = Ba at low Eu

concentration to the red spectral region for high Eu concentration up to a maximum

wavelength of ~ 680 nm which perfectly agrees with a linear decrease of the unit cell

volume with varying Eu content for both M = Sr and M = Ba (Figs. 3(b)-3(c)). The

red-shift in Ba2Si5N8:Eu2+ (~ 107 nm) is larger than in Sr2Si5N8:Eu2+ (~ 71 nm) also in

agreement with a large shrinkage of the unit cell of Ba2Si5N8 due to the incorporation of

the smaller Eu2+ ion [30].

0.45 0.50 0.55 0.60 0.65 0.70 0.750.25

0.30

0.35

0.40

0.45

0.50

0.55

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.80.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

x=1.6

x=0.80x=1.2

x=0.40

x=0.40

620 nm

600 nm

x=0.20

x=0.02

x=0.02

x=0.10

x=0.20

x=0.02

590 nm

585 nm

580 nm

Y

X

M2-xEuxSi5N8

M = Ba M = Sr M = Ca Sbcose-F593 Sarnoff-Red

λexc= 465 nm

Fig. 2.4. CIE colour coordinates of the emission band of M2-xEuxSi5N8 (λexc = 465 nm).

41

Chapter 2

In M2Si5N8 (M = Ca, Sr, Ba), there are two crystallographic M sites [28, 29].

Therefore, if the Eu ions occupy two M sites as found with the Rietveld refinement to be

the case, theoretically two emission bands should be observed. However, the presence of

only a single broad emission band suggests that the environment of both Eu2+ ions is very

similar or the Eu2+ ions are not very sensitive to the changes of the local structure,

eventually resulting in large overlap of the two emission band of Eu2+ in M2Si5N8:Eu2+,

especially for higher-Eu content (Fig. 2.3). This is consistent with the fact that varying

the excitation wavelength yields similar emission spectra. However, after Gaussian

deconvolution on an energy scale, these broad emission bands can be well decomposed

into two Gaussian components for M2Si5N8:Eu2+ (5 mol%) as typical examples (Fig. 2.5).

Our results are consistent with the observation of Höppe et al., they observed two

emission bands for Ba1.89Eu0.11Si5N8 corresponding to the two crystallographic BaEu sites

in Ba2Si5N8 [24]. Evidently, for Ba2Si5N8:Eu2+ (5 mol%) the emission band is

significantly broad (FWHM ~ 125 nm) due to the largest difference of two BaEu sites

compared to Ca2Si5N8:Eu2+ (FWHM ~ 104 nm) and Sr2Si5N8:Eu2+ (FWHM ~ 88 nm) (Fig.

2.5).

The Stokes shift increases with increasing the Eu concentration in all cases. The

changes of the Stokes shift as function of the Eu concentration are about 300, 1200 and

2100 cm-1 for M = Ca, Sr and Ba, respectively. Based on the fact that the center of gravity

and crystal-field splitting have no significant change, therefore, the observed red-shift of

the emission band of M2Si5N8:Eu2+ with increasing the Eu concentration can be mainly

ascribed to an increase of the Stokes shift for M = Sr, Ba (Fig. 2.3). Here we also cannot

exclude the possibility of the reabsorption by Eu2+.

Although the excitation spectra are almost independent of the M type, the position of

the emission bands is strongly dependent on the M type. For example, the emission bands

of M2Si5N8:Eu2+ (1 mol%) peak at about 605, 610 and 574 nm for M = Ca, Sr and Ba,

respectively. In addition, the relative luminescence intensity is about 71%, 87% and 84%

42

Chapter 2

450 500 550 600 650 700 750 800 850

0.0

0.2

0.4

0.6

0.8

1.0

450 500 550 600 650 700 750 800 850

0.0

0.2

0.4

0.6

0.8

1.0

450 500 550 600 650 700 750 800 850

0.0

0.2

0.4

0.6

0.8

1.0

22 21 20 19 18 17 16 15 14 13 12

22 21 20 19 18 17 16 15 14 13 12

22 21 20 19 18 17 16 15 14 13 12

(c)

659

Ca2Si5N8:Eu2+ 5%

Wavelength (nm)

605

(b)

657

615

Sr2Si5N8:Eu2+ 5%

Emis

sion

inte

nsity

(a.u

.)

646

594

Ba2Si5N8:Eu2+ 5%(a)

Wavenumber (cm-1)

Fig. 2.5. Observed (solid) and fitted (dashed) emission spectra and decomposed Gaussian

components (dotted) for M2Si5N8:Eu (5 mol%), (a) M = Ca, (b) = Sr, (c) = Ba (λexc = 395

nm).

for M = Ca, Sr and Ba, respectively, with an excitation wavelength at 465 nm. The same

tendency is also found for the conversion efficiency (Table 2.1). The same observation

also holds for high Eu concentration which generally shows a lower efficiency for

Ca2Si5N8:Eu2+ and the higher efficiency for Sr2Si5N8:Eu2+ and Ba2Si5N8:Eu2+.

43

Chapter 2

Table 2.1. Luminescence data for Eu-doped M2-xEuxSi5N8 (M = Ca, Sr, Ba)

M2-xEuxSi5N8

M = Ca M = Sr M = Ba

Crystal system Monoclinic Cc Orthorhombic Pmn21 Orthorhombic Pmn21

Maximum solubility of Eu2+ x = 0.14 x = 2.0 x = 2.0

Excitation band (nm)* 297, 355, 394, 460, 496 294, 334, 395, 465, 505 295, 334, 395, 460, 504

Emission band (nm) 605 - 615 609 - 680 570 - 680

Center of gravity (cm-1)* 1 25800 26000 26100

Crystal field splitting (cm-1)* 2 13500 14200 14100Stokes shift (cm-1)* 3Luminescence intensity# 4

Conversion efficiency# 4

3800 71% 26%

3700 87% 37%

3500 84% 36%

* x = 0.10; # x = 0.02 1. Center of gravity calculated from averaging the energies of the observed 5d excitation levels of Eu2+. 2. Stokes shift calculated from the energy difference between the lowest 5d excitation band and emission band of Eu2+. 3. Crystal-field splitting estimated from the energy difference between highest and lowest observed 5d excitation levels of Eu2+

4. The conversion efficiency of the Sbcose and Sarnoff phosphors is 42% and 41%, respectively.

44

Chapter 2

2.4. Conclusions

Undoped and Eu2+-doped M2Si5N8 (M = Ca, Sr, Ba) materials were obtained by a

solid-state reaction of MNx, EuNx and α-Si3N4 powder at 1300 – 1400 °C under a N2-H2

(10%) atmosphere. The maximum solubility of Eu2+ is about 7 mol% in the Ca2Si5N8

lattice. In contrast, Eu2+ can be completely incorporated into the M2Si5N8 (M = Sr, Ba)

lattice because the M2Si5N8 (M = Sr, Ba) compounds are isostructural with Eu2Si5N8 in

which Eu2+ is statistically distributed over the two M sites. Eu-doped M2Si5N8 shows a

typical broad band emission of Eu2+ in the spectral range from orange to red (580 – 680

nm). With increasing Eu concentration, the emission band shifts to long-wavelength

depending on the type of M ion and the Eu concentration which is found to be proportional

to the changes of the unit cell volume for M = Sr, Ba. This red-shift possibly can be

attributed to an increase of the Stokes shift and reabsorption by Eu2+. Both the relative

luminescence intensity and the conversion efficiency of M2Si5N8:Eu2+ (M = Sr, Ba) are

higher than that of Ca2Si5N8:Eu2+ excited by 465 nm. The absorption and excitation bands

of M2Si5N8:Eu2+ ranging from 370 to 460 nm perfectly match with the radiation of the

InGaN based LEDs showing high potential for white-LED applications.

References:

1. S. Nakamura, Appl. Phys. Lett., 1994, 64, 1687.

2. S. Aanegola, J. Petroski and E. Radkov, SPIE, 2003, 10, 16.

3. Y. Narukawa, Optics & Photonics News, 2004, 4, 25.

4. L.S. Rohwer, A.M. Srivastava, The Electrochemical Society Interface, 2003, 36.

5. J.Y. Taso, Ed., in Light Emitting Diodes (LEDs) for General Illumination Update

2002 Optoelectronics Industry Development Association, Washington, DC, (2002).

6. G. Blasse and B.C. Grabmaier, Luminescent materials, Springer-Verlag, Berlin,

1994.

7. C. Feldmann, T. Justel, C. R. Ronda and P. J. Schmidt, Adv. Funct. Mater., 2003,

13, 511.

8. Keith H. Butler, Fluorescent Lamp Phosphors, Technology and Theory,

The Pennsylvania State University Press, University Park and London, PA 1981.

45

Chapter 2

9. T. Jüstel, H. Nikol, C. Ronda, Angew. Chem. Int. Ed., 1998, 37, 3084.

10. P. Schlotter, J. Baur, Ch. Hielscher, M. Kunzer, H. Obloh, R. Schmidt and J.

Schneider, Mater. Sci. Eng., 1999, B59, 390.

11. S. Nakamura and G. Fasol, The Blue Laser Diodes, GaN Based Light Emitters and

Lasers, Springer, Berlin, 1997. pp. 216–221.

12. R. Mueller-Mach, G.O. Mueller, M.R. Krames, J. IEEE, 2002, 8, 339.

13. G. Blasse and A. Bril, Appl. Phys. Lett., 1967, 11, 53.

14. M. Yamada, T. Naitou, K. Izuno, H. Tamaki, Y. Murazaki, M. Kameshima and

T. Mukai, Jpn. J. Appl. Phys., 2003, 42, L20.

15. S. Neeraj, N. Kijima, A.K. Cheettham, Solid State Comm., 2004, 131, 65.

16. H.T. Hintzen, J.W.H. van Krevel and G. Botty, EP-1104 799 A1, 1999.

17. J.W.H. van Krevel, Ph.D. Thesis, Eindhoven University of Technology, 2000.

18. J.W.H. van Krevel, H.T. Hintzen, R. Metselaar, and A. Meijerink,

J. Alloys Comp., 1998, 268, 272

19. J.W.H. van Krevel, H.T. Hintzen, R. Metselaar, Mater. Res. Bull., 2000, 35, 747.

20. J.W.H. van Krevel, J.W.T. van Rutten, H. Mandal, H.T. Hintzen, and R. Metselaar,

J. Solid State Chem., 2002, 165, 19.

21. Y.Q. Li, G. de With and H.T. Hintzen, J. Alloys Comp., 2004, 385, 1.

22. Y.Q. Li, C.M. Fang, G. de With and H.T. Hintzen, J. Solid State Chem., 2004, 177,

4687.

23. Y.Q. Li, G. de With and H.T. Hintzen, J. Lumin., 2005, in press.

24. H.A. Höppe, H. Lutz, P. Morys, W. Schnick and A. Seilmeier, J. Phys. Chem.

Solids, 2000, 61, 2001

25. H.M. Rietveld, J. Appl. Cryst., 1969, 2, 65.

26. A.C. Larson and R.B. Von Dreele, Report LAUR 86-748, Los Alamos National

Laboratory, Los Alamos, NM, 2000.

27. B. H. Toby, J. Appl. Cryst. 2001, 34, 210.

28. T. Schlieper, W. Milius and W. Schnick, Z. Anorg. Allg. Chem, 1995, 621, 1380.

29. T. Schlieper and W. Schnick, Z. Anorg. Allg. Chem., 1995, 621, 1037.

30. R.D. Shannon, Acta Cryst., 1976, A 32, 751.

31. H. Huppertz and W. Schnick, Acta Cryst., 1997, C53, 1751.

46

Chapter 3

The effect of replacement of Sr by Ca on the structural and

luminescence properties of red-emitting Sr2Si5N8:Eu2+ phosphor

ABSTRACT The influence of the replacement of Sr by Ca on the structural and luminescence

properties of Eu2+-doped Sr2Si5N8 is reported. The Rietveld refinement of the powder

X-ray diffraction data shows that the Ca2+ ion preferentially occupies the larger Sr site in

Sr2Si5N8:Eu2+. Although the excitation spectrum is hardly modified, the position of the

emission band of Eu2+ can be tailored through partial replacement of Sr by Ca in

Sr2Si5N8:Eu2+, resulting in red-emission shifting from 620 to 643 nm. Furthermore, (Sr,

Ca)2Si5N8:Eu2+ shows high potential for white-light LED applications due to a limited

decrease in the conversion efficiency after the introduction of the Ca ion.

Keywords: luminescence, alkaline-earth-silicon-nitride, calcium, strontium, europium,

phosphor, X-ray powder diffraction, Rietveld refinement, white-light LEDs.

47

Chapter 3

3.1. Introduction

Red-emitting M2Si5N8:Eu2+ (M = Ca, Sr, Ba) phosphors are attracting extensive

attention due to its excellent performance for white-LED lighting applications [1-4].

M2Si5N8:Eu2+ (M = Ca, Sr, Ba) shows unusual long-wavelength broad band emission

between 575 to 680 nm depending on the type of the M ion and the Eu concentration [1, 5,

6]. It can be efficiently excited in the UV-blue range (370 – 465 nm) and in this way

convert absorbed UV-blue light from the InGaN-based LED to orange – red light. The

long-wavelength emission is attributed to a high covalency of the host-lattice and a large

crystal-field splitting effect on the Eu2+ 5d band due to the presence of nitrogen [5, 6].

When excited by 465 nm, the conversion (quantum) efficiency is higher for M = Sr, Ba

than M = Ca [7]. As conversion phosphors for use in white-light LEDs, a high chemical,

thermal and radiation stability is necessary for achieving long lifetime of the devices [8].

However, with the ionic radius of M increasing going from Ca to Ba, the alkaline-earth

compound generally tends to become less stable towards O2, H2O and CO2 as well as

elevated temperatures [5]. As evidence of this, the sensitivity towards oxidation of

M2Si5N8:Eu2+ (M = Ca, Sr, Ba) exposed to air at 300 – 600 °C increases in the sequence

of Ca < Sr < Ba [5]. Therefore, partial replacement of Sr or Ba by Ca in M2Si5N8:Eu2+ (M

= Sr, Ba) is expected to improve their stability, as also found for Sr1-xCaxS:Eu2+ [2]. In

addition, it is well-established that the luminescence properties can be tuned by not only

the Eu concentration but also partial cross-substitution between alkaline-earth ions, for

example replacement of Sr by Ca and Ba by Sr [9 - 12]. With respect to M2Si5N8, there

are two crystallographic M sites in the lattice, thus better understanding of whether or not

Ca and Eu ions have site preference in the Sr2Si5N8 lattice is valuable for further

improvement of the performance of Sr2Si5N8:Eu2+ phosphor from the practical and

scientific point of views. In this paper, the focus is on the investigation of the effect of the

substitution of Ca for Sr in Sr2Si5N8:Eu (5mol%) on the structural and luminescence

properties.

48

Chapter 3

3.2. Experimental

3.2.1. Starting materials

The binary nitride precursors SrNx (x ≈ 0.6) and EuNx (x ≈ 0.94) were pre-prepared

by the reaction of the pure strontium metal (Aldrich, 99.9%, pieces) and Eu metal (Csre,

99.9%, lumps) under flowing dried nitrogen at 800°C for 8 – 16 h in a horizontal tube

furnace. In addition, calcium nitride powder Ca3N2 (Alfa, 98%) and α-Si3N4 powder

(Permascand, P95H, α content 93.2%; Oxygen content: ~1.5%) are used as the

as-received raw materials.

3.2.2. Synthesis of undoped M2Si5N8, M1.9Eu0.1Si5N8 and Sr1.3Ca0.6Eu0.1Si5N8

M2Si5N8, M1.9Eu0.1Si5N8 (M = Ca, Sr) and Sr1.3Ca0.6Eu0.1Si5N8 were prepared by a

solid state reaction at high temperature. The Eu concentration is fixed at 5 mol% with

respect to the divalent lattice site. The Ca3N2, SrNx and EuNx as well as α-Si3N4 starting

powders were weighed out, thoroughly mixed and ground together in the appropriate

molar ratio in an agate mortar. The powder mixtures were then transferred into

molybdenum crucibles. All processes were carried out in a purified-nitrogen-filled

glove-box. Subsequently those powder mixtures were fired twice in a horizontal tube

furnace at 1300 – 1400 oC for 12 and 16 h, respectively, under flowing 90%N2-10%H2

atmosphere with an intermediate grinding in between. After firing, the samples were

cooled down in the furnace.

3.2.3. X-ray diffraction data collection and structure refinement

All measurements were performed on finely ground samples, which were analyzed

by X-ray powder diffraction (Rigaku, D/MAX-B) using Cu-Kα radiation at 40 kV and 30

mA with a graphite monochromator. The phase formation of all samples was checked by

a routine scan (2 °/min). For structure refinement, X-ray diffraction data were collected

49

Chapter 3

from 10-120o 2θ at 0.01° intervals, counting for 20 s per step. Structure refinement was

carried out by the Rietveld method [13], using the program GSAS [14, 15]. The structural

parameters of M2Si5N8 (M = Ca, Sr) [16, 17] were used as the initial parameters for

structural refinement of M1.9Eu0.1Si5N8 (M = Ca, Sr) and Sr1.3Ca0.6Eu0.1Si5N8. Site

preferences of Ca2+ and Eu2+ were examined by manually varying the occupancies of Ca

and Eu over two M sites within the stoichiometric constraints.

3.2.4. Optical measurements

The diffuse reflection, emission and excitation spectra of the samples were obtained at

room temperature by a Perkin Elmer LS 50B spectrophotometer equipped with a Xe flash

lamp. The reflection spectra were calibrated with the reflection of black felt (reflection

3%) and white barium sulfate (BaSO4, reflection ~100%) in the wavelength region of

230-700 nm. The excitation and emission slits were set at 5 nm. The emission spectra

were corrected by dividing the measured emission intensity by the ratio of the observed

spectrum of a calibrated W-lamp and its known spectrum from 300 to 900 nm. Excitation

spectra were automatically corrected for the variation in the lamp intensity by a second

photomultiplier and a beam-splitter. All the spectra were measured with a scan speed of

100 nm/min.

The relative conversion efficiency ηc for λexc = 450 nm was estimated from the diffuse

reflection and the emission spectra according to the following equation (1):

0

emc

ref

η Φ=

Φ − Φ (1)

where Φem is the integrated emission intensity at λexc = 450 nm; Φ0 and Φref are the

reflection intensities of undoped and Eu-doped materials, respectively, at a wavelength of

450 nm.

50

Chapter 3

3.3. Results and discussion

3.3.1. Effect of incorporation of Ca2+ on the structural characteristics of

Sr2Si5N8:Eu2+

M1.9Eu0.1Si5N8 (M = Ca, Sr) is obtained as a single phase compound. For (Sr,

Eu)2Si5N8 this is as expected, because Sr2Si5N8 and Eu2Si5N8 are isostructural [16, 17].

For (Ca, Eu)2Si5N8 the Eu concertration is below solubility limit. According to our

preliminary studies, Ca2Si5N8 and Sr2Si5N8 can form a limited solid-solution due to the

fact that Ca2Si5N8 and Sr2Si5N8 have different crystal structures [16, 17] and the

maximum solubility of Ca in Sr2Si5N8 is about 40 mol% with respect to strontium. For

the purpose of the structure investigation, in this work we select the composition

Sr1.3Ca0.6Eu0.1Si5N8. Incorporation of the Ca2+ ions is found to form a nearly single phase

Sr1.3Ca0.6Eu0.1Si5N8 material. As expected, the lattice parameters of Sr1.3Ca0.6Eu0.1Si5N8

are smaller than those of Sr2Si5N8:Eu2+ as a matter of fact that the ionic radius of Ca2+

(1.00 Å, CN = 6) is smaller than that of Sr2+ (1.16 Å, CN = 6) [18]. Correspondingly, a

significant shrinkage of the lattice is observed (Table 3.1).

In the Sr2Si5N8 lattice, there are two nonequivalent Sr sites, located in a channel

along [100] formed by three-dimensional framework of corner-sharing SiN4 tetrahedra

[16]. The two Sr ions are six-fold (Sr(I)) and seven-fold (Sr(II)) coordinated with

nitrogen atoms [16]. Since the ionic radius of the Eu2+ ion is similar to that of the Sr2+ ion,

one can reasonably assume that Eu2+ statistically distributes over the two available Sr

sites, which indeed we found to be really the case [7]. However, one would expect that

Ca2+, being smaller than Sr2+, preferentially occupies the smaller Sr site (i.e. Sr(I)).

Initially the Ca ion was fixed on the Sr(I) and Sr(II) sites with the occupancies of Ca

ranging from 1:0 to 1:1. However, our refinement results clearly indicated that these

assignments of Ca yield unreasonable structures, in which some Si-N distances are

unusually small (< 1.6 Å) in comparison with the normal Si-N distances (1.6 – 1.9 Å), as

51

Chapter 3

Table 3.1. Crystallographic and optical data for Eu-doped M1.9Eu0.1Si5N8 (M = Ca, Sr) and Sr1.3Ca0.6Eu0.1Si5N8

Material Ca1.9Eu0.1Si5N8 Sr1.9Eu0.1Si5N8 Sr1.3Ca0.6Eu0.1Si5N8

Space group Monoclinic Cc Orthorhombic Pmn21 Orthorhombic Pmn21 Lattice parameters

a (Å) b (Å) c (Å) V (Å3) β (°)

Z Rwp Rp χ2

14.3377(2) 5.6087(1) 9.6835(2) 721.52(2) 112.09(1) 4 0.092 0.059 2.71

5.7069(1) 6.8142(1) 9.3269(1) 362.71(1) 2 0.082 0.057 3.80

5.6966(4) 6.7864(5) 9.3161(7) 360.16(1) 2 0.082 0.061 3.55

Av M(I)-N (Å) (CPV (Å3))1 2.624 ± 0.323 2.783 ± 0.148 (16.408 ± 0.285) 2.736 ± 0.169 (15.703 ± 0.527) Av M(II)-N (Å) (CPV (Å3))1 2.715 ± 0.256 2.811 ± 0.159 (18.939 ± 0.521) 2.838 ± 0.183 (17.816 ± 0.851) Excitation band (nm) 297, 355, 394, 460, 496 294, 334, 395, 463, 505 295, 334, 395, 462, 505 Emission band (nm) 612 620 643 Center of gravity (cm-1)2 25800 26000 26100Crystal field splitting (cm-1)3 13500 14200 14100Stokes shift (cm-1)4 3800 3700 43001. CPV M(Eu, Ca): coordination polyhedral volume calculated by the program IVTON [23] within a distance of the M ion around 0 - 3.05 Å. 2. Center of gravity calculated from the average energy of the observed 5d excitation levels of Eu2+. 3. Stokes shift calculated from the energy difference between the lowest 5d excitation band and emission band of Eu2+. 4. Crystal-field splitting estimated from the energy difference between highest and lowest observed 5d excitation levels of Eu2+

52

Chapter 3

found for such nitride and oxynitride compounds [16, 17, 19]. This suggests that if Ca

predominantly occupies on the Sr(I) site it will result in a largely distorted network and

eventually makes the crystal structure unstable. Surprisingly, a reasonable crystal

structure (e.g. all the interatomic distances are in the normal range) can only be obtained

when the occupancy of Ca on the smaller Sr(I) site is much lower than on the larger Sr(II)

site. Evidently, in case that Ca preferentially occupies the larger Sr site in

Sr1.3Ca0.6Eu0.1Si5N8 all the Si-N and Sr(Ca, Eu)-N distances are in the expected range

(Table 3.3). In addition, the Eu2+ ions are found to be almost statistically incorporated on

the two Sr sites as we found in Sr2Si5N8:Eu2+ [7]. The total lattice energy based on the

refined structures calculated by Gulp [20] are -84.56, -86.01 and -86.13 eV for Ca2+ on

only Sr(I), equally distributed over Sr(I) and Sr(II) and dominantly on Sr(II), respectively.

These results strongly support our assignment for the preferential substitution of Ca2+

ions in Sr2Si5N8:Eu2+. Fig. 3.1 shows the crystal structure of Sr1.3Ca0.6Eu0.1Si5N8 and

coordination of the Sr(Ca, Eu) atoms with nitrogen atoms. The final refinement structural

parameters are shown in Tables 3.1 to 3.3.

Sr(Ca,Eu) (I)

Sr(Ca,Eu) (II)

b

c

(a)

53

Chapter 3

(b)

Fig. 3.1. (a) Crystal structure view along [100], (b) coordination of the Sr(Ca, Eu) atoms

(black sphere) and the Sr(Ca, Eu)-N distances (Å) in Sr1.3Ca0.6Eu0.1Si5N8.

10 20 30 40 50 60 70 80 90 100 110 120

Inte

nsity

(Cou

nts)

2θ (deg.)

Fig. 3.2. Observed (+), calculated (solid) X-ray powder diffraction patterns and the

difference profile of the Rietveld refinement of Sr1.3Ca0.6Eu0.1Si5N8.

54

Chapter 3

The observed pattern is in fair agreement with the calculated X-ray diffraction pattern for

Sr1.3Ca0.6Eu0.1Si5N8 (Fig. 3.2). As mentioned before, Ca2Si5N8 and Sr2Si5N8 have different

crystal structures (Table 3.1). When a large amount of Ca is introduced, the Sr2Si5N8

lattice has to adapt itself to counteract the contraction due to the replacement of Sr2+ by

Ca2+. If Ca preferentially occupies the smallest Sr (I) site, the Sr (I) – N distances will

become too much smaller than those Sr (II) – N distances resulting in unreasonable short

Si-N bands. Obviously, this results in a large lattice stress which is not in favour of the

stability of the crystal structure.

3.2. Effect of Ca2+ substitution on the luminescence properties of Sr2Si5N8:Eu2+

Similar to Sr2Si5N8:Eu2+, the diffuse reflection spectrum of Sr1.3Ca0.6Eu0.1Si5N8 has

two absorption bands of Eu2+ centered at about 300 nm and 425 nm, as shown in Fig. 3.3.

With incorporation of Ca the reflection intensity decreases from about 16% for

Sr1.9Eu0.1Si5N8 to 11% for Sr1.3Ca0.6Eu0.1Si5N8 in the range of 400 - 465 nm, showing the

same tendency as an increase of the Eu concentration to enhance the absorption [7].

Five Eu2+ excitation bands at about 295, 334, 395, 462, 505 nm can be discriminated

(Fig. 3.4). The weak excitation band peaking at about 250 nm is attributed to the

host-lattice excitation by the valence to conduction band transitions at the absorption

edge of the host lattice (Fig. 3.3). Obviously, the replacement of Sr by Ca also does not

significantly change the shape and the position of the excitation bands of Sr1.9Eu0.1Si5N8

(Fig. 3.4), indicating that the effect of Ca on the covalency and the crystal field strength

related to the Eu2+ ions is neglectable (Table 3.1). This further verifies our previous

conclusion that the excitation characteristics of Eu2+-doped M2Si5N8 are almost

independent of the type of M (M = Ca, Sr, Ba) [7]. In contrast, the emission band of Eu2+

significantly shifts to long-wavelength from 620 to 643 nm as Ca is incorporated into the

Sr2Si5N8 lattice (Fig. 3.4). As the changes of the center of gravity and the crystal-field

splitting of Eu2+ are so small for the case with and without Ca substitution (Table 3.1),

this demonstrates that an increase of the Stokes shift is mainly responsible for the

red-shift of the Eu2+ emission band. At the same Eu concentration, this can be explained

by the shrinkage of the Sr sites as Ca is incorporated, especially for the larger Sr(II) site

due to preferential Ca substitution on it. This lattice shrinkage corresponds to the obvious

55

Chapter 3

Table 3.2. Refined atomic coordinates and isotropic displacement parameters for

Sr1.3Ca0.6Eu0.1Si5N8

Atom Wyck. S.O.F. x/a y/b z/c U [Å2] Sr1 2a 0.86 0.0 0.8747(3) 0.0015(2) 0.02811

Sr2 2a 0.44 0.0 0.8822(3) 0.3677(1) 0.00699

Ca1 2a 0.1 0.0 0.8747(3) 0.0015(2) 0.02811

Ca2 2a 0.5 0.0 0.8822(3) 0.3677(1) 0.00699

Eu1 2a 0.05 0.0 0.8747(3) 0.0015(2) 0.02811

Eu2 2a 0.05 0.0 0.8822(3) 0.3677(1) 0.00699

Si1 4b 1.0 0.2529(4) 0.6671(2) 0.6844(6) 0.01072

Si2 2a 1.0 0.0 0.0570(3) 0.6727(6) 0.00659

Si3 2a 1.0 0.0 0.4219(8) 0.4641(7) 0.00748

Si4 2a 1.0 0.0 0.4033(8) 0.9051(7) 0.01461

N1 2a 1.0 0.0 0.1874(2) 0.5255(2) 0.00542

N2 4b 1.0 0.2494(7) 0.9064(6) 0.6708(9) 0.01052

N3 4b 1.0 0.2513(9) 0.4549(8) 0.0201(8) 0.01499

Table 3.3. Selected interatomic distances (Å) for Sr1.3Ca0.6Eu0.1Si5N8. Sr1-N5 Sr1-N2 Sr1-N1 Sr1-N4 Sr1-N3 Sr1-N2 Sr2-N1 Sr2-N2 Sr2-N5 Sr2-N3 Sr2-N2 Sr2-N6

2.557(16) 2.595(6) x2 2.888(3) x2 2.892(7) 3.193(5) x2 3.399(8) x2 2.540(15) 2.731(7) x2 2.890(3) x2 3.042(6) x2 3.165(7) x2 3.247(7)

Si1-N2 Si1-N6 Si1-N3 Si1-N4 Si2-N1 Si2-N2 Si2-N5 Si3-N1 Si3-N3 Si3-N6 Si4-N5 Si4-N4 Si4-N3

1.629(5) 1.722(4) 1.740(7) 1.762(4) 1.632(18) 1.750(4) x2 1.753(18) 1.691(13) 1.726(5) x2 1.830(8) 1.647(14) 1.723(9) 1.822(6) x2

56

Chapter 3

200 300 400 500 600 700

0

10

20

30

40

50

60

70

80

90

Ref

lect

ion

(%)

Wavelength (nm)

Sr2Si5N8

Sr1.9Eu0.1Si5N8

Sr1.3Ca0.6Eu0.1Si5N8

Fig. 3.3. Diffuse reflection spectra of Sr2Si5N8, Sr1.9Eu0.1Si5N8 and Sr1.3Ca0.6Eu0.1Si5N8.

200 300 400 500 600 700 800 900

0.0

0.2

0.4

0.6

0.8

1.0

Inte

nsity

(a.u

.)

Wavelength (nm)

Sr1.3Ca0.6Eu0.1Si5N8

λem = 640 nm λexc = 395 nm λexc = 450 nm

Sr1.9Eu0.1Si5N8

λem = 620 nm λexc = 450 nm

Fig. 3.4. Excitation (left) and emission (right) spectra of Sr1.9Eu0.1Si5N8 (λexc = 450 nm,

λem = 620 nm) and Sr1.3Ca0.6Eu0.1Si5N8 (λexc = 395, 450 nm, λem = 640 nm).

57

Chapter 3

decrease of the coordination polyhedral volume of the Sr(Eu,Ca)N(n) polyhedron (n = 6,

7, respectively, within a distance 0 – 3.05 Å for the Sr ions) in Sr.9Eu0.1Si5N8 after

incorporation of the Ca ions (Table 3.1). In an isotypic lattice, on a smaller

crystallographic site a larger Stokes shift of Eu2+ luminescence is expected, thus resulting

in a longer wavelength emission (643 vs. 620 nm). This observation is in agreement with

a number of previous investigations where such a relationship is established between the

Stokes shift and the site size [10, 21, 22]. In addition, the conversion efficiency (λexc =

450 nm) of Sr1.3Ca0.6Eu0.1Si5N8 only slightly decreases (~ 5%) compared to Sr.9Eu0.1Si5N8.

Therefore, by incorporation of Ca the emission-wavelength can be varied, while keeping

the conversion efficiency about the same levels. It is expected that the conversion

efficiency of Sr2-x-yCaxEuySi5N8 can be further increased by optimization of the Ca and

Eu concentrations.

3.4. Conclusions

The incorporation of Ca has a great influence on the structural and luminescence

properties of Sr2Si5N8:Eu2+. The Rietveld refinement of X-ray powder diffraction shows

that the Ca2+ ions preferentially occupy the larger crystallographic Sr site, while the Eu2+

ions statistically distribute over two available Sr sites in the Sr2Si5N8 lattice. In addition,

the replacement of Sr by Ca in Sr2Si5N8:Eu2+ results in a remarkable red-shift of the

emission band from about 620 to 643 nm due to an increase of the Stokes shift.

58

Chapter 3

References:

1. H.T. Hintzen, J.W.H. van Krevel and G. Botty, EP 1104 799 A1, 1999.

2. R.B. Muller-Mach, G.O. Mueller, T. Juestel, P. Schmidt, US 6680569 B2, 2004.

3. C. Feldmann, T. Justel, C. R. Ronda and P. J. Schmidt, Adv. Funct. Mater., 2003,

13, 511.

4. L.S. Rohwer and A.M. Srivastava, The Electrochem. Soc., Inferface, 2003, 36.

5. J.W.H. van Krevel, Ph.D. Thesis, Eindhoven University of Technology, 2000.

6. H.A. Höppe, H. Lutz, P. Morys, W. Schnick and A. Seilmeier, J. Phys. Chem.

Solids, 2000, 61, 2001

7. Y.Q. Li, J. E.J. van Steen, A.C.A. Delsing, G. de With and H.T. Hintzen, to be

published (Chapter 2).

8. J.Y. Taso, Ed., in Light Emitting Diodes (LEDs) for General Illumination Update

2002 Optoelectronics Industry Development Association, Washington, DC, (2002).

9. G. Blasse and B.C. Grabmaier, Luminescent materials, Springer-Verlag, Berlin,

1994.

10. T.L. Barry, J. Electrochem. Soc., 1968, 115, 1181.

11. K.Kato, F. Okamoto, Jap. J. Appl. Phys., 1983, 22, 76.

12. H, Kasano, K. Megumi, H. Yamamoto, J. Electrochem. Soc., 1984, 131, 1954.

13. H.M. Rietveld, J. Appl. Crystallogr., 1969, 2, 65.

14. A.C. Larson and R.B. Von Dreele, Report LAUR 86-748, Los Alamos National

Laboratory, Los Alamos, NM, 2000.

15. B. H. Toby, J. Appl. Cryst. 2001, 34, 210.

16. T. Schlieper, W. Milius and W. Schnick, Z. Anorg. Allg. Chem, 1995, 621, 1380.

17. T. Schlieper and W. Schnick, Z. Anorg. Allg. Chem., 1995, 621, 1037.

18. R.D. Shannon, Acta Cryst., 1976, A 32, 751.

19. W. Schnick and H. Huppertz, Chem. Eur. J., 1997, 3, 679.

20. J.D. Gale, JCS Faraday Trans., 1997, 93, 629.

21. Y.Q. Li, G. de With and H.T. Hintzen, J. Alloys Comp., 2004, 385, 1.

22. Y.Q. Li, C.M. Fang, G. de With and H.T. Hintzen, J. Solid State Chem., 2004, 177,

4687.

23. T. Balic Zunic, I. Vickovic, J. Appl. Cryst. 1996, 29, 305.

59

Chapter 4

Synthesis, structural and luminescence properties

of Eu2+ and Ce3+ activated BaYSi4N7

ABSTRACT

BaYSi4N7 and its phosphors activated with Eu2+ and Ce3+ were synthesized by solid-state

reaction at 1400 – 1650 °C under nitrogen mixed with hydrogen atmosphere. The crystal

structure of BaYSi4N7 was solved by direct methods and refined by the Rietveld method

from X-ray powder diffraction data. BaYSi4N7 crystallizes in the hexagonal space group

P63mc (No.186), with a = 6.0550 (2) Å, c = 9.8567 (1) Å, V = 312.96 (2) Å3, and Z = 2,

which is isotypic with BaYbSi4N7. The photoluminescence properties have been studied

for the solid solutions of Ba1-xEuxYSi4N7 (x = 0 - 0.4) and BaY1-xCexSi4N7 (x = 0 - 0.1) at

room temperature. Eu2+- doped BaYSi4N7 gives a broad green emission band centered

between 503 and 527 nm depending on the Eu2+ concentration. The Eu2+ emission band

shows a red shift formulation with increasing Eu2+ concentration mainly caused by the

change of the crystal field strength and Stokes shift. Concentration quenching of Eu2+

emission is observed for x = 0.05 due to energy transfer between Eu2+ ions by electric

dipole-dipole interactions with a critical interaction distance of about 20 Å. Ce3+-doped

BaYSi4N7 exhibits a bright blue emission band with a maximum at about 417 nm, which

is independent of Ce3+ concentration. This is ascribed to a lower solubility of Ce3+ ions in

BaYSi4N7 lattice as shown by X-ray powder diffraction analysis.

Keywords: barium yttrium silicon nitride, europium, cerium, crystal structure, X-ray

powder diffraction, Rietveld refinement, luminescence.

60

Chapter 4

4.1. Introduction

It is of considerable interest to develop advanced luminescent materials with high-

brightness and high efficiency for applications in fluorescent lamps, light emitting diodes

(LED) and various kinds of display devices. Up to date, most of them are dominated by

oxides, sulfides, halides and phosphides doped with transition metal or rare-earth ions [1-

4]. Recent work has shown that nitride or oxynitride compounds are promising host

lattices for luminescent materials [5-8], the presence of a significant covalent character of

nitrogen atoms in the lattice may bring about some peculiar optical properties with

respect to the traditional host lattice [9, 10]. Hence, it is necessary to explore some new

nitride compounds and furthermore build-up the relationships between the chemical

composition, crystal structure and the resulting optical properties.

A series of quaternary compounds containing trivalent Yb formed with alkaline-earth

ions and silicon nitride, MYbSi4N7 (M = Sr, Ba and divalent Eu) have already been

reported [11-13]. All of them are isotypic in space group P63mc, with Z = 2. This

structure is composed of a network of corner-sharing SiN4 tetrahedra. M2+ (M = Sr, Ba,

and Eu) and Yb3+ reside within the Si6N6-ring channels. Besides common two-fold

coordinated nitrogen atom or N[2] bridges, unusual four-fold coordinated N[4] atoms are

also present, with significantly longer bond lengths than those of the Si-N bonds of the

N [2] atoms [14].

Considering the comparable ionic radius (Y3+: 0.9 Å; Yb3+: 1.02 Å) and the

similarity of some Y and Yb containing silicon oxynitride compounds, it is therefore

interesting to explore the possibility of isostructural compounds with substitution of Y for

Yb with the intention of design of promising host lattices for doping with luminescent

ions. Recently we have reported about ab initio calculation of the crystal structure and

electronic structure of MYSi4N7 (M = Sr, Ba) [15]. To systematically study those

luminescent materials, powder samples of Ba1-xEuxYSi4N7 (0 ≤ x ≤ 1) and BaY1-

xCexSi4N7 (0 ≤ x ≤ 0.1) were prepared by the conventional solid-state reaction approach.

In this paper, we describe the synthesis, and give further details about the crystal structure

of BaYSi4N7. Another goal of this work is to investigate the structure using Eu or Ce as a

structural (spectroscopic) probe and emphasize the structure-luminescence properties

61

Chapter 4

relationships of the powder samples activated with Eu2+ or Ce3+ ions.

4.2. Experimental

Powder samples of BaYSi4N7 and solid solutions of several Ba1-xEuxYSi4N7 (0 ≤ x ≤

1) and BaY1-xCexSi4N7 (0 ≤ x ≤ 0.2) compounds were prepared by solid-state reaction

from stoichiometric quantities of high purity grade Si3N4 (Cerac S-1177, measured β

content: ~91%, N content: 38.35%, 99.5%), and the metals Y (Csre, 99.9%), Ba (Aldrich,

99%, pieces), Ce (Alfa, 99%) and Eu (Csre, 99.9%, pieces). The large pieces of the Ba

and Eu metals rendered a homogeneous mixing procedure impossible, therefore, Ba3N2

and EuN were pre-synthesized by nitriding the Ba and Eu metals under a flowing pure

nitrogen atmosphere at 550 and 850 °C, respectively, and subsequently grinding them

into fine powders. The mixtures of raw materials were thoroughly mixed and ground with

an agate mortar and pestle. Subsequently the well-mixed powders were placed in a

molybdenum crucible covered with a lid and fired twice at 1400 and 1650 °C for 12-24 h

under a flowing gas of 5%H2 - 95%N2 in horizontal tube furnaces with an intermediate

grinding between the firing steps. All manipulations were carried out in a nitrogen filled

dry glove box due to the great air sensitivity of most of the raw materials.

Powder X-ray diffraction (XRD) data were collected at room temperature on a

Rigaku D/Max-γB diffractometer operating at 40 kV, 30 mA with Bragg-Brentano

geometry (flat graphite monochromator, Scintillation counter) using CuKα radiation.

The sample was mounted on a standard flat plate aluminum sample holder. For the lattice

parameters determination of both undoped and doped samples, powder diffraction data

were recorded in the 2θ range of 10-90 ° with step scan mode (step size 0.01° 2θ,

counting time per step 6 s) while 15 wt% silicon powder was used as an internal standard.

For indexing and crystal structure determination XRD data were recorded with step scan

within a 2θ range of 10-120° with a step size of 0.01° 2θ and a counting time of 20 s per

step on the finely ground samples. A 1° divergence and scatter slit together with a 0.3°

receiving slit were employed for measurement.

The photoluminescence spectra were determined at room temperature on the powder

samples by a Perkin-Elmer LS-50B luminescence spectrometer with Monk-Gillieson type

62

Chapter 4

monochromators and a 20 kW Xenon discharge lamp as excitation source. The radiation

was detected by a red sensitive photomultiplier R928. The spectra were obtained in the

range of 200 – 900 nm with a scanning speed of 100 nm/min and the selected excitation

and emission slit widths of 2.5 nm. Excitation spectra were automatically corrected;

however, all the emission spectra were corrected by taking into account the effect of the

combined spectral response of the detector of R928 and the monochromator using the

measured spectra of a calibrated W-lamp as the light source.

Diffused reflectance spectra were recorded in the range of 230 – 700 nm with BaSO4

white powder and black felt as the references.

4.3. Results and discussion

4.3.1. Structure determination of undoped BaYSi4N7

The prepared samples as examined by X-ray powder diffraction appeared to be single

phase. The accurate position and integrated intensities of the first 20 Bragg peaks were

obtained by profile fitting with the program XFIT [16] using a split Pearson VII function.

The powder X-ray diffraction pattern of BaYSi4N7 was then indexed on the basis of a

primitive hexagonal cell with unit cell parameters a = 6.0525(3) Å, c = 9.8525(7) Å, and

V = 312.57 Å3 (M20 = 225.5, F20 = 193.3(0.0038, 27)) by the powder indexing program of

DICVOL91 [17, 18] in the CRYSFIRE suite [19]. This result was also confirmed with

TREOR90 [20]. The final refined lattice parameters using Si powder as an internal

standard are listed in Table 4.1. Two formula units per primitive unit cell can be deduced

from the lattice parameters and the measured density (4.105 g·cm-3).

The systematic absences (2h-hl: l=2n; h-2hl: l = 2n and hhl: l = 2n) suggest that the

possible space groups could be P31c, P 3 1c, P63mc, P 6 2c and P63/mmc.

The crystal structure elucidation of BaYSi4N7 was carried out by ab initio crystal

determination with the program EXPO [21] using EXTRA [22] for extraction of the

integrated intensities by the Le Bail method [23] and SIRPOW97 [24] optimized for

solving crystal structure by powder data for direct methods. All the possible space groups

were used as input to the EXPO program to derive the atomic position with direct

methods in the range of 10 - 70° 2θ because of the strong intensity decrease at large 2θ

63

Chapter 4

range.

With photoluminescence spectroscopy (see photoluminescence section) on Eu2+- and

Ce3+-doped BaYSi4N7 (Eu and Ce probes are assumed to partially replace Ba and Y atom

in the lattice) only one relatively high symmetric emission band can be observed for each

of them. It is clearly suggested that there is only one Ba and Y site in the BaYSi4N7

primitive lattice and those sites should have high point symmetry. From the primary

results of EXPO program, the most probable point symmetry group for both Ba and Y

ions might be C3v with the highest site symmetry, consistent with the space group of

P63mc and P63/mmc mentioned above. Combining this information together with the

output results of the EXPO program, the position of all Ba, Y, Si and N atoms in the unit

cell was obtained. The results using the space group P63mc gave a lower structural R

factor as compared with that of the P63/mmc. Consequently, in the further structure

refinement stage only the space group P63mc was applied to carry out the Rietveld

refinement. All of the above results are similar to those obtained from a single crystal

study of BaYbSi4N7 [10-13].

The structure of BaYSi4N7 was refined by the Rietveld method [25] using the initial

coordination of atoms obtained from the above mentioned direct methods based on the

space group P63mc. Rietveld refinement was performed using the program GSAS [26，

27] in the range 10 – 120° 2θ. The scaling factor, the zero point, the background and the

lattice parameters were refined initially. The profile fitting was used a pseudo-Voigt

function corrected for asymmetry. The preferential orientation was also refined using the

March-Dollase function because of the needle-like morphology of BaYSi4N7 particle. All

atom positions and thermal displacement factors were refined and the final refinement

converged to the residual factors Rwp = 8.60% and Rp = 5.38%. Fig. 4.1 shows the final

simulation of the calculated and observed diffraction patterns. The crystallographic data

are listed in Table 4.1 and the atomic coordinates are given in Table 4.2, and some

selected bond distances and angles are summarized in Table 4.3.

BaYSi4N7 is isostructural with BaYbSi4N7 and contains a three-dimensional network

structure of corner-sharing SiN4 tetrahedra 3 [4] [2] [4] 53 6[(Si N N ) ]−

∞ . In this network the N[4]

atoms connect four Si atoms and the N[2] atoms connect two Si atoms without the

64

Chapter 4

10 20 30 40 50 60 70 80 90 100 110 120-2

-1

0

1

2

3

4

5

6

7

8

9

10

Inte

nsity

(104 c

ount

s)

2 θ (degree)

Fig. 4.1. Observed (crossed) and calculated (line) X-ray powder diffraction pattern as

well as difference profile (bottom line) between observed and calculated intensity of the

Rietveld refinement of BaYSi4N7. The positions of the Bragg reflections are marked by

vertical short lines.

Table 4.1. Crystallographic data for BaYSi4N7 Formula BaYSi4N7Formula weight 436.64 Crystal system hexagonal Space group P 63 m c (no. 186) Unit cell dimensions a = 6.0550(2) Å c = 9.8567(1) Å Cell volume 312.96(2) Å3

Z 2 Density, calculated 4.634 g/cm3

T 298 K 2 θ (deg.) range 10 – 120 Scan condiations step size 0.01, 20 s / per step R-factors wRp 0.0860 Rp 0.0538

65

Chapter 4

Table 4.2. Atomic coordinates and isotropic displacement parameters of BaYSi4N7

Atom Wyckoff Symmetry x y z U (Å 2) position

Ba 2b C3v 1/3 2/3 0.3627(3) 0.0035(3) Y 2b C3v 1/3 2/3 0.7372(3) 0.0016(4) Si1 2a C3v 0 0 0.3136(4) 0.0022(9) Si2 6c Cs 0.1732(2) 0.3464(3) 0.0465(3) 0.0036(4) N1 6c Cs 0.0276(10) 0.5138(5) 0.0996(4) 0.0020(13) N2 6c Cs 0.8461(3) 0.6921(7) 0.3730(5) 0.0022(11) N3 2a C3v 0 0 0.1230(8) 0.0096(19)

presence of N[3] atoms as generally observed in metal-silicon nitrides. The structure can

be considered as an infinite building of tetrahedral units of [N(SiN3)4] joined by sharing

N[2] atoms along the b axis (Fig. 4.2a). Both Ba2+ and Y3+ ions occupy one site in the

primitive lattice and are located in channels along [100] formed by Si6N6 rings, as shown

in Fig. 4.2. The Ba atom is surrounded by 12 nearest nitrogen neighbours: six N2 with

long-distances are in the same planar hexagonal array ((Ba1)(N2)6 layer), and six N1 with

short-distances (three above and three below the (Ba1)(N2)6 layer) form a hexagonal anti-

prism (or a distorted octahedron) around the central atom of Ba in the cubic closest

packing (CCP) framework composed of the stacking of four (Ba2+)(N3-)n layers. The Y

atom is sixfold coordinated by N (3 x N1, 3 x N2) forming a slight distorted octahedron.

The local coordination environments of Ba and Y by N atoms are presented in Fig. 4.2b

and 2c. As compared with BaYbSi4N7 the main structural difference is that the bond

lengths of most of Ba-N and Y-N as well as Si-N in BaYSi4N7 become longer since the

ionic radius of Y3+ is larger than that of Yb3+ ion (see Table 4.3) which results in the unit

cell volume increase. The Ba-N bond lengths vary from 3.006 to 3.049 Å and the Y-N

bond lengths vary from 2.309 to 2.329 Å (three long and three short bonds).

4.3.2. Solubility of Eu and Ce ions in the BaYSi4N7 host lattice

The position of the emission band and efficiencies can be tuned by varying the Eu or

Ce concentration due to altering the lattice parameters of the BaYSi4N7 host, which

results in changing crystal field strength and covalency. Therefore it is of interest to know

the solubility limit of Eu or Ce ions in the BaYSi4N7 lattice. Considering the cation radius

66

Chapter 4

Table 4.3. Selected interatomic distances (Å) and angles (deg.) for BaYSi4N7 Ba—N1 Ba—N1vii

Ba—N1viii

Ba—N1ix

Ba—N1xi

Ba—N1xii

Ba—N2xiv

Ba—N2 Ba—N2xv

Ba—N2viii

Ba—N2xi

Ba—N2xvi

Y—N1vii

Y—N1ix

Y—N1xii

Y—N2vi

Y—N2x

Y—N2xiii

Ba—Y Ba—Yi

Ba—Yii

Ba—Yiii

N1—Ba—N1vii

N1—Ba—N1viii

N1—Ba—N1ix

N1—Ba—N1xi

N1—Ba—N1xii

N1—Ba—N2xiv

N1—Ba—N2 N1—Ba—N2xv

N1vii—Ba—N1viii

N1vii—Ba—N1ix

N1vii—Ba—N1xi

N1vii—Ba—N1xii

N1vii—Ba—N2xiv

N1vii—Ba—N2 N1vii—Ba—N2xv

N1viii—Ba—N1ix

N1viii—Ba—N1xi

N1viii—Ba—N1xii

N1viii—Ba—N2xiv

3.048(6) 3.006(6) 3.049(6) 3.006(6) 3.048(6) 3.006(6) 3.0317(1) 3.0326(1) 3.0323(1) 3.0320(1) 3.0323(1) 3.0320(1) 2.328(5) 2.328(5) 2.329(5) 2.309(5) 2.309(5) 2.309(5) 3.6919(9) 3.7083(1) 3.7083(1) 3.7078(1) 145.74(3) 54.19(16) 109.26(1) 54.19(16) 145.72(3) 65.51(13) 119.67(14) 90.32(10) 145.72(3) 66.08(15) 109.26(14) 66.08(15) 86.92(10) 86.91(10) 122.13(14) 145.72(34) 54.19(16) 109.24(14) 119.68(14)

N1viii—Ba—N2 N1viii—Ba—N2xv

N1ix—Ba—N1xi

N1ix—Ba—N1xii

N1ix—Ba—N2xiv

N1ix—Ba—N2 N1xi—Ba—N1xii

N1xi—Ba—N2xiv

N1xi—Ba—N2 N1xi—Ba—N2xv

N1xii—Ba—N2xiv

N1xii—Ba—N2 N1xii—Ba—N2xv

N2xiv—Ba—N2 N2xiv—Ba—N2xv

N2—Ba—N2xv

N1vii—Y—N1ix

N1vii—Y—N1xii

N1vii—Y—N2vi

N1vii—Y—N2x

N1vii—Y—N2xiii

N1ix—Y—N1xii

N1ix—Y—N2vi

N1ix—Y—N2x

N1ix—Y—N2xiii

N1xii—Y—N2vi

N1xii—Y—N2x

N1xii—Y—N2xiii

N2vi—Y—N2x

N2vi—Y—N2xiii

N2x—Y—N2xiii

65.50(13) 145.74(3) 66.08(15) 56.17(12) 122.13(14) 86.91(10) 145.72(4) 90.33(10) 90.32(10) 119.67(14) 122.14(14) 56.16(12) 56.16(12) 173.66(17) 119.89(1) 64.99(13) 89.49(20)

89.48(20) 90.39(13) 179.81(1)

90.38(13) 89.48(20) 90.39(13) 90.38(13) 179.81(1) 179.82(1)

90.37(13) 90.37(13) 89.76(19) 89.76(19) 89.75(19)

67

Chapter 4

* Symmetry codes: (i) x-y, x, -0.5+z; (ii) 1+x-y, x, -0.5+z; (iii) 1+x-y, 1+x, -0.5+z; (iv) x, 1+y, z; (v) 1+x, 1+y, z; (vi) x-y, x, 0.5+z; (vii) 1+x-y, 1+x, 0.5+z; (viii) 1-y, 1+x-y, z; (ix) -x, 1-y, 0.5+z; (x) 1-x, 1-y, 0.5+z; (xi) -x+y, 1-x, z; (xii) y, -x+y, 0.5+z; (xiii) y, 1-x+y, 0.5+z; (xiv) -1+x, y, z; (xv) 1-y, x-y, z; (xvi) 1-x+y, 2-x, z; (xvii) 1+x-y, x, 0.5+z; (xviii) x, y, 1+z; (xix) 1-y, 1+x-y, 1+z; (xx) -x+y, 1-x, 1+z; (xxi) -1+x, -1+y, z; (xxii) x, -1+y, z; (xxiii) -x, -y, 0.5+z; (xxiv) x, y, -1+z; (xxv) 1-x, 1-y, -0.5+z; (xxvi) 1+x, y, z; (xxvii) -y, x-y, z; (xxviii) -x+y, -x, z.

(a)

68

Chapter 4

(b) (c)

Fig. 4.2. Schematic views of the crystal structure of BaYSi4N7. (a) Projection of

tetrahedral representation of the crystal structure of BaYSi4N7 along [100] direction. The

Ba2+ ions are shown as large grey spheres and the Y3+ ions as small black spheres. (b)

Coordination of the Ba atoms with twelve nitrogen atoms, and (c) Coordination of the Y

atoms with six nitrogen atoms.

[28] and the valence states, we supposed that Eu2+ ions prefer to occupy Ba sites, while

Ce3+ ions prefer to occupy Y sites. Accordingly, series of doped Eu and Ce samples were

investigated, respectively.

Fig. 4.3 shows the lattice parameters as function of the concentration for Ba1-

xEuxYSi4N7 (0 ≤ x ≤ 1) (Fig. 4.3a) and BaY1-xCexSi4N7 (0≤ x < 0.1) (Fig. 4.3b). For Ba1-

xEuxYSi4N7 the lattice parameters a, c and the unit cell volume V, nearly linearly decrease

with the Eu2+ concentration in the whole range because of the substitution of small Eu2+

for the large Ba2+ in ionic radius [28], while above x ≥ 0.4 a secondary phase YSi3N5 can

be observed. This implies that Ba1-xEuxYSi4N7 forms a limited solid solution between

BaYSi4N7-EuYSi4N7 and the maximum solubility is about x = 0.4. On the other hand, the

69

Chapter 4

lattice parameters for BaY1-xCexSi4N7 exhibit only a slight tendency to increase with an

increase of x due to the ionic radius of Ce3+ being significantly larger than that of Y3+. It

also can be seen from Fig. 4.3b that variation in Ce concentration does change much less

the lattice parameters. Due to this limited solubility, an unknown secondary phase is

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.06.00

6.02

6.04

6.06

9.799.809.819.829.839.849.859.869.87

304

305

306

307

308

309

310

311

312

313

314

315

316

a, c

(Å)

x

a cV

Uni

t cel

l vol

ume

(Å3 )

(a)

0.00 0.01 0.02 0.03 0.04 0.05 0.066.050

6.052

6.054

6.056

6.058

9.840

9.845

9.850

9.855

9.860

9.865

312.0

312.2

312.4

312.6

312.8

313.0

313.2

313.4

313.6

a, c

(Å)

x

a c

Uni

t cel

l vol

ume

(Å3 )

v

(b)

Fig. 4.3. The Eu and Ce concentration dependence of the lattice parameters of (a) Ba1-

xEuxYSi4N7 (0≤ x ≤ 1) and (b) BaY1-xCexSi4N7 (0≤ x < 0.1).

70

Chapter 4

present in Ce-doped BaYSi4N7 for x > 0.05.

In order to differentiate the interatomic distances and angles between undoped and

Eu2+ and Ce3+ doped BaYSi4N7, the Rietveld refinement was performed using powder

XRD data with the structural model based on the results for BaYSi4N7. The observed and

difference Rietveld plots are given in Fig. 4.4.

10 20 30 40 50 60 70 80 90 100 110 120

-5

0

5

10

15

20

25

30

35

40

45

Inte

nsity

(103 c

ount

s)

2θ (degree) (a)

10 20 30 40 50 60 70 80 90 100 110 120

-5

0

5

10

15

20

25

30

35

40

45

Inte

nsity

(103 c

ount

s)

2θ (degree) (b)

Fig. 4.4. Observed (crossed), calculated (line) X-ray powder diffraction pattern and the

difference profile (bottom line) between observed and calculated intensity of the Rietveld

refinement of (a) Ba0.9Eu0.1YSi4N7, and (b) BaY0.97Ce0.03Si4N7 samples. The vertical

markers show the positions calculated for Bragg reflections.

71

Chapter 4

The summary of crystallographic data, including the atomic coordination for Eu2+ and

Ce3+ doped BaYSi4N7, are given in Table 4.4 and selected distances and bond angles

given in Table 4.5. The Eu-N2 bond becomes shorter, whereas the distances of the Eu-N1

bonds are similar with the Ba-N1 distances in undoped BaYSi4N7 lattice. On the other

hand the average Y/Ce-N distances are not significantly changed in comparison with Y-N

distances of the undoped sample (as shown in Tables 3, 4 and 5). The site-occupancy

factors (Table 4.4), reveal that the Y3+ sites reject Ce3+ ions surpassing 1 mol% to occupy

its 2b sites in the BaYSi4N7 lattice, which confirms that only a small amount of Ce3+ ions

can be incorporated into BaYSi4N7.

4.3.3. Diffuse reflection of Eu- and Ce-doped BaYSi4N7

The daylight colour of undoped, Eu, and Ce-doped samples are gray-white, green-

yellow and antique-white, respectively. The typical diffuse reflectance spectra for

undoped BaYSi4N7, Ba0.9Eu0.1YSi4N7 and BaY0.97Ce0.03Si4N7 are shown in Fig. 4.5. The

spectrum of Eu-doped BaYSi4N7 is described by one broad absorption feature centered

between 310-350 nm depending on the concentration of the Eu ions. Because the

undoped sample does not present such absorption, it is implied that the absorption

originates from Eu2+. Addition of Eu2+ to form Ba1-xEuxYSi4N7 solid solution has a

significant influence on the onset of absorption. The onset of the absorption band of Eu

ions systemically shifts to longer wavelength up to x ≈ 0.3 (see inset in Fig. 4.5a)

corresponding to the solubility limit of Eu ions in the BaYSi4N7 host (Fig. 4.3a). In

contrast, except for the absorption intensity enhancement with the Ce concentration

increasing no significant effects on the onset of absorption for all BaY1-xCexSi4N7

samples (x = 0 – 0.1) could be observed. Clearly, the absorption below 260 nm is

attributed to the valence to conduction band transitions of the host lattice. This is in

agreement with our calculated predictions as reported earlier [15]. In addition, the

absorption band of BaY1-xCexSi4N7 shows two distinctly separated sub-bands (313 nm

and 334 nm, see inset in Fig. 4.5b), which is ascribed to splitting of the 4f 5d

excitation band of the Ce3+ ion.

72

Chapter 4

260 280 300 320 340 36010

15

20

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45340

350

360

370

380

390

400

410

200 250 300 350 400 450 500 550 600 650 7000

10

20

30

40

50

60

70 BaYSi4N7

BaY0.97Ce0.03Si4N7

Wavelength (nm)

BaYSi4N7

Ba0.9Eu0.1YSi4N7

(b)0

10

20

30

40

50

60

70

Ref

lect

ion

(%)

(a)

313 334

x

Ons

et o

f abs

orpt

ion

(nm

)

Fig. 4.5. Typical diffuse reflectance spectra of BaYSi4N7, Ba0.9Eu0.1YSi4N7 and

BaY0.97Ce0.03Si4N7. The inset in (a) is the onset of absorption dependence of the Eu2+

concentration. The inset in (b) shows the magnified part of the reflection pattern around

the Ce3+ absorption band.

73

Chapter 4

Table 4.4. Crystal data and refined atomic coordinations for Ba0.9Eu0.1YSi4N7 and

BaY0.97Ce0.03Si4N7

Formula Formula weight Crystal system Space group Unit cell dimensions

a Å c Å

Cell volume, V Å3 Z Density, calculated T K 2 θ (deg.) range Scan condiations R-factors wRp Rp

Ba0.9Eu0.1YSi4N7 hexagonal P 63 m c (no. 186) 6.0520 (4) 9.8540 (1) 312.56 (3) 2 298 10 – 120 step size 0.01, 20 s /per step 0.0979 0.0643

BaY0.97Ce0.03Si4N7 hexagonal P 63 m c (no. 186) 6.0550(1) 9.8563 (1) 312.90 (2) 2 298 10 – 120 step size 0.01, 20 s /per step 0.0896 0.0560

Ba0.9Eu0.1YSi4N7Atom Wyckoff x y z U g* position Ba 2b 1/3 2/3 0.3625(1) 0.0052(4) 0.90(6) Eu 2b 1/3 2/3 0.3625(1) 0.0052(4) 0.10(6) Y 2b 1/3 2/3 0.7370(1) 0.0023(6) 1.0 Si1 2a 0 0 0.3127(6) 0.0019(11) 1.0 Si2 6c 0.1729(2) 0.3458(4) 0.0458(4) 0.0047 (6) 1.0 N1 6c 0.0290(12) 0.5145(6) 0.0985(7) 0.0048(17) 1.0 N2 6c 0.8477(4) 0.6952(8) 0.3708(7) 0.0035(14) 1.0 N3 2a 0 0 0.1195(13) 0.0085(22) 1.0 BaY0.97Ce0.03Si4N7 Atom Wyckoff x y z U g position Ba 2b 1/3 2/3 0.3625(1) 0.0045(4) 1.0 Y 2b 1/3 2/3 0.7373(1) 0.0030(6) 0.993(8) Ce 2b 1/3 2/3 0.7373(1) 0.0030(6) 0.007(8) Si1 2a 0 0 0.3144(5) 0.0026(9) 1.0 Si2 6c 0.1730(2) 0.3460(3) 0.0464(3) 0.0052(5) 1.0 N1 6c 0.0265(10) 0.5133(5) 0.0989(6) 0.0011(13) 1.0 N2 6c 0.8478(4) 0.6955(8) 0.3700(6) 0.0031(12) 1.0 N3 2a 0 0 0.1248(10) 0.0076(20) 1.0 * g : Constraint on occupancy : g(Ba)+g(Eu) = 1.0 for Ba0.9Eu0.1YSi4N7 g(Y)+g(Ce) = 1.0 for BaY0.97Ce0.03Si4N7

74

Chapter 4

Table 4.5 Selected interatomic distances (Å) for Ba0.9Eu0.1YSi4N7 and BaY0.97Ce0.03Si4N7

Ba0.9Eu0.1YSi4N7 BaY0.97Ce0.03Si4N7Bond Length (Å) Bond Length (Å) Eu-N1 Eu-N1 Eu-N1 Eu-N1 Eu-N1 Eu-N1 Eu-N2 Eu-N2 Eu-N2 Eu-N2 Eu-N2 Eu-N2 Y-N1 Y-N1 Y-N1 Y-N2 Y-N2 Y-N2

3.051(7) 3.003(7) 3.051(7) 3.003(7) 3.051(7) 3.003(7) 3.030(1) 3.031(1) 3.031(1) 3.031(1) 3.031(1) 3.031(1) 2.338(7) 2.338(7) 2.339(7) 2.309(6) 2.310(6) 2.310(6)

Ba-N1 Ba-N1 Ba-N1 Ba-N1 Ba-N1 Ba-N1 Ba-N2 Ba-N2 Ba-N2 Ba-N2 Ba-N2 Ba-N2 Ce-N1 Ce-N1 Ce-N1 Ce-N2 Ce-N2 Ce-N2

3.055(6) 2.999(6) 3.055(6) 2.999(6) 3.055(6) 2.999(6) 3.032(1) 3.032(1) 3.032(1) 3.032(1) 3.032(1) 3.031(1) 2.328(5) 2.328(5) 2.328(5) 2.306(5) 2.306(5) 2.306(5)

4.3.4. Luminescence of BaYSi4N7:Eu2+

Figure 6 shows the room-temperature emission spectra of Ba1-xEuxYSi4N7 (0 ≤ x ≤ 0.4).

The inset displays the corresponding excitation spectra from bottom to top. Since we

could not obtain single-phase EuYSi4N7 and Eu-rich solid solution samples (0.5 ≤ x < 1),

the luminescence properties in this range will not be described in this paper.

The excitation spectra for Ba1-xEuxYSi4N7 (0 ≤ x ≤ 0.4) exhibit two remarkable broad

excitation bands with maxima around 342 and 386 nm together with a weak band near

283 nm (Fig. 4.6). The latter band is ascribed to host-lattice excitation (Fig. 4.5a), in

agreement with the fact that its position is independent of the Eu concentration in contrast

to the other bands (Table 4.6). As the Eu2+ concentration increases the long-wavelength

excitation band shifts from about 383 to 388 nm, while the short-wavelength excitation

band shifts from about 348 to 346 nm (Table 4.6). This is a consequence of a larger

crystal field splitting (CFS) due to shrinkage of the lattice when the Ba2+ ion is replaced

by the smaller Eu2+ ion. As compared with Eu-doped Ba2Si5N8, the excitation bands

above 400 nm [8, 9] are absent in BaYSi4N7:Eu, which is related to a different crystal and

75

Chapter 4

electronic structure as well as number of cross-linking SiN4 tetrahedra (N [x]).

Table 4.6. Spectral parameters of the Ba1-xEuxYSi4N7 (x = 0 - 0.4) and BaY1-xCexSi4N7 (x

= 0 - 0.5)

Sample Excitation maximum

(nm)

Emission maximum

(nm)

Stokes shift (cm-1)

CFS (cm-1)

Ba1-xEuxYSi4N7 0.02 283, 348, 383 503 6200 2600 0.10 283, 349, 385 508 6300 2700 0.20 283, 349, 388 517 6400 2900 0.30 283, 348, 389 526 6800 3000 0.40 283, 346, 388 537 7200 3100

BaY1-xCexSi4N7 0.01 285, 297, 317, 339 416 5500 4100 0.03 285, 297, 318, 338 417 5600 4100 0.05 285, 297, 319, 338 419 5700 4100

The emission spectra of Ba1-xEuxYSi4N7 (0 < x ≤ 0.4) consist of a single broad band

with a nearly symmetric profile in the green spectral region. It is well known that Eu2+

ions show emission bands, while Eu3+ displays sharp emission lines due to 4f 4f

transitions of 5D0 to 7FJ (J = 0 – 6) around 580 - 630 nm [1]. As no 4f 4f emission

lines originating from Eu3+ in the red spectral area can be observed, the broad green

emission band can be assigned to the 4f65d 4f7 transition of Eu2+, indicating that the

Eu ions in the nitride or oxynitride compounds are reduced to Eu2+ [9].

The position of the broad emission band shifts to longer wavelengths (Fig. 4.6) with

an increase of Eu2+ concentration (from 503 to 527 nm), as expected for Eu2+ in a

shrinking lattice. Due to the associated smaller interatomic distances, the crystal field

strength around Eu2+ increases (Table 4.6), which results in increasing splitting of the 5d

levels and lowering of the level from which emission occurs. In addition, a larger Stokes

shift is induced (Table 4.6), because a stronger relaxation is promoted for Eu2+ on a site

becoming smaller. Consequently the red-shift of the emission band for higher Eu contents

can be understood. The interaction between the host lattice and the activator Eu2+

becomes stronger with increasing Eu2+ concentration which results in broadening of the

76

Chapter 4

400 500 600 700 8000.0

0.2

0.4

0.6

0.8

1.0

200 300 400 5000.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

Emis

sion

inte

nsity

(a.u

.)

Wavelength (nm)

0.02 0.1 0.2 0.3 0.4

Exci

tatio

n in

tens

ity (a

.u.)

Wavelength (nm)

Fig. 4.6. The emission spectra of BaYSi4N7:Eu with varying Eu concentration (385 nm

excitation wavelength at room temperature). In the inset the corresponding excitation

spectra are shown from bottom to top (monitoring at 510 nm emission wavelength).

emission band (Fig. 4.7). Such behavior is ascribed to the strong coupling of the

electronic states of the Eu2+ center with vibrational modes of the host lattice [29]. The

width of the emission band is also related to the Stokes shift, generally, that is a broad

emission band corresponds to a large Stokes shift [1]. This relation is in satisfactory

agreement with the results observed for Ba1-xEuxYSi4N7 (0 ≤ x ≤ 0.4) phosphors (Fig. 4.7,

Table 4.6).

Concentration quenching of the luminescence becomes effective for Eu2+ contents

surpassing 5 mol% because the distance between Eu2+ ions becomes smaller due to the

replacement of Ba with Eu ions, which leads to the energy transfer between Eu2+ centers.

The critical distance for the energy transfer between identical Eu2+ centers in BaYSi4N7

can be estimated by the formula (1) [30]

13

32( )4c

c

VRX Nπ

≈ (1)

77

Chapter 4

0.0 0.1 0.2 0.3 0.40.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

70

80

90

100

110

120

Rel

ativ

e in

tens

ity (a

.u.)

x

FW

HM

(nm

)

Fig. 4.7. The Eu2+ concentration dependence of the emission intensity and FWHM of the

emission bands of Ba1-xEuxYSi4N7 (0 ≤ x ≤ 0.4).

-2.0 -1.8 -1.6 -1.4 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0-0.8

-0.4

0.0

0.4

0.8

1.2

1.6

2.0

log(

I obs/x

)

log x

S = 5.8

Fig. 4.8. Observed relative intensity of Eu2+ emission dependence of the concentration of

Ba1-xEuxYSi4N7 (0≤ x ≤ 0.4) under 385 nm excitation wavelength.

where Xc is the critical concentration, N is the number of Ba2+ ions in the unit cell, and V

is the volume of the unit cell. The estimated value for the energy transfer distance Rc

78

Chapter 4

between Eu2+ ions in BaYSi4N7 is approximately 20 Å, which is similar to the Rc value

for Eu2+ centers in several oxide and apatite lattices [31, 32].

Energy transfer is generally associated with multipolar interactions, radiation

reabsorption or exchange [33]. Based on the calculated Rc value and the broad emission

band with high symmetry and small spectral overlap with the excitation band, it seems

that multipolar interactions are the most relevant to Eu2+ energy transfer [1, 30, 33]. The

type of multipolar interactions between the Eu2+ ions can be identified by examining the

concentration dependence of the emission intensity from the emitting level which has

multipolar interaction [34, 35]. The emission intensity per activator ion can be expressed

by the equation for weakly absorbed exciting radiation:

3 1/ (1 '( ) )obsI x x θβ −= + (2)

where x is the activator concentration; Iobs is the observed relative emission intensity; β’

is a constant for each interaction for a given host crystal and θ = 6, 8, 10 for dipole-dipole,

dipole-quadrupole, quadrupole-quadrupole interaction, respectively.

The Iobs/x vs. x curve of the 4f65d 4f7 emission from Ba1-xEuxYSi4N7 (0 ≤ x ≤ 0.4)

is shown in Fig. 4.8. The Iobs/x data of the emission over the Eu2+ concentration range of

0.1 to 0.4 mole fraction, can be fitted with a straight line with a slope of about -1.94. This

corresponds to θ ≈ 5.8, which indicates that electric dipole-dipole interaction is

responsible for the concentration quenching of Eu2+ emission.

4.3.5. Luminescence of BaYSi4N7:Ce3+

The excitation spectra of BaYSi4N7:Ce3+ exhibit four bands at 338, 318, 297 and 285

nm (Fig. 4.9). Similar to BaYSi4N7:Eu2+ the band at about 285 nm is ascribed to host-

lattice excitation, and the remaining peaks to splitting of the 5d band into 3 levels as

expected for the incorporation of Ce3+ on the slightly distorted octahedral YN6 site in

BaYSi4N7 with point symmetry C3v.

A relatively narrow emission band centered at about 417 nm can be distinguished in the

emission spectra, in agreement with the substitution of Ce3+ ions on a single site. A

decomposition of the emission band into two Gaussian-shaped bands is displayed in an

inset in Fig. 4.9. The energy gap between the fitted two maxima is about 2009 cm-1,

which is in agreement with the value of the spin-orbit splitting of Ce3+ ground state of 4f

79

Chapter 4

configuration. Therefore, the emission of the Ce3+ in BaYSi4N7 can be assigned to the

transition of 5d 4f (2F5/2, 2F7/2).

As can be seen from Fig. 4.9, no significant shifts of both Ce3+ excitation and

emission bands are observed upon increasing the Ce3+ concentration. The lowest

excitation band is located at 338 nm, thus resulting for all Ce concentrations in about the

same Stokes shift (≈ 5600 cm-1, Table 4.6) as well as crystal field splitting (CFS) data

(Table 4.6). An untunable Ce3+ emission band by varying the Ce3+ concentration is

related to the limited solubility of Ce3+ ions in BaYSi4N7 lattice. As a consequence, the

effect of Ce3+ substitution for Y3+ ions is very slight.

200 250 300 350 400 450 500 550 600 6500

100

200

300

400

500

600

700

800

900

300 350 400 450 500 550 6000

1

2

3

4

532 30 28 26 24 22 20 18

Inte

nsity

(a.u

.)

Wavelength (nm)

x 0.01 0.03 0.05

Em

issi

on in

tens

ity (a

.u.)

Wavelength (nm)

411

448

Wavenumber (103 cm-1)

Fig. 4.9. The excitation (λem = 420 nm) and emission spectra (λexc = 338 nm) of BaY1-

xCexSi4N7 (0≤ x < 0.05) with varying Ce3+ concentration. The dashed curves (inset

figure )for the emission spectrum x = 0.01 represent deconvoluted Gaussians.

80

Chapter 4

4.4. Conclusions

A new compound, BaYSi4N7, has been synthesized by solid-state reactions and the

crystal structure was determined from X-ray powder diffraction data with direct methods.

BaYSi4N7 exhibits strong structural similarities to the already known BaYbSi4N7. The

compound crystallizes in the hexagonal crystal system, space group P63 mc, Z = 2, unit

cell parameters a = 6.0550 (2) Å, c = 9.8567 (1) Å, and V = 312.96 (2) Å3. The

refinement was carried out using the Rietveld method and the residual factors of the final

refinement are Rwp = 0.0860, Rp = 0.0538. The structure of BaYSi4N7 contains one

crystallographically distinct site for Ba and Y atoms, respectively. The Ba atoms are

twelvefold coordinated by nearest nitrogen neighbours and the Y atoms are located inside

a slightly distorted octahedron consisting of nitrogen atoms.

The optical properties of the Ba1-xEuxYSi4N7 (x = 0 - 0.4) and BaY1-xCexSi4N7 (x = 0

- 0.1) have been studied using diffuse reflectance, UV excitation and emission

spectroscopy. The interatomic distances for the local coordination of Eu and Ce atoms in

Eu2+- and Ce3+-doped samples were also obtained by the Rietveld analysis. One broad

Eu2+ green emission band with a maximum intensity around 503 - 527 nm emission

center was observed depending on the Eu2+ concentration which can be assigned to the

transition 4f65d1 4f7. Varying the Eu2+ concentration results in a significant red-shift

and broadening of the Eu2+ emission as well as a decrease of the emission intensity. The

changes in the emission spectra have been associated with changes in the crystal field

strength, Stokes shift and possibly the covalency around Eu2+ ions, as concluded from the

variation of lattice parameters with Eu concentration and the Rietveld refinement data. As

possible concentration quenching mechanism, electric dipole-dipole interaction is

proposed for Eu2+ emission. In BaY1-xCexSi4N7 (x = 0 - 0.1), a rather narrow Ce3+

emission around 417 nm is observed, its position almost independent of Ce3+

concentration.

81

Chapter 4

References:

1. G. Blasse, B. C. Grabmaier, Luminescent Materials; Springer-Verlag: Berlin, 1994.

2. Keith H. Butler Fluorescent Lamp Phosphors, The Pennsylvania State University Press:

University Park, PA, 1980.

3. Justel, T, Nikol, H. and Ronda, C., Angew. Chem. Int. Ed., 1998, 37, 3084.

4. C. Feldmann, T. Justel, C. R. Ronda and P. J. Schmidt, Adv. Funct. Mater., 2003,

13, 511.

5. J.W.H. van Krevel, H.T. Hintzen, R. Metselaar, and A. Meijerink, J. Alloys

Comp., 1998, 268, 272

6. J.W.H. van Krevel, J.W.T. van Rutten, H. Mandal, H.T. Hintzen, and R. Metselaar,

J. Solid State Chem., 2002, 165, 19.

7. K. Uheda, H. Takizawa, T. Endo, et al., J. Lumin., 2000, 87-89, 967.

8. H.A. Hoppe, H. Lutz, P. Morys, W. Schnick and A. Seilmeier, J. Phys. Chem. Solids,

2000, 61, 2001.

9. J.W.H. van Krevel, Ph.D. thesis, Eindhoven University of Technology, 2000.

10. R. Marchand, Franck Tessier, André Le Sauze and Nadège Diot., Interational J.

Inorg. Mater. 2001, 3, 1143.

11. H. Huppertz, W. Schnick, Angew. Chem. Int. Ed., 1996, 108, 2115.

12. H. Huppertz, W. Schnick, Z. Anorg. Allg. Chem., 1997, 212, 623

13. H. Huppertz, W. Schnick, Acta Cryst., 1997, C53, 1751.

14. W. Schnick, and H. Huppertz, Chem. Eur. J. 1997, 3, 679.

15. C. M. Fang, Y.Q. Li; H.T. Hintzen, and G. de With. J. Mater. Chem., 2003,13, 1480.

16. R. W. Cheary, and A.A. Coelho, J. Appl. Cryst., 1992, 25, 109.

17. D. Louer, and M. Louer, J. Appl. Cryst.,1972, 5, 271.

18. A. Boultif, and D. Louer, J. Appl. Cryst., 1991, 24, 987.

19. R. Shirley The Crysfire 2002 System for Automatic Powder Indexing: User’s Manual,

The Lattice Press, 41 Guildford Park Avenue, Guildford, Surrey GU2 7NL, England,

2002.

20. P.E. Werner, Z. Krist., 1964, 120, 375.

21. A. Altomare, M.C. Burla, M. Carmalli, B. Carrozzini, G.L. Cascarano, C. Giacovazzo,

82

Chapter 4

A. Guagliardi, A. Moliterni, G. Polidori, R. Rizzi, J. Appl. Crystallogr., 1999, 32, 339.

22. A. Altomare, M.C. Burla, G. Cascarano, C. Giacovazzo, A. Guagliardi,

A.G.G. Moliterni, and G. Polidori, J. Appl. Crystallogr., 1995, 28, 842.

23. A. Le Bail, H. Duroy and J. Fourquet, Mater. Res. Bull., 1988, 238, 447.

24. A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi, M.C. Burla, G. Polidori,

and M. Camalli, J. Appl. Crystallogr., 1994, 27, 435.

25. H.M. Rietveld, Acta Cryst., 1967, 22, 151.

26. A.C. Larson, R.B. von Dreele, GSAS: General Structure Analysis System, Los

Alamos National Laboratory, 2000.

27. B.H. Toby, J. Appl. Cryst., 2001, 34, 210.

28. R.D. Shannon, Acta Cryst., 1976, A32, 751

29. F.C. Palilla, A.K. Levine, and M.R. Tomkus, J. Electrochem. Soc., 1968, 115, 642.

30. G. Blasse, Philips Res. Rep., 1969, 24, 131.

31. S.H.M. Poort, W.P. Blokpoel, G. Blasse, Chem. Mater. 1995, 7, 1547.

32. M. Kottaisamy, R. Jagannathan, P. Jeyagopal, R.P. Rao, and R.L. Narayanan, J. Phys.

D: Appl. Phys., 1994, 27, 2210.

33. D.L. Dexter, J. Chem. Phys., 1953, 21, 836.

34. L.G. van Uitert, J. Electrochem. Soc., 1967, 114, 1048.

35. L. Ozawa, P.M. Jaffe, J. Electrochem. Soc., 1971, 118, 1678.

83

Chapter 5

Preparation, structure and photoluminescence properties of

Eu2+ and Ce3+-doped SrYSi4N7

ABSTRACT Undoped and Eu2+ or Ce3+-doped SrYSi4N7 were synthesized by solid-state reaction

method at 1400 – 1660 °C under nitrogen/hydrogen atmosphere. The crystal structure

was refined from the X-ray powder diffraction data by the Rietveld method. SrYSi4N7

and EuYSi4N7, being isotypic with the family of compounds MYbSi4N7 (M = Sr, Eu, Ba)

and BaYSi4N7, crystallize with the hexagonal symmetry: space group P63mc (No.186), Z

= 2, a = 6.01597 (3) Å, c = 9.78939 (2) Å, V = 306.83(3) Å3; and a = 6.0123 (1) Å, c =

9.7869 (1) Å, V = 306.37(1) Å3, respectively. Photoluminescence properties have been

studied for Sr1-xEuxYSi4N7 (x = 0 – 1) and SrY1-xCexSi4N7 (x = 0 - 0.03) at room

temperature. Eu2+- doped SrYSi4N7 shows a broad yellow emission band peaking around

548 - 570 nm, while Ce3+ -doped SrYSi4N7 exhibits a blue emission band with a

maximum at about 450 nm. SrYSi4N7: Eu2+ can be very well excited by 390 nm radiation,

which makes this material attractive as conversion phosphor for LED lighting

applications.

Keywords: strontium yttrium silicon nitride, europium, cerium, crystal structure, X-ray

powder diffraction, Rietveld refinement, luminescence

84

Chapter 5

5.1. Introduction

Rare-earth ions have an important role in silicon nitride, Sialon and related nitrogen

materials. Rare-earth oxides are often used as sintering aids not only lowering the

sintering temperature but also improving high-temperature and creep properties [1-2]. On

the other hand, rare-earth doped phosphors are also used in modern fluorescent lamps,

displays and X-ray intensifying/scintillation screens, such as LaPO4: Ce, Tb, Y2O3: Eu,

Y2O2S:Eu and Y2SiO5:Ce, as well as YTaO4:Nb [3-4]. In recent years, it has been shown

that rare earth ion-doped silicon/aluminium nitrides and oxynitrides might be promising

phosphor materials [5-11].

In recent years several new quaternary rare earth containing silicon nitride compounds

like MYbSi4N7 (M = Sr, Ba, Eu) were prepared and characterized [12-13]. The crystal

structure of these compounds is different from the conventional metal-silicon nitride

compounds. Although the framework in MYbSi4N7 is also built up by the basic building

block of corner-sharing SiN4 tetrahedra, no NSi3 (N[3]) units are present. Instead, an

unusual N[4] atom which is coordinated by four silicon atoms is found, besides N[2] ions

[12-14]. The bond lengths to the N[4] atoms are significantly larger than those to the

twofold coordinated N[2] atoms. Recently, we successfully prepared the pure BaYSi4N7

compound which is isotypic with MYbSi4N7 and determined the optical properties of the

undoped [15] and rare-earth doped [16] materials.

Based on the fact that an amount of Y3+ and Yb3+, Sr2+ and Eu2+ containing compounds

are isostructural in oxynitrides and nitrides, like LnSi3O3N4 (Ln = Y, Yb), MYbSi4N7 (M

= Sr, Eu) and MSi2N5 (M = Sr, Eu), it is interesting to synthesize the SrYSi4N7 and

EuYSi4N7 compounds. If the two compounds really exist with the same crystal structure,

it is expected that complete solid solutions are formed and the optical properties can be

tuned by varying the Eu2+ concentration. Apart from the Sr2+ site, also the Y3+ site in

SrYSi4N7 is available for trivalent activator ions like Ce3+, and it is of interest to compare

the results for this material with our previous results obtained for BaYSi4N7: Ce [16].

In the present study, we report the preparation of MYSi4N7 (M = Sr, Eu) compounds by

the solid-state method and the determination of their crystal structure by Rietveld

refinement of X-ray diffraction measurements. In addition, a series of varying

85

Chapter 5

compositions Sr1-xEuxYSi4N7 and SrY1-xCexSi4N7 are synthesized and we report on the

luminescence properties of Eu2+ or Ce3+-doped SrYSi4N7.

5.2. Experimental

Powder samples of Sr1-xEuxYSi4N7 (0 ≤ x ≤ 1) and SrY1-xCexSi4N7 (0 ≤ x ≤ 0.05)

compounds were prepared by solid-state reaction from stoichiometric quantities of high

purity grade Si3N4 (two types, viz. 1. Cerac S-1177, measured β content: 91%, N content:

38.35%; O content: ~0.7%, with purity 99.5%；2. Permascand Grade P95H, measured α

content: 91%, N content: 38.08%; O content: ~1.5%, with purity 99%), Y (Csre, 99.9%,

powder), Sr (Aldrich, 99%, pieces), Ce (Alfa, 99%, pieces) and Eu (Csre, 99.9%, pieces).

SrNx (x ~ 0.65) and Eu-nitride (approximately EuN) were pre-synthesized by a nitriding

reaction of Sr and Eu metals under nitrogen atmosphere at 800 - 850 °C, and then

grinding them into fine powders. The starting mixtures were thoroughly mixed and

ground with an agate mortar and pestle. All manipulations were carried out in a

protecting atmosphere in a glove box filled with dry nitrogen because of the great air

sensitivity of most of the raw materials. Subsequently the well-mixed starting powders

were placed in a molybdenum crucible and fired at 1400 and 1660 °C for 12 and 16 h,

respectively, under a flowing gas mixture 5%H2-95%N2 in horizontal tube furnaces with

an intermediate grinding between the firing steps.

Powder X-ray diffraction (XRD) data were collected on a Rigaku D/Max-γB

diffractometer with Bragg-Brentano geometry (flat graphite monochromator, scintillation

counter) using CuKα radiation operating at 40 kV, 30 mA at room temperature. The

lattice parameters were determined in the 2θ range of 10-90 ° using step scan mode (step

size 0.01°, counting time per step 10 s) using Si powder as an internal standard reference.

For the Rietveld analysis, the powder diffraction data were recorded in the 2θ range of

10-120 ° using step scan mode (step size 0.01°, counting time per step 20 s) on finely

ground samples. A 1° divergence and scatter slit together with a 0.3° receiving slit were

employed for the measurements. Rietveld refinement [17] was performed using the

program GSAS [18, 19]. The scaling factor, the zero point, the background and the lattice

parameters were refined initially. A pseudo-Voigt function was chosen to fit the profile.

The preferential orientation was also refined with the March-Dollase function.

86

Chapter 5

The photoluminescence spectra were recorded at room temperature on powder samples

by a Perkin-Elmer LS-50B luminescence spectrometer with Monk-Gillieson type

monochromators and a 20 kW Xenon discharge lamp as excitation source. The spectra

were obtained in the range of 200 – 900 nm with a scanning speed of 100 nm/min.

Diffuse reflectance spectra were recorded in the range of 230 – 700 nm with BaSO4 white

powder and black felt as the reference materials. Excitation spectra were automatically

corrected for the lamp intensity by a second photomultiplier. All the emission spectra

were corrected by taking into account the combined effect of the spectral response of the

detector and the transmission of the monochromator using the measured spectra of a

calibrated W-lamp as the light source.

5.3. Results and discussion

5.3.1. Preparation

In the introductory investigation of the synthesis processes, we used α-Si3N4 (O

content: ~1.5%) powder as the raw material. However, a large amount of secondary

phases like, Sr2Si5N8 or Eu2Si5N8 and unidentified phases was present in the final product,

even when the sample was fired at high temperatures for a long time. After changing the

starting Si3N4 powder from α to the normally less reactive β-Si3N4 (O content: ~0.7%),

high phase purity compounds of SrYSi4N7 and EuYSi4N7 could be obtained. X-ray

powder diffraction analysis showed that a small amount (< 9%) of YSi3N5 [20] is present

in the final samples. Apart from YSi3N5, still some peaks due to traces of Sr2Si5N8 or

Eu2Si5N8 are also detected in the SrYSi4N7 and EuYSi4N7 samples as well as their solid

solutions. It is well known that α-Si3N4 has a higher oxygen content (O content: 1.2 ~

2.5%) than β-Si3N4 powder (O content: <1%) [21]. In addition, the Y powder also

contains a small amount of oxygen impurity. On the contrary for BaYSi4N7 single phase

material can be obtained easily with both α- and β-Si3N4 [15, 16]. Our experiments have

shown that with the cation size decreasing going from Ba to Ca the preparation of

MYSi4N7 becomes more difficult, CaYSi4N7 could even not be obtained at all. Possibly

oxygen can be incorporated by replacing nitrogen when the divalent ion is sufficiently

small to be incorporated on the Y-site for charge compensation, ultimately resulting in

structure breakdown. The sensitivity to oxygen was confirmed by our experiments

87

Chapter 5

showing that the sample became purer by firing several times at high temperatures for a

long time using α-Si3N4 in an atmosphere of mixed N2-H2 (10%) with the purpose to

remove oxygen from its lattice. It has been proved that long term heat treatment under

reducing atmosphere is an effective way to eliminate the lattice oxygen from AlN

ceramics [22]. Based on the same considerations, only β-Si3N4 starting powder with low

oxygen content was used as raw material in the further investigations.

5.3.2. Structure determination

The X-ray powder diffraction patterns of SrYSi4N7 and EuYSi4N7 are found to be

similar to those of MYbSi4N7 (M = Sr, Ba, Eu) [12-14] and BaYSi4N7 [16]. Therefore,

the structure of SrYbSi4N7 (space group: P63mc) was employed as the starting model for

the Rietveld refinement of the structures of SrYSi4N7 and EuYSi4N7.

Since the prepared samples contained a small amount of impurity phases (YSi3N5,

Sr2Si5N8 or Eu2Si5N8), some parts of the data were excluded for the refinement. All

atomic positions and equivalent isotropic displacement parameters were refined

converging to the residual factors Rwp = 9.33%, Rp = 5.96%; and Rwp = 11.38%, Rp =

8.33% for SrYSi4N7 and EuYSi4N7, respectively. The resulting crystallographic data are

summarized in Table 5.1. The atomic coordinates and equivalent isotropic displacement

parameters are given in Table 5.2. A list of selected bond distances and angles is gathered

in Table 5.3. The final calculated and observed diffraction patterns are presented in Fig.

5.1 for SrYSi4N7 and Fig. 5.2 for EuYSi4N7, respectively.

The structure of MYSi4N7 being isostructural with MYbSi4N7 consists of SiN4

tetrahedra which share corners, in this way forming a three-dimensional network

structure with large channels along [100] and [010] formed by Si6N6 rings. Both Sr2+ (or

Eu2+) and Y3+ ions occupy a site in the above mentioned channels. The Sr2+ (or Eu2+) ion

is coordinated by twelve nitrogen atoms (SrN12 or EuN12) and the Y3+ ion is coordinated

by six nitrogen atoms (YN6). In the network one N atom (N3) connects four Si atoms

(N[4]) and the other N atoms (N1 and N2) connect two Si atoms (N[2]) without the

presence of N[3] atoms as is usual the case in metal silicon nitrides [12-14]. The building

tetrahedral units of [N(SiN3)4] are linked by sharing N[2] atoms extending along the a and

88

Chapter 5

Table 5.1. Crystallographic data for SrYSi4N7 and EuYSi4N7 Formula SrY Si4N7 EuYSi4N7Formula weight 486.92 451.26 Crystal system hexagonal hexagonal Space group P 63 m c (no. 186) P 63 m c (no. 186) Cell parameters, Å a = 6.0160(1) a = 6.0123 (1) c = 9.7894(1) c = 9.7869 (2) Cell volume, Å3 306.83 (1) 306.37 (1) Z 2 2 Density, calculated, g/cm3 4.188 4.891 T , K 298 298 2 θ (deg.) range 10 – 120 10 - 120 Scan condiations step size 0.01, 20 s / per step R-factors wRp 0.0933 0.1138 Rp 0.0596 0.0833 RF

2 0.0484 0.0991 χ2 8.04 8.96 Table 5.2. Atomic coordinates and isotropic displacement parameters of SrYSi4N7 and

EuYSi4N7.

Atom Wyckoff Symm. x y z Uiso (Å ) Y 2b 3m. 1/3 2/3 0.4552(4) 0.0032 Sr 2b 3m. 1/3 2/3 0.08448(35) 0.0065 Si1 2a 3m. 0 0 0.5297(5) 0.0031 Si2 6c .m. 0.82655(18) 0.6531(4) 0.26008(27) 0.0034 N1 6c .m. 0.5138(4) 0.4862(4) 0.3125(4) 0.0013 N2 6c .m. 0.1512(5) 0.3025(9) 0.5855(5) 0.0065 N3 2a 3m. 0 0 0.3292(8) 0.0028 Atom Wyckoff Symm. x y z Uiso (Å ) Y 2b 3m. 1/3 2/3 0.4590(6) 0.0029 Eu 2b 3m. 1/3 2/3 0.0896(6) 0.0100 Si1 2a 3m. 0 0 0.5156(12) 0.0040 Si2 6c .m. 0.8271(4) 0.6543(8) 0.2685(7) 0.0026 N1 6c .m. 0.5177(11) 0.4823(11) 0.2986(10) 0.0052 N2 6c .m. 0.1479(8) 0.2960(16) 0.6004(12) 0.0033 N3 2a 3m. 0 0 0.3250(17) 0.0108

89

Chapter 5

Table 5.3. Selected bond distances (Å) and angles (°)

SrYSi4N7 EuYSi4N7

Sr-N1 Sr-N1 Sr-N2 Sr-N2 Y-N1 Y-N2 Si1-N2 Si1-N3 Si2-N1 Si2-N2 Si2-N3

2.918(5) (x3) 3.103(5) (x3) 3.013(1) (x3) 3.012(1) (x3) 2.342(3) (x3) 2.286(5) (x3) 1.668(5) (x3) 1.962(9) 1.710(1)) (x2) 1.724(5) 1.930(4)

N1-Sr-N1 Si1-N2-Si2 Si1-N3-Si2 Si2-N3-Si2 N1-Y-N1 N1-Y-N2 N1-Y-N2 N1-Y-N2 N2-Y-N2 N2-Si1-N2 N2-Si1-N3 N1-Si2-N3 N1-Si2-N1 N1-Si2-N2 N2-Si2-N3

67.82(18) (x3) 116.8(4) 110.5 (3) (x3) 108.4(3) (x3) 88.06(13) (x3) 89.99(13) (x 3) 89.98(13) (x 3) 177.29(18) (x3) 91.89 (19) (x3) 109.80(27) (x3) 109.14(27) (x3) 111.78(22) (x2) 107.62(34) 111.43(19) (x2) 102.82(32)

Eu-N1 Eu-N1 Eu-N2 Eu-N2 Y-N1 Y-N2 Si1-N2 Si1-N3 Si2-N1 Si2-N2 Si2-N3

2.806(12) (x 3) 3.243(13) (x 3) 3.015(1) (x 3) 3.014(1) (x 3) 2.481(11) (x 3) 2.375(11) (x 3) 1.750(10) x 3 1.866(15) 1.641(4) (x 2) 1.665(13) 1.883(7)

N1-Eu-N1 72.7(4) (x 3) Si1-N2-Si2 127.2(7) Si1-N3-Si2 107.1(6) (x 3) Si2-N3-Si2 111.8(5) (x 3) N1-Y-N1 84.2(4) (x 3) N1-Y-N2 176.4(5) (x 3) N1-Y-N2 93.12(28)(x 3) N1-Y-N2 93.11(28)(x 3) N2-Y-N2 89.5(4) (x 3) N2-Si1-N2 99.4(6) (x 3) N2-Si1-N3 118.3(5) (x 3) N1-Si2-N3 118.0(5) (x 2) N1-Si2-N1 109.9(9) N1-Si2-N2 105.2(5) (x 2) N2-Si2-N3 98.1(7)

90

Chapter 5

10 20 30 40 50 60 70 80 90 100 110 120

-10

-5

0

5

10

15

20

25

30

35

40

20 30 40 50 60

-10

-5

0

5

10

15

20

25

30

Inte

nsity

(x10

3 cou

nts)

2θ (deg.)

* *

*

*

Fig. 5.1. The Rietveld refinement pattern for SrYSi4N7 using X-ray powder diffraction

data. Plus (+) marks represent the observed intensities, and the solid line calculated

patterns. A difference (obs.– cal.) plot is shown in the bottom. The tick marks above the

difference data indicate the positions of Bragg reflections for SrYSi4N7. The asterisk (*)

in inset indicates the impurity peaks.

10 20 30 40 50 60 70 80 90 100 110 120

-20

-10

0

10

20

30

40

50

10 20 30 40 50-20

-10

0

10

20

30

40

50

Inte

nsity

(x10

3 cou

nts)

2θ (deg)

* *

*

**** *

** *

Fig. 5.2. The Rietveld refinement pattern for EuYSi4N7 using X-ray powder diffraction

data. Plus (+) marks represent the observed intensities, and the solid line calculated

patterns. A difference (obs.– cal.) plot is shown in the bottom. The tick marks above the

difference data indicate the positions of Bragg reflections for EuYSi4N7. The asterisk (*)

in inset indicates the impurity peaks.

91

Chapter 5

b axis (Fig. 5.3). The mean Si-N distance in EuYSi4N7 is about 0.03 Å smaller as

compared with SrYSi4N7. The Si2-N1 and Si2-N2 distances almost reach the minimal

value of the range typical for metal silicon nitride compounds [23, 24] which indicates

that the Si-N network in EuYSi4N7 is distorted as compared with that of SrYSi4N7, as

shown in Fig. 5.4. It is worth noting that such significantly different interatomic distances

of Si-N in SrYSi4N7 and EuYSi4N7 may be responsible for the formation of the small

amounts of the impurity phases observed in the solid solution samples (see below),

especially at Eu-rich side. Some individual atomic displacement parameters show a little

bit difference in SrYSi4N7 and EuYSi4N7 (Table 5.2). Because these parameters represent

the combined total of several effects in addition to displacements caused by thermal

motion, we will not give a further discussion.

Y

Eu

Y

M

Y

Y

M N2 M

Y

M

Y

M

Y

Y

N2

Y

M

M

Y

M M

Y Y

Y

M N2 M

Y

M

Y

M

Y

Y

M

Y

M

M

Y

N2 M

Y Y

Y

M M

Y

M

Y

Y

Y

M

Y

ab

c

Fig. 5.3. Schematic illustration of crystal structure of MYSi4N7 (M = Sr, Eu) along [100].

92

Chapter 5

(a) (b)

Fig. 5.4. The building groups of [N(SiN3)4] in (a) SrYSi4N7 and (b) EuYSi4N7

Accurate cell parameters of Sr1-xEuxYSi4N7 (0 ≤ x ≤ 1) compounds were determined

by the Rietveld method. Fig. 5.5(a) shows the variation of the a, c parameters and unit

cell volume (V) versus x. The a and V parameters decrease with increasing Eu

concentration going from 6.0160 (1) Å, 306.83 (3) Å3 to 6.0138 (1) Å, 306.46 (1) Å3 for

undoped and Eu-doped (x = 0.3) samples. The decrease is in agreement with the fact that

the ionic radius of the Eu2+ ion is slightly smaller than that of the Sr2+ ion [25]. The c/a

ratio of Sr1-xEuxYSi4N7 is almost constant for all x values (~ 1.627). Therefore, it can be

concluded that the overall structural shrinkage of Sr1-xEuxYSi4N7 lattice is isotropic.

When x > 0.3 the lattice parameters still slightly decrease but a significant amount of

secondary phase, like YSi3N5 and (Sr,Eu)2Si5N8, is present in the samples. Therefore, the

lattice parameters in the x range between 0.3 and 0.9 are not given.

The lattice parameters of SrY1-xCexSi4N7 (0 ≤ x ≤ 0.05) compounds are shown in Fig.

5.5(b). As expected, with increasing Ce concentration the lattice parameters show a slight

increase because of Ce3+ being larger than the Y3+ ion [25]. A very limited solubility of

Ce in SrYSi4N7 is found around x = 0.03.

5.3.3. Reflection spectra of the undoped and doped SrYSi4N7 compounds

The diffuse reflectance spectra of undoped, Eu-doped and Ce-doped SrYSi4N7 samples

93

Chapter 5

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.06.011

6.012

6.013

6.014

6.015

6.016

6.017

9.7729.7769.7809.7849.7889.7929.796

306.0

306.1

306.2

306.3

306.4

306.5

306.6

306.7

306.8

306.9

307.0

307.1

307.2

a, c

(Å)

x

a c V

Uni

t cel

l vol

ume

(Å3 )

(a)

0.00 0.01 0.02 0.03 0.04 0.05 0.066.005

6.010

6.015

6.020

6.025

6.0309.780

9.785

9.790

9.795

9.800

9.805

9.810

305

306

307

308

309

310

a c

a, c

(Å)

x

Uni

t cel

l vol

ume

(Å3 )

V

(b)

Fig. 5.5. The lattice parameters as a function of x in (a) Sr1-xEuxYSi4N7 and (b) SrY1-

xCexSi4N7.

are shown in Fig. 5.6. The reflection spectrum of undoped SrYSi4N7 shows an absorption

edge at about 350 - 375 nm (corresponding with the valence to conduction band

transitions of the host lattice) indicating that the band gap of the SrYSi4N7 compound is

about 3.3 – 3.5 eV, in fair agreement with our results (2.9 eV) estimated by the first-

principles calculations [15].

94

Chapter 5

250 300 350 400 450 500 550 600 650 7000

10

20

30

40

50

60

70

Ref

lect

ion

(%)

Wavelength (nm)

[Eu] x 0.00 0.02 0.05 0.10 0.20 0.30 1.00

(a)

200 300 400 500 600 7000

10

20

30

40

50

60

70

342

Ref

lect

ion

(%)

Wavelength (nm)

SrYSi4N7

SrY0.97Ce0.03Si4N7

320

(b)

Fig. 5.6. Reflection spectra of (a) Sr1-xEuxYSi4N7, and (b) SrY1-xCexSi4N7 (x = 0.03).

For comparison, EuYSi4N7 is also present in Fig. 5.6(a).

The incorporation of Eu2+ ions into the SrYSi4N7 lattice results in broad absorption

bands in the range of 300 – 450 nm. With increasing Eu concentration the onset of the

absorption band gradually extends to longer wavelengths into the visible part of the

spectrum (Fig. 5.6a). Correspondingly, the daylight color of the samples changes from

yellow green to orange and dark red for the Eu-rich samples.

95

Chapter 5

The Ce3+-doped SrYSi4N7 displays a pronounced doublet absorption band in the region

of 310 to 350 nm (Fig. 5.6b), similar to BaYSi4N7: Ce3+ [16]. The doublet band peaking

at about 320 and 342 nm correspond to splitting of the 4f 5d excitation band (see next

section). These absorption bands thereby demonstrate a good ability to be stimulated in

this region by UV.

5.3.4. Luminescence properties

5.3.4.1. SrYSi4N7:Eu2+

The emission spectra of SrYSi4N7:Eu2+ at room temperature show a broad band at

about 550 nm (Fig. 5.7a). This emission band corresponds to the 4f65d 4f7 transition of

Eu2+. Obviously, the emission bands of samples doped with a low Eu concentration

appear to be symmetric indicating that only a single Eu site is present in the SrYSi4N7

lattice. With increasing Eu content the emission band exhibits a red-shift from 548 to 570

nm. In addition, an emission shoulder around 660 to 680 nm in the spectra becomes

evident. Because SrYSi4N7 and EuYSi4N7 are isotypic and only one crystallographic site

is available for the divalent cation, this shoulder (second band) probably is originating

from the second phase (Sr, Eu)2Si5N8 as described before.

The excitation spectra of SrYSi4N7:Eu2+ (Fig. 5.7b) show broad bands peaking at about

340 and 390 nm, matching the absorption range as observed in the reflection spectra (Fig.

5.6a). Normally, when the Eu2+ ion occupies a lattice site with C3v symmetry a splitting

into three 5d bands is expected in the excitation spectra. Due to serious overlap,

especially for high Eu concentrations, only two distinct 5d bands can be observed in the

excitation spectra. SrYSi4N7: Eu can thus be well excited with a GaN-based LED, which

makes this material promising for LED lighting applications. The relative intensity of

excitation systematically decreases with increasing Eu concentration. The emission

intensity monitored at an excitation wavelength of 390 nm shows a maximum at around x

= 0.05. When x > 0.05 the emission intensity decreases dramatically (Fig. 5.7c).

The Stokes shift, roughly estimated from the maxima in the excitation and emission

spectra, increases from about 7900 cm-1 for lower Eu contents to 8300 cm-1 for higher Eu

concentrations (Fig. 5.8). The Stokes shift becomes larger as expected for total lattice

96

Chapter 5

400 450 500 550 600 650 700 750 800 850

0.0

0.2

0.4

0.6

0.8

1.0 λexc= 390 nm

Emis

sion

inte

nsity

(a.u

.)

Wavelength (nm)

[Eu] x 0.02 0.05 0.10 0.20 0.30

(a)

200 250 300 350 400 450 500 5500

100

200

300

400

500

600

700

800

Rel

ativ

e ex

cita

tion

inte

nsity

(a.u

.)

Wavelength (nm)

[Eu] x 0.02 0.05 0.10 0.20 0.30

λem = 550 nm

(b)

97

Chapter 5

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

0

1

2

3

4

5

6

7

8

9

Rel

ativ

e em

issi

on in

tens

ity (a

.u.)

x

(c)

Fig. 5.7. (a) Emission, (b) excitation spectra and (c) concentration of Eu2+ dependence of

emission intensity for Sr1-xEuxYSi4N7 (0 < x ≤ 1) (λexc = 390 nm)

contraction arising from the substitution of Eu2+ for Sr2+ ions, while the mean Eu-N

distance has a negligible influence (Fig. 5.8, Table 5.3). The mean bond length, i.e. EuSr-

N, has a limited effect in the x range of 0 to 0.3. It can be seen (Fig. 5.7b) that with

increasing Eu concentration the covalency slightly increases (shift of excitation band to

lower energy), whereas the crystal field is hardly affected (as expected from the similar

sizes Eu2+ and Sr2+). Compared with Eu-doped BaYSi4N7 [16], the emission band is

shifted to longer wavelength (about 30 nm) because the metal-ligand distances are

smaller in SrYSi4N7:Eu. As a consequence, the crystal field strength and the 5d splitting

(as deduced from the splitting of the 5d excitation band, Table 5.4) is larger. Moreover,

because Ba2+ is larger than Sr2+ [25], the relaxation of the Eu2+ ion in the excited state

possibly is larger in SrYSi4N7 as compared to BaYSi4N7, resulting in a higher Stokes shift

[26 - 28]. Both effects possibly can explain that the emission in SrYSi4N7: Eu2+ is at

lower energy (i.e. longer wavelength) than in BaYSi4N7: Eu2+.

98

Chapter 5

Fig. 5.8. Relation between x，the maxima of the emission band and the mean SrEu-N

distance, the unit cell volume and the Stokes shift of Sr1-xEuxYSi4N7 (0< x ≤ 0.3).

5.3.4.2. SrYSi4N7:Ce3+

SrYSi4N7:Ce3+ exhibits an intense blue emission under ultraviolet excitation. A broad

symmetric emission band with a maximum at about 450 nm can be observed (Fig. 5.9a).

No distinguishable emission doublet due to transitions from the lowest 5d level to the 2F5/2 and 2F7/2 spin-orbit split 4f ground state level could be observed in the spectra,

similar to the case of BaYSi4N7:Ce3+ [16]. However, the emission band can be fit two

Gaussians centered at 435 nm and 473 nm, respectively (Fig. 5.9b), whose difference is

about 1847 cm-1 which is in agreement with the theoretical difference between the 2F5/2

and 2F7/2 levels (~2000 cm-1 [29]). With Ce concentration increasing from 1 to 3%, no

emission band shift was observed consistent with the low Ce solubility in SrYSi4N7.

The excitation spectrum of SrYSi4N7:Ce3+ shows three intense bands at 280, 320 and

340 nm, respectively (Fig. 5.9a). The bands at longer wavelengths correspond with

transitions from the 2F5/2 ground state to levels of the Ce3+ 5d configuration split by the

crystal field interaction, in fair agreement with the absorption bands in the reflection

spectra (Fig. 5.6b).

99

Chapter 5

200 250 300 350 400 450 500 550 600 650 700

0

50

100

150

200

250

300

350

400

450

[Ce] x 0.01 0.03

Inte

nsity

(a.u

.)

Wavelength (nm)

λ em = 340 nmλ exc = 445 nm

(a)

32 30 28 26 24 22 20 18 16

300 350 400 450 500 550 600 650 700

0.0

0.5

1.0

1.5

2.0

2.5

473

Emis

sion

inte

nsity

(a.u

.)

Wavelength (nm)

435

x = 0.01

Wavenumber (cm-1)

(b) Fig. 5.9. (a). Excitation and emission spectra of SrY1-xCexSi4N7 (0 < x ≤ 0.03);

(b) Emission spectrum of SrY1-xCexSi4N7 (x = 0.01) fit to two Gaussians.

100

Chapter 5

Table 5.4. Luminescence data of Eu2+ - or Ce3+ -doped MYSi4N7 (M = Sr, Ba)

Eu2+-doped Ce3+-doped MYSi4N7

Excitation band (nm)

Emission band (nm)

Stokes shift (cm-1)

Excitation band (nm)

Emission band (nm)

Stokes shift (cm-1)

Sr

(this work)

340, 382-386 548 - 570 7900 - 8300 280, 320, 340 450 7200

Ba

(Ref. 16)

348, 385 505 - 537 6200 - 7200 285, 318, 338 415 - 420 4100

Chapter 5

For the Ce3+ ion incorporated on the Y3+ site in MYSi4N7 (M = Sr, Ba), the

replacement of Sr by Ba has a negligible effect on the crystal field, as deduced from the

similar position of the excitation bands (Table 5.4). Similarly to the Eu-doped case [26 –

28], because Ce shrinks during excitation, this shrinkage is more obstructed in an

expanded host lattice, resulting in a smaller Stokes shift, and consequently the emission

band is at a lower wavelength in BaYSi4N7: Ce3+ (415 - 420 nm).

5.4. Conclusions

Sr1-xEuxYSi4N7 (x = 0 - 1) and SrY1-xCexSi4N7 (x = 0 – 0.05) have been synthesized by

a solid-state reaction method. The crystal structure of MYSi4N7 (M = Sr, Eu) isostructural

with MYbSi4N7 (M = Ba, Sr, Eu), was refined from the X-ray powder diffraction pattern

by the Rietveld method. SrYSi4N7 and EuYSi4N7 crystallize in the hexagonal symmetry:

space group P63mc (No.186), Z = 2, a = 6.01597 (3) Å, c = 9.78939 (2) Å, V = 306.83(3)

Å3; and a = 6.0123 (1) Å, c = 9.7869 (1) Å, V = 306.37(1) Å3 for SrYSi4N7 and EuYSi4N7,

respectively. The Eu2+ emission was found at 548-570 nm in Eu-doped SrYSi4N7 for low

Eu content. Its excitation maximum is at about 390 nm, which is a favourable position for

LED lighting purposes. With increasing Eu concentration the Eu2+ emission band shifts to

longer wavelength and the emission intensity decreases. Ce3+-doped SrYSi4N7 exhibits a

narrow blue emission band with a maximum at about 450 nm.

References:

1. G. Petzow, M. Herrmann, Silicon nitride ceramics. In: D.M.P. Mingos (ed.) Struct

Bond (High Performance Non-Oxide Ceramics II.), 2002, 102, 47.

2. V.A. Izhevskiy, L.A. Genova, J.C. Bressiani and F. Aldinger, J. Eur. Ceram. Soc.,

2000, 20, 2275.

3. T. Jüstel, H. Nikol and C. Ronda, Angew. Chem. Int. Ed. 1998, 37, 3084.

4. C. Feldmann, T. Jüstel, C.R. Ronda and P.J. Schmidt, Adv. Funct. Mater. 2003, 13, 511.

5. J.W.H. van Krevel, Ph.D. thesis, Eindhoven University of Technology, 2000.

6. J.W.H. van Krevel, H.T. Hintzen, R. Metselaar and A. Meijerink, J. Alloys

102

Chapter 5

Comp., 1998, 268, 272

7. J.W.H. van Krevel, H.T. Hintzen and R. Metselaar, Mater. Res. Bull., 2000, 35, 747.

8. J.W.H. van Krevel, J.W.T. van Rutten, H. Mandal, H.T. Hintzen and R. Metselaar,

J. Solid State Chem., 2002, 165, 19.

9. S.R. Jansen, J. Migchels, H.T. Hintzen and R. Metselaar, J. Electrochem. Soc., 1999,

146, 800.

10. K. Uheda, H. Takizawa and T. Endo, J. Lumin., 2000, 87-89, 967.

11. H.A. Höppe, H. Lutz, P. Morys, W. Schnick and A. Seilmeier, J. Phys. Chem.

Solids, 2000, 61, 2001.

12. H. Huppertz and W. Schnick, Angew. Chem. Int. Ed. Engl. 1996, 108, 2115.

13. H. Huppertz and W. Schnick, Z. Anorg. Allg. Chem., 1997, 623, 212.

14. H. Huppertz and W. Schnick, Acta Cryst. 1997, C53, 1751.

15. C. M. Fang, Y.Q. Li, H.T. Hintzen and G. de With, J. Mater. Chem., 2003, 13, 1480.

16. Y.Q. Li, G. de With and H.T. Hintzen, J. Alloys. Comp., 2004, 385, 1.

17. H.M. Rietveld, J. Appl. Crystallogr., 1969, 2, 65.

18. A.C. Larson and R.B. Von Dreele, Report LAUR 86-748, Los Alamos National

Laboratory, Los Alamos, NM, 2000

19. B. H. Toby, J. Appl. Cryst. 2001, 34, 210.

20. Thommy C. Ekström, Kenneth J. D. Mackenzie, Martin J. Ryan, Ian W. M. Brown

and G. Vaughan White, J. Mater. Chem., 1997, 7, 505.

21. Samuel Natansohn, Arvid E. Pasto and William J. Rouke, J. Am. Ceram. Soc., 1993,

16, 2273.

22. Anil V. Virkar, T.B. Jackson and R.A. Cutler, J. Am. Ceram. Soc., 1989, 72, 2031.

23. W. Schnick and H. Huppertz, Chem. Eur. J., 1997, 3, 679

24. W. Schnick, International J. Inorg. Mater., 2001, 3, 1267.

25. R.D. Shannon, Acta Cryst., 1976, A32, 751.

26. G. Blasse, J. Chem. Phys., 1969, 51, 3529.

27. G. Blasse and A. Bril, Philips Tech. Review, 1970, 31, 314.

28. A. Meijerink and G. Blasse, J. Lumin., 1989, 43, 287.

29. G. Blasse and B.C. Grabmaier, Luminescent materials, Springer-Verlag, Berlin,

1994.

103

Chapter 6

Structure and luminescence properties of YTbSi4N6C ABSTRACT The synthesis, structural and luminescence properties of a new rare-earth-silicon-nitride-

carbide YTbSi4N6C, prepared by the reaction of TbN, Y, α-Si3N4 and SiC in N2

atmosphere at 1550 - 1650 oC, have been investigated. The crystal structure, determined

by X-ray powder diffraction and refined by Rietveld profile analysis, is monoclinic (P21/c,

No. 14) with a = 5.9339(1) Å, b = 9.8925(2) Å, c = 11.8870(3) Å, β = 119.62(1), Z = 4,

i.e. YTbSi4N6C is isostructural with Re2Si4N6C (Re = Ho, Tb). YTbSi4N6C shows an

unusual long-wavelength 4f-5d excitation band of Tb3+ at about 300 nm due to the highly

covalent silicon-nitride-carbide network. The emission spectrum of Tb3+ reveals typical

strong lines in the green region with high efficiency, originating from the 5D4 7FJ (J =

0 – 6) transitions. In 1 mol% Ce3+-doped YTbSi4N6C, the energy transfer from Ce3+ to

Tb3+ is observed. For the first time, a green Tb3+ emission can be realized by the

excitation of Ce3+ ions in the visible range (i.e. 390 – 480 nm) via Ce3+ Tb3+ energy

transfer, which demonstrates a new approach to use the line emission of the rare-earth

ions for white-light LED applications.

Keywords: Crystal structure, Luminescence properties, Rare-earth silicon nitride carbide, Terbium, Cerium, X-ray powder diffraction, Rietveld refinement, Energy transfer, White- Light-LEDs.

104

Chapter 6

6.1. Introduction

Recently, several quaternary rare-earth containing silicon-nitride-carbides have been

found and characterized [1-4]. Unlike quaternary silicon nitride compounds, such as

MReSi4N7 (M = Ba, Sr, Eu; Re = Y, Yb) [5-11], the silicon-nitride-carbide, Re2Si4N6C

(Re = Ho, Tb, La) [1-4], consists of a three-dimensional network of star like [C(SiN3)4]

units, which are isoelectronic to the characteristic building [N(SiN3)4] units in MReSi4N7

(M = Ba, Sr, Eu; Re = Y, Yb) compounds [5-11]. These units are connected by sharing

N[2] (i.e. one nitrogen coordinates with two silicon, NSi2) atoms to form two kinds of

layers with diametrical orientation of the SiN3C tetrahedra. Along [001] these two types

of layers are alternately connected by further N[2] atoms to build up the three-dimensional

condensed framework [Si4N6C]6-. Similar to MReSi4N7 (M = Ba, Sr, Eu; Re = Y, Yb), the

rare-earth ions are located at the channels along [100] [1, 12]. Because of the substitution

of N3- by C4- in MReSi4N7, i.e. the replacement of the fourfold coordinated nitrogen

atoms by carbon in the framework, the lattice becomes more rigid due to the Si-C bond

which has a higher covalence than the Si-N bond. Subsequently, improved mechanical

properties and unique optical as well as magnetic properties are expected [1, 12].

Most recently, the quaternary silicon-nitride-carbide have been extended to the

pentad silicon-nitride-carbide system by replacement of M2- by a different rare-earth ion

and N[4] (i.e. four-coordinated nitrogen) by a carbon ion, respectively, in MYSi4N7 (M =

Ba, Sr) compounds [3,4,13]. For example, CeYSi4N6C and LaYSi4N6C can be derived

from the substitution of Ba by Ce or La and N[4] by C, simultaneously. These compounds

are isostructural with MReSi4N7 (M = Ba, Sr, Eu; Re = Y, Yb) with a space group of

P63mc (no.186) and shows interesting luminescence properties with a long wavelength

excitation band (~ 395 nm) for Ce3+-doped materials [13]. When we used two small rare-

earth ions (as compared to the Ce3+ ion), like Tb, Y, to replace the alkaline-earth ions and

a carbon to replace N[4] in MYSi4N7 (M = Ba, Sr), no isomorphic compounds were

obtained [14]. The resulting XRD pattern is very similar to Ho2Si4N6C, Tb2Si4N6C [1]

and Y3Si6N11 [15] which implies that its crystal structure possibly isotypic with

Ho2Si4N6C and Tb2Si4N6C [1]. In the present study, we report the synthesis, structural

and luminescence properties of a new YTbSi4N6C compound. Normally, Tb3+ shows

green emission, but the Tb3+ ions usually have to be excited by UV light at rather short

105

Chapter 6

wavelength (< 230 nm). Sensitization of the Tb3+ emission (5D4 7FJ) for UV-blue to

visible light through energy transfer from a sensitizer (i.e. Ce3+) to Tb3+ is of great

interesting for white-light LED applications. Therefore, the feasibility of using Ce3+ as a

sensitizer of Tb3+ is also investigated here.

6.2. Experimental 6.2.1. Preparation

The samples were synthesized by a high temperature solid-state reaction method. The

YTbSi4N6C compound was prepared using Tb metal (Csre, >99%, ingots), Y metal

powder (Csre, YH2 ~ 97%, O content ~ 5%), α-Si3N4 powder (Permascand, P95H,

measured α content 93%; Oxygen content: ~1.5%), β-Si3N4 powder (Cerac, S1173,

measured β content: 91%, O content <1%, , N content: 38.35%), carbon black powder

(Cabot, efflex 125) and SiC (Alfa, 99%) as starting materials. Firstly, the binary terbium

nitrides TbNx (x ≈ 0.99, ~ TbN) were synthesized by nitridation of Tb metal at 1200 °C

for 12 hours in a horizontal tube furnace under pure nitrogen atmosphere in a closed

molybdenum crucible. Subsequently, appropriate amounts of terbium nitride TbNx (x ≈

0.99) or Tb metal, Y metal powder, Si3N4 as well as SiC or carbon black powders were

weighed out, mixed and ground in an agate mortar. All manipulations were performed in

a glove box flushed with dry nitrogen because of the metal and the metal nitride materials

being air and moisture sensitive. The powder mixtures were fired in a molybdenum

crucible at 1550 - 1650 °C for 10 h in a horizontal tube furnace under pure nitrogen

atmosphere. After firing, the samples were cooled down in the furnace.

For the YTbSi4N6C:Ce3+ samples in which Ce3+ was employed as a sensitizer of the

Tb3+ ions, 0.5 and 1 mol% Ce (Alfa, 99%, lumps) was used as a starting material to

substitute for Y in YTbSi4N6C with the same preparation processes as that of YTbSi4N6C.

In order to compare the luminescence properties between YTbSi4N6C and Tb2Si4N6C,

the Tb2Si4N6C compound was also prepared in a similar way as mentioned above.

106

Chapter 6

6.2.2. Structure determination

All measurements were performed on finely ground powder samples, which were

analyzed by X-ray powder diffraction (Rigaku, D/MAX-B) using Cu-Kα radiation at 40

kV and 30 mA with a graphite monochromator.

The X-ray diffraction data were collected from 10-120o 2θ at 0.02° intervals using

CuKα radiation with a counting time of 12 s per step. Structure refinement was carried

out by the Rietveld method [16], using the program GSAS [17, 18]. The structural

parameters of Ho2Si4N6C [1] and MYbSi4N7 (M = Ba, Sr, Eu) [5-7] were used as the

initial parameters for the crystal structure model of YTbSi4N6C. The second phase of

Y2Si3O3N4 was also taken into account in the refinement processing. The refined

parameters include the scale factor, zero shift, background, lattice parameters, peak

profile parameters, fractional coordinates of individual atoms, and isotropic displacement

parameters.

6.2.3. Optical measurements

The diffuse reflectance, emission and excitation spectra of the samples were measured

at room temperature by a Perkin Elmer LS 50B spectrophotometer equipped with a Xe

flash lamp. The reflection spectra were calibrated with the reflection of black felt

(reflection 3%) and white barium sulfate (BaSO4, reflection ~100%) in the wavelength

region of 230-700 nm. The excitation and emission slits were set at 2.5 nm. The emission

spectra were corrected by dividing the measured emission intensity by the ratio of the

observed spectrum of a calibrated W-lamp and its known spectrum from 300 to 900 nm.

Excitation spectra were automatically corrected for the variation in the lamp intensity by

a second photomultiplier and a beam-splitter. All the spectra were measured with a scan

speed of 100 nm/min.

6.3. Results and discussion

6.3.1. Synthesis of YTbSi4N6C

In order to avoid oxygen contamination from the raw materials and the processing as

usual for most of the nitride based compounds, initially, we tried to use β-Si3N4 powder

107

Chapter 6

(oxygen content < 1 wt%) and/or carbon black powder as the raw materials instead of the

α-Si3N4 and/or SiC powders. However, for these starting materials a large amount of

secondary phase (YSi3N5 [15] or Y6Si11N20O [19] like-phase) was present in the final

products. We also found that the amount of the second phase significantly increased with

prolonging firing time, firing in a mixture of N2 – 10 %H2 and removing the cover of the

crucibles. After changing starting Si3N4 from β- to α- type, using fused SiC powder

instead of the active carbon black and making use of a closed Mo crucible, a highly

phase-pure YTbSi4N6C was obtained with only a small amount of Y2Si3O3N4. These

observations suggest that the YTbSi4N6C lattice is an oxygen containing compound,

which means that YTbSi4N6C possibly requires some level of oxygen for structural

stabilization. Normally, the oxygen contamination partially replaces the nitrogen atom

N[2] in the nitride network in an ordered or disordered manner. With the X-ray diffraction

data alone, it is not possible to distinguish between O2- and N3- because of their similar

scattering factors. Therefore, in this study we use the ideal YTbSi4N6C formula to denote

this new compound.

Figure 1 shows the X-ray powder diffraction patterns of YTbSi4N6C and Tb2Si4N6C

for comparison. Apparently, YTbSi4N6C has a powder diffraction pattern which is similar

to Ho2Si4N6C [1]. The pattern was indexed on a monoclinic unit cell [20] with the

possible space groups P21/c (No.14), P21/n (No.14) and P21 (No.4), respectively.

Inte

nsity

(cou

nts)

Tb2Si4N6C

TbYSi4N6C

Ho2Si4N6C

** *

10 20 30 40 50 60 70

2θ (deg)

Fig. 6.1. X-ray powder diffraction diagrams of YTbSi4N6C, Tb2Si4N6C and calculated powder diffraction pattern of Ho2Si4N6C (*indicates the second phase).

108

Chapter 6

6.3.2. Structure determination

Of all possible space groups, only P21/c (No. 14) yielded the most reasonable results

using Ho2Si4N6C [1] as the initial model for YTbSi4N6C. However, the refinements using

the structural model reported for MYbSi4N7 (M = Ba, Sr, Eu) [5-7] were not successful.

For a space group P21/c, there are three possibilities for the arrangement of Tb and Y

atoms in YTbSi4N6C. That is, (1) Tb and Y are on the Ho(1) and Ho(2) sites, respectively;

(2) in reverse, Y and Tb are on the Ho(1) and Ho(2) sites, respectively; (3) Tb and Y are

statistically distributed on both two Ho sites. The best result was obtained when the Tb

and Y ion are equally shared over both Ho sites (i.e. site occupancies of 50% for both Y

and Tb ions), and the final refinement converges with Rwp = 0.089, Rp = 0.063, χ2 = 8.2.

The other two possible models were hardly convergent and gave unreasonable

interatomic distances. The refined results are shown in Table 6.1 and the atomic

coordination and the selected bond distances are listed in Table 6.2. Fig. 6.2 shows the

observed and calculated powder diffraction patterns of YTbSi4N6C along with the

difference profile of measured and calculated diffraction patterns.

YTbSi4N6C belongs to the family of Re2Si4N6C (Re = Ho, Tb) compounds. Fig. 6.3

shows a section of this structure. This structure is closely related to MReSi4N7 (M = Ba,

Sr, Re = Yb, Y), as described in ref [1], its network structure [Si4N6C]6- can be derived

from MYSi4N7 (M = Ba, Sr) by substitution of N[4] by C[4] and by chemical twinning

perpendicular to [001] performing a symmetry operation on an inversion centre.

YTbSi4N6C contains a three-dimensional framework of corner-connected SiN3C

tetrahedra, which are condensed on one carbon ligand CSi4. The remaining three nitrogen

atoms per tetrahedra connect only two silicon atoms (N[2]) finally resulting in the

tetrahedron units. The (Tb/Y)(I) ions are coordinated by five nitrogen

atoms and the (Tb/Y)(II) ions are coordinated by six nitrogen atoms within a distance of

2.28 – 2.61 Å. The (Tb/Y)(I) and (Tb/Y)(II) atoms are situated in the channels formed by

the 6-rings silicon-nitride-carbide network with average distances of 2.4390 and 2.3871

Å for the (Tb/Y)(I)-N and (Tb/Y)(II)-N bonds, respectively (see Table 6.2 and Fig.

6.3(b)). The observed interatomic distances, for example Si-N, Si-C and (Tb/Y)-N, are in

the normal range as found in silicon nitride, silicon carbide, Tb and Y containing nitride

based compounds except for Si1-N2 (1.676 Å) and Si2-N4 (1.664 Å) which are

[2]3 4[C(SiN ) ]

109

Chapter 6

Table 6.1. Crystallographic data and refinement results for YTbSi4N6C. Formula Space group Lattice constants Formula units per cell Rwp Rp χ2

YTbSi4N6C P21/c (No. 14) Monoclinic a = 5.9339(1) Å b = 9.8925(2) Å c = 11.8870(3) Å β = 119.62(1) V = 606.62(4) Å3 Z = 4 0.089 0.063 8.2

Table 6.2. Atomic coordinates, isotropic displacement parameters (Å2) and selected

interatomic distances (Å) for YTbSi4N6C.

Atom Wyckoff S.O.F. x/a y/b z/c Uiso Tb1 Y1 Tb2 Y2 Si1 Si2 Si3 Si4 N1 N2 N3 N4 N5 N6 C1

4e 4e 4e 4e 4e 4e 4e 4e 4e 4e 4e 4e 4e 4e 4e

0.5 0.5 0.5 0.5

0.3351(6) 0.3351(6) 0.6644(7) 0.6644(7) 0.0047(22) 0.1678(22) 0.1933(18) 0.6728(18) 0.0450(4) 0.0970(4) 0.2292(34) 0.4900(6) 0.5190(5) 0.6850(4) 0.0200(6)

0.5640(2) 0.5640 (2) 0.4154(2) 0.4154(2) 0.4780(6) 0.2128(9) 0.2094(9) 0.2245(10) 0.2086(23) 0.0404(22) 0.0266(22) 0.2843(20) 0.2541(19) 0.0544(21) 0.2925(16)

0.0947 (3) 0.0947(3) 0.4206(3) 0.4206(3) 0.2488(10) 0.4223(9) 0.1761(9) 0.1561(9) 0.0001(23) 0.3997(19) 0.2194(18) -0.0057(26) 0.2470(19) 0.1232(18) 0.2663(25)

0.0212 0.0212 0.0225 0.0225 0.0177 0.0187 0.0224 0.0181 0.0149 0.0167 0.0063 0.0167 0.0156 0.0088 0.0158

Tb1/Y1 N2 1x 2.2804(19) N4 1x 2.3523(34) N3 1x 2.4612(15) N5 1x 2.4930(19) N2 1x 2.6079(26) Average 2.4390

Tb2/Y2 N1 1x 2.3213(22) N6 1x 2.3225(23) N3 1x 2.3271(24) N6 1x 2.3696(24) N5 1x 2.4058(20) N4 1x 2.5764(30) Average 2.3871

110

Chapter 6

0 10 20 30 40 50 60 70 80 90 100 110 120

Inte

nsity

(cou

nts)

2θ (deg.)

Y2Si3O3N4

YTbSi4N6C

Fig. 6.2. Observed (crosses), calculated (solid line) X-ray powder diffraction patterns and

the difference profile of the Rietveld refinement of YTbSi4N6C. The vertical bars below

the diffraction pattern represent the possible diffraction position of YTbSi4N6C.

somewhat shorter than the typical Si-N bond length [1-4, 8-9, 21]. As discussed in the

previous section, this observation may be ascribed to the partial occupation of oxygen on

the nitrogen sites probably resulting in local structure distortion in which the Si-N bond

length becomes short based on the fact that the Si-O or Si-N/O bond is shorter than Si-N

[21, 22]. Because of the limited capability of the X-ray powder diffraction technique and

an uncertain amount of oxygen impurity in the samples, we can not assign those oxygen

atoms to the specific nitrogen atom positions in YTbSi4N6C. For this purpose neutron

diffraction experiments need to be performed.

6.3.3. Luminescence properties of YTbSi4N6C and Tb2Si4N6C

Fig. 6.4 shows the diffuse reflection, the excitation and emission spectra of YTbSi4N6C.

The diffuse reflection spectrum of YTbSi4N6C shows a broad absorption band at about

111

Chapter 6

(a)

(b)

Fig. 6.3. Crystal structure of YTbSi4N6C: (a) view along a-axis; (b) coordination of the

Tb/Y (I) and Tb/Y (II) atoms.

112

Chapter 6

300 nm originating from the Tb3+ ions (Fig. 6.4). A short-wavelength band (i.e. at higher

energy) around 240 - 250 nm in the reflection spectrum can be assigned to the absorption

of the host lattice. The band gap derived from the reflection spectrum is about 240 – 260

nm which is close to that of MYSi4N7 (~ 250 nm for M = Sr and 260 nm for M = Ba) [10,

11].

200 250 300 350 400 450 500 550 600 650 700 750

0.0

0.2

0.4

0.6

0.8

1.0

0.0

0.2

0.4

0.6

0.8

1.0

7F6→ 5D3

5 D4 -

7 F 0

5 D4 -

7 F 1

Ref

lect

ion

(%)

f → d

5 D4 -

7 F 6

Inte

nsity

(a. u

.)

Wavelength (nm)

5 D4 -

7 F 5

5 D4 -

7 F 4

5 D4 -

7 F 3

5 D4 -

7 F 2

f → f

λem = 544 nm λexc = 298 nm

Fig. 6.4. Diffuse reflection, excitation (left) and emission (right) spectra of

YTbSi4N6C (λem = 544 nm, λexc = 298 nm).

The excitation spectrum of Tb3+ 5D4 7F5 emission of YTbSi4N6C consists of a very

strong band having an unusual long wavelength centered at about 300 nm corresponding

with the absorption band observed in the diffuse reflection spectrum. In addition, a

number of sharp lines in the region from 350 – 500 nm (some of them are not present in

figure 4). The broad band corresponds to a transition between the 4f8 ground state and

4f75d1 excited states of Tb3+. The weak lines at low energy are attributed to the transition

between the energy levels within the 4f8 configuration, i.e. 7F6 5D3 and 7F6 5D4. The

5d excitation band of Tb3+ at longer wavelength is rather particular [23]. It can be

explained by a highly covalent host lattice due to the presence of N and in addition C in

the silicon-nitride-carbide network. As a consequence, the positions of the 5d excitation

of Tb3+ shift to lower energies, i.e. longer wavelengths.

113

Chapter 6

The emission spectrum of Tb3+ in YTbSi4N6C is composed of a series of sharp lines in

the region of 470 – 700 nm which originates from the 5D4 7Fj (J = 6,5,4,3,2) transitions

of Tb3+, while the emission from the 5D3 level is very weak in YTbSi4N7. This is possibly

attributed to the well-known cross-relaxation between the Tb3+ ions at a high Tb3+

concentration with the Tb3+-Tb3+ distance becoming short resulting in quenching of the 5D3 emission in YTbSi4N6C.

Although Tb3+ occupies two rare-earth metal sites, no change in position of the

excitation band by varying the monitoring emission wavelength was observed indicating

that the environment of the two Tb3+ ions is very similar or the Tb3+ ions are not so

sensitive to their neighbors.

The spectra of Tb2Si4N6C are very similar to that of YTbSi4N6C. The position of the

excitation band shows a slight shift to longer wavelength with a maximum at about 310

nm which is longer than that found for YTbSi4N6C. With a similar network of the two

compounds, the Tb3+ 5d band at lower energy can possibly be explained by a slight

expansion of the lattice due to the replacement of small Y3+ ions (0.90 Å, CN = 6) by the

larger Tb3+ ions (0.923 Å, CN = 6) [24]. In Tb2Si4N6C, the integrated emission intensity

(5D4 7FJ = 6,5,4,3,2) of Tb3+ is remarkably decreased due to the obvious reason of

concentration quenching for the 5D4 level emission of Tb3+ in Tb2Si4N6C. That is, the

emission intensity of YTbSi4N6C from the 5D4 7FJ (J = 6,5,4,3,2) transitions is about 7

times larger than that of Tb2Si4N6C at the same excitation wavelength of 300 nm.

6.3.4. Energy transfer from Ce3+ to Tb3+ in YTbSi4N6C

Codoping of Tb3+-activated phosphors with Ce3+ acting as the sensitizer is well

established for mercury gas-discharge lamps, e.g. CeMgAl10O19:Tb3+ (CAT),

GdMgB5O10:Ce3+,Tb3+ (CBT) and LaPO4: Ce3+,Tb3+ (LAP) [23, 25-27]. The 254 nm

radiation of the gas-discharge is absorbed by Ce3+, which then transfers its energy to the

4f levels, i.e., 5D3, 5L10 and higher energy levels of Tb3+, finally resulting in mainly 5D4 7FJ green emission. For efficient energy transfer, a high concentration Tb3+ is

usually required for quenching of the 5d 4f of Ce3+ and 5D3 7FJ of Tb3+ (i.e. the

cross-relaxation process) emissions, respectively. As from a comparison with Tb3+, the

absorption and excitation band of Ce3+ in YTbSi4N6C is expected to be in the range 370-

114

Chapter 6

450 nm, the challenging question is whether the Tb3+ emission of YTbSi4N6C can be

sensitized for blue-UV LED (InGaN) emission by co-doping with Ce3+. Different from

the traditional way of sensitization, the Ce3+ Tb3+ energy transfer then has to take

place to the emitting the 5D4 level of Tb3+ itself.

The diffuse reflection, excitation and emission spectra of Y0.99Ce0.01TbSi4N6C are

illustrated in Fig. 6.5. Both the reflection (Fig. 6.5 a) and excitation (Tb3+ 5D4 emission at

about 544 nm) spectra (Fig. 6.5 b) of Y0.99Ce0.01TbSi4N6C clearly show two main bands

with the maxima at about 300, and a second band composed of two subbands centered at

390 and 420 nm (Fig. 6.5). Definitely, the first excitation band belongs to the Tb3+ 4f8 -

4f75d transition and the others originate from the Ce3+ 4f - 5d transition. For excitation

into the Ce3+ absorption band below 380 nm, the emission spectrum mainly shows Tb3+

emission together with a very weak Ce3+ emission because the Tb3+ ions can be directly

excited themselves via 7F6 5D3 transition in this range [23]. On the contrary, for

excitation into the Ce3+ absorption bands above 380 nm (i.e. from 390 to 470 nm) both

Ce3+ band and (superimposed on it) Tb3+ line emission are present in the emission

spectrum (Fig. 6.5 b). The Ce3+ emission band is estimated to have its maximum at about

530 - 560 nm (Fig. 6.5c). Based on the fact that hardly any Tb3+ emission (5D4 7FJ, J =

6,5,4,3,2) can be observed with excitation wavelengths above 420 nm in both YTbSi4N6C

and Tb2Si4N6C compounds, it can be concluded that energy transfer really occurs from

Ce3+ to Tb3+ in Y0.99Ce0.01TbSi4N6C as for these wavelengths (> 420 nm) only direct

excitation of the Ce3+ ions (and not Tb3+) is possible (see Fig. 6.5b). Unlike most Tb and

Ce co-doped materials (for example oxide host lattices), in this case the 7F6 - 5D4 energy

level of Tb3+ is partially overlapping with the Ce3+ emission band (see inset in Fig. 6.5b)

which results in energy transfer directly from Ce3+ 5d band to the 5D4 (and not the 5D3

and 5L10) level of Tb3+. As a schematic diagram shown in Fig. 6.6, the primary 5d

excitation levels are just situated between 5D3 and 5D4 levels of Tb3+. In YTbSi4N6C:Ce3+

the 420 - 490 nm excitation energies can be absorbed by Ce3+ through the 4f 5d

transition. After relaxation the Ce3+ ion transfers its excitation energy to the nearest

neighbour Tb3+ ions which then are pumped to the 5D4 level from which the 5D4 7FJ (J

115

Chapter 6

200 250 300 350 400 450 500 550 600 650 700

0

10

20

30

40

50

60

Ref

lect

ion

(%)

Wavelength (nm)

(a)

(b)

Ce3+ f - d

Tb3+

(a)

200 250 300 350 400 450 500 550 600 650 700 750 800

0.0

0.2

0.4

0.6

0.8

1.0

420 440 460 480 500 520 5400.0

0.1

0.2

0.3

0.4

0.5

Tb3+

Inte

nsity

(a. u

.)

Wavelength (nm)

λem = 544 nm λexc = 420 nm

Ce3+

Ce3+

7F6 - 5D4

(b)

116

Chapter 6

450 500 550 600 650 700 750

0.0

0.2

0.4

0.6

0.8

1.0

450 500 550 600 650 700 750

0.0

0.1

0.2

0.3

0.4

0.5

Emis

sion

inte

nsity

(a.u

.)

Wavelength (nm)

5 D4 -

7 F 5

5 D4 -

7 F 6

5 D4 -

7 F 4

5 D4 -

7 F 3

5 D4 -

7 F 2

Inte

nsity

(a. u

.)

Wavelength (nm)

Tb3+

Ce3+

(c)

Fig. 6.5. (a) Diffuse reflection (a: YTbSi4N6C, b: Y0.99Ce0.01TbSi4N6C), and (b) excitation

(left) and emission (right) spectra of Y0.99Ce0.01TbSi4N6C. Inset shows an enlarged image

of overlapped region between the Tb3+ excitation and the Ce3+ emission spectra; (c) the

fitted emission spectra: the dashed line shows the 5d – 4f emission band of Ce3+; the inset

shows the emission of Tb3+after subtraction the emission of Ce3+ (λexc = 420 nm).

= 6 – 2) emission occurs; meanwhile, the Ce3+ ion also transfer excitation energy to itself,

as we found in Ce3+-doped Y2Si4N6C [28], from which the 5d 4f emission of Ce3+

takes place.

Several compositions have been examined for better understand this kind of the energy

transfer. We found that in the case of low Tb concentrations (i.e., Y2-xTbxSi4N6C:Ce3+ (1

mol%), x < 0.5) the dominant emission comes from the Ce3+ ions and the Tb3+ line

emission is extremely weak in the excitation range above 420 nm possibly because the

absorption edge of Tb3+ upward shifts to higher energy levels of the 5d band of Ce3+.

While in the case of Y1.98Tb0.02Si4N6C:Ce3+ (10 mol %), only the Ce3+ emission can be

117

Chapter 6

0

5

10

15

20

25

30

35

2

7F62F5/2

5D3

5

432

7F0

Ener

gy (x

103 c

m-1)

1

5D4

5d

2F7/2

Ce3+ Tb3+

1

Fig. 6.6. Schematic representation of the luminescence of YTbSi4N6C:Ce3+. 1 indicates

the 4f 5d emission of Ce3+ after relaxation from the 5d excitation levels; 2 indicates

energy transfer between Ce3+ and Tb3+ from which 5D4 7FJ emission occurs.

observed indicating a lower energy-transfer rate from Ce3+ to Tb3+. Therefore, in reverse

to CeMgAl12O19:Tb3+ and LaPO4: Ce3+, Tb3+, the concentration of Tb3+ should be higher

than that of Ce3+. Because the Ce3+-to-Tb3+ energy transfer is just limited to several

angstroms [23], high Tb3+ concentration is understandable. In addition, an appropriate

Ce3+ concentration (i.e. 1 - 3 mol %) is also helpful for quenching of the Ce3+ emission by

energy migration [28], evidently from our found in Y1-xCexTbSi4N6C (x = 0.01, 0.02)

cases, in order to obtain efficient energy-transfer from Ce3+ to Tb3+ in YTbSi4N6C.

6.4. Conclusions

A new rare-earth silicon-nitride-carbide compound YTbSi4N6C has been synthesized

by a solid-state reaction at high temperature. YTbSi4N6C crystallizes in the monoclinic

space group P21/c (No. 14) with a = 5.9339(1) Å, b = 9.8925(2) Å, c = 11.8870(3) Å, β =

118

Chapter 6

119.62(1), Z = 4, and V = 606.62(4) Å3, which is isostructural with Re2Si4N6C (Re = Ho,

Tb). YTbSi4N6C is a very efficient green-emitting phosphor material under UV excitation

~ 300 nm, in which the green emission originates from the 5D4 7FJ (J = 6, 5, 4, 3, 2)

transitions. An unusual long-wavelength 4f - 5d excitation band of Tb3+ at about 300 nm

is observed due to the high covalency of the silicon-nitride-carbide network. Thus,

YTbSi4N6C can be effectively excited at rather long-wavelength. The energy transfer

from Ce3+ to Tb3+ is observed in Ce3+-doped YTbSi4N6C. YTb0.99Ce0.01Si4N6C is the first

compound, in which a green Tb3+ emission can be realized by excitation Ce3+ in the

visible range (390 – 480 nm) due to energy transfer from Ce3+ to Tb3+, demonstrating that

this new material is an interesting phosphor for white-light LED applications.

References:

1. Henning A. Höppe, Gunter Kotzyba, Rainer Pottgen and Wolfgang Schnick, J. Mater.

Chem., 2001, 11, 3300.

2. K. Liddell, D.P. Thompson, J. Mater. Chem., 2001, 11, 507.

3. K. Liddell, D.P. Thompson, T. Brauniger, R.K. Harris, J. Eur. Ceram. Soc., 2005,

25, 37.

4. K. Liddell, D.P. Thompson, S.J. Teat, J. Eur. Ceram. Soc., 2005, 25, 49.

5. H. Huppertz and W. Schnick, Angew. Chem. Int. Ed. Engl. 1996, 108, 2115.

6. H. Huppertz and W. Schnick, Z. Anorg. Allg. Chem., 1997, 623, 212.

7. H. Huppertz and W. Schnick, Acta Cryst. 1997, C53, 1751.

8. Henning A. Höppe, Henning Trill, Gunter Kotzyba, Bernd D. Mosel, Rainer Pottgen

and Wolfgang Schnick, Z. Anorg. Allg. Chem., 2004, 630, 224.

9. C. M. Fang, Y.Q. Li, H.T. Hintzen and G. de With, J. Mater. Chem., 2003, 13, 1480.

10. Y.Q. Li, G. de With and H.T. Hintzen, J. Alloys Comp., 2004, 385, 1.

11. Y.Q. Li, C.M. Fang, G. de With and H.T. Hintzen, J. Solid State Chem., 2004, 177,

4687.

12. W. Schnick, R. bettenhausen, B. Gotze, H.A. Hoppe, H. Huppertz, E. Irran, K.

Kollisch, R. Lauterbach, M. Orth, S. Rannabauer, T. Schlieper, B. Schwarze,

119

Chapter 6

F. Wester, Z. Anorg. Allg. Chem., 2003, 629, 902.

13. H.T. Hintzen, K.V. Ramanujachary, Y.Q. Li, A.C.A Delsing, to be published.

14. Y.Q. Li, G. de With and H.T. Hintzen, unpublished results.

15. Thommy C. Ekstrom, Kenneth J.D. MacKenzie, Martin J. Ryan, Ian W.M. Brown

and G. Vaughan White, J. Mater. Chem., 1997, 7, 505.

16. H.M. Rietveld, J. Appl. Crystallogr., 1969, 2, 65.

17. A.C. Larson and R.B. Von Dreele, Report LAUR 86-748, Los Alamos National

Laboratory, Los Alamos, NM, 2000.

18. B. H. Toby, J. Appl. Cryst. 2001, 34, 210.

19. Michael Woike, Wolfgang Jeitschko, J. Solid State Chem., 1997, 129, 312.

20. Materials Data, Inc., MDI Jade version 5.0.

21. G. Petzow, M. Herrmann, Silicon nitride ceramics. In: D.M.P. Mingos (ed.) Struct

Bond (High Performance Non-Oxide Ceramics II.), 2002, 102, 47.

22. F. Liebau, Structural Chemistry of Silicates, Springer, Berlin, 1985.

23. G. Blasse and B.C. Grabmaier, Luminescent materials, Springer-Verlag, Berlin,

1994.

24. R.D. Shannon, Acta Cryst., 1976, A32, 751.

25. T. Jüstel, H. Nikol and C. Ronda, Angew. Chem. Int. Ed. 1998, 37, 3084.

26. C. Feldmann, T. Jüstel, C.R. Ronda and P.J. Schmidt, Adv. Funct. Mater., 2003,

13, 511.

27. A.M. Srivastave and C.R. Ronda, The Electrochem. Soc., Interface, 2003, 48.

28. Y.Q. Li, G. de With and H.T. Hintzen, to be published (Chapter 7).

120

Chapter 7

Structure and luminescence properties of Ce3+-doped Y2Si4N6C ABSTRACT The structure and luminescence properties of undoped and Ce3+-doped yttrium silicon-

nitride-carbide, Y2Si4N6C, are reported. The crystal structure of Y2Si4N6C, prepared by a

solid-state reaction from Y metal, α-Si3N4 and SiC at 1650 oC in N2 atmosphere, has been

determined by X-ray powder diffraction and refined by Rietveld profile analysis.

Y2Si4N6C crystallizes in the monoclinic cell with the space group P21/c, a = 5.9339(1) Å,

b = 9.8925(2) Å, c = 11.8870(3) Å, β = 119.62(1)°, and Z = 4. Ce3+-doped Y2Si4N6C

shows an unusual long-wavelength 4f-5d excitation band of Ce3+ in the range of 380 -

450 nm due to the highly covalent silicon-nitride-carbide network combined with large

crystal field splitting due to coordinating N3- ions. For excitation in the UV-blue range

(370 – 450 nm), Y2Si4N6C:Ce3+ gives rise to a green emission in the range of 530 – 560

nm showing high promise for use as a conversion phosphor in white-emitting LEDs.

Keywords: Crystal structure, X-ray powder diffraction, Rietveld refinement,

Luminescence, Yttrium silicon-nitride-carbide, Cerium, White-emitting LEDs.

121

Chapter 7

7.1. Introduction

Light emitting diodes (LEDs) have greatly impacted our daily life, from traffic lights,

outdoor signs to automobile lights and backlights of liquid crystal displays and mobile

phone. In particular, the efficiency of white-light LEDs has increased significantly during

the last years, and white LED lighting has already shown high potential to replace the

traditional lamps (i.e. it has already surpassed the incandescent lamps and is competitive

with fluorescent lamps) [1-5].

Conversion phosphor materials play a key role in the type of white-emitting LEDs

based on gallium-indium nitride (InGaN). Apart from high quantum efficiency and high

stability, a perfect wavelength-matching with the excitation source is a challenging

requirement. That means that phosphors should strongly absorb blue light (~ 465 nm)

from InGaN chip and efficiently convert this blue radiation into green, yellow and red

visible light. In addition, good thermal quenching behaviors are also necessary [4-5]. The

yellow emitting phosphor (Y,Gd)3(Al,Ga)5O12:Ce3+ is a well known example [6-8], which

is used for generating white light combined with a blue-emitting LED chip. Additionally,

a green- and a red-emitting phosphor in combination with a blue-emitting LED is an

alternative approach, such as green-emitting (Sr,Ca,Ba)(Al,Ga)2S4:Eu2+ and red-emitting

(Sr,Ca)S:Eu2+ phosphors [9]. In this way, white-light with high color temperature and

high color rendering index (CRI) can be achieved. However, a low conversion efficiency,

low chemical stability and high thermal quenching characteristics remarkably limit the

quality of the obtained white LED lighting [5]. To meet the requirements of the white-

light LEDs, recently, some novel nitride-based phosphor materials with improved

properties have been developed [10-23]. In particular, the red-emitting phosphor

M2Si5N8:Eu2+ (M = Ca, Sr, Ba) has demonstrated to be an excellent conversion-phosphor

for white LED lighting which strongly absorbs the primary blue light and then converts it

into light in the orange-red spectral range [17]. However, so far, stable green-emitting

phosphors (520 - 560 nm) which can efficiently be excited by UV- blue light (370 – 450

nm) are very limited. Tailorable green emission can be realized by doping Eu2+ or Ce3+

ions into suitable host lattices because the 5d excitation band generally depends on the

local structure around Eu2+ or Ce3+ (i.e. the type of ligand and coordination number).

122

Chapter 7

Nevertheless, a large amount of experiments have proved that it is very difficult to

extend the dominant 5d excitation band of Eu2+ or Ce3+ in most traditional host lattices

like oxide or fluoride based materials from the normal UV into visible range (370 – 450

nm). In contrast, nitride-based materials have opened a new area by breaking down the

above mentioned restriction due to its higher covalency and larger crystal field splitting.

Examples of suitable host lattices are alkaline-earth-silicon-nitrides [11, 15-19], alkaline-

earth-silicon-oxynitrides and α-Ca-Sialon [11, 14, 20, 21, 22] and rare-earth silicon

oxynitrides [12, 13]. Most recently, we have shown that rare-earth silicon-nitride-carbides

also have high potential for such applications [23, 24]. For example, in Ce3+-doped

YTbSi4N6C, we can observe Tb3+ green line-emission by excitation of Ce3+ in the visible

range (390 – 480 nm) due to the energy transfer from Ce3+ to Tb3+. Sufficient resonant

energy transfer between Ce3+ and Tb3+ is believed to be responsible for this unusual

characteristic, evidenced from the fact that the Ce3+ emission band (5d 4f transition)

partially overlaps with the Tb3+ excitation band (i.e. 7F6 5D4 transition) [24].

Furthermore, a longer wavelength Ce3+ excitation band (~ 425 nm) in the spectrum of

YTbSi4N6C:Ce3+ is very attractive. Accordingly, complete replacement of Tb with Y in

YTbSi4N6C and using Ce3+ as an activator is of great interest. Y2Si4N6C has already been

reported [25] and was suggested to be analogous in structure with the MYbSi4N7 series

(M= Sr, Ba, Eu) [26-28]. More recently, Liddel et al. further reported that the Y2Si4N6C

phase was similar to La2Si4N6C and indexed its structure on an orthorhombic unit cell

with the lattice parameters a = 5.9677(7) Å, b = 10.2648(13) Å, c = 9.8937(13) Å [29,

30]. However, the detailed structure is still unclear up to date. According to our previous

work, YCeSi4N6C [23] is isostructural with MReSi4N7 (M = Sr, Ba; Re = Yb, Y)) [15, 16,

26-28], while YTbSi4N6C [24] is isostructural with Ln2Si4N6C (Ln = Ho, Tb) [31]. Based

on the fact that the ionic radius of Y3+ (0.9 Å, CN = 6) is significantly smaller than that of

Ce3+ (1.01 Å, CN = 6) and similar to Tb3+ (0.923 Å, CN = 6) [32], it is expected that the

structure of Y2Si4N6C is more close to YTbSi4N6 (i.e. Ln2Si4N6C (Ln = Ho, Tb [31]))

rather than YCeSi4N6C (i.e. MLnSi4N7, M = Ba, Sr; Ln = Y, Yb [15,16,26-28]).

Clarification of the crystal structure of Y2Si4N6C and its relationship with YCeSi4N6C

and YTbSi4N6C can be helpful for further improvement of the luminescence properties of

123

Chapter 7

these nitride-carbide materials. In the present study, we therefore investigate the structure

and the luminescence properties of undoped and Ce3+-doped Y2Si4N6C compounds.

7.2. Experimental 7.2.1. Preparation

Undoped and Ce3+-doped Y2Si4N6C compounds were synthesized by a high

temperature solid-state reaction method. The Y2Si4N6C compound was prepared using Y

metal powder (Csre, ~ 97%), α-Si3N4 (Permascand, P95H, α content 93%; Oxygen

content: ~1.5%), β-Si3N4 (Cerac S-1177, content: 91%, Oxygen content: ~ 0.7%, with

purity 99.5%) and SiC (Alfa, 99%) as starting materials. The appropriate amounts of Y

metal powder, Si3N4 and SiC powders were weighed out, mixed and ground in an agate

mortar. All manipulations were performed in a dry glove box flushed with dry nitrogen

because some starting materials are air and moisture sensitive. The powder mixtures were

then transferred into a closed molybdenum crucible and fired at 1650 °C for 10 h in a

horizontal tube furnace under nitrogen atmosphere to form the desired compound. After

firing, the samples were cooled down in the furnace.

Ce3+-doped solid-solutions of Y2-xCexSi4N6C (x = 0.01 – 0.2) were prepared with the

same processes using Ce (Alfa, 99%, lumps) as a starting material. The Ce3+

concentrations were restricted below 10 mol% with respect to Y in Y2Si4N6C in order to

keep structural stability.

7.2.2. Structure determination

All measurements were performed on finely ground powder samples, which were

analyzed by X-ray powder diffraction (Rigaku, D/MAX-B) using Cu-Kα radiation at 40

kV and 30 mA with a graphite monochromator.

The phase formation and purity was checked with a scan speed 2 °/min in the range of

10 – 70 2θ. With respect to the structure determination, the X-ray diffraction data were

collected from 10-120o 2θ with 0.02° step size using CuKα radiation with a counting time

12 s per step. Structure refinement was carried out by the Rietveld method [33], using the

program GSAS [34, 35]. The structures of Ho2Si4N6C [31], MYSi4N7 (M = Sr, Ba) [15,

16] and La2Si4N6C [30] were used as the initial models for the refinement of Y2Si4N6C.

124

Chapter 7

Impurity phases like Y2Si3O3N4 and SiC were also taken into account in the course of the

refinement. The refined parameters include the scale factor, zero shift, background, lattice

parameters, peak profile parameters, fractional coordinates of individual atoms, and

isotropic displacement parameters.

The lattice parameters of Ce3+-doped Y2Si4N6C were determined by a least-squares

method from X-ray diffraction data collected between 10 and 90° 2θ in step scan with a

step size of 0.02 o in 2θ and a count time of 10 seconds using about 15 wt% silicon

powder as an internal standard.

7.2.3. Optical measurements

The diffuse reflectance, excitation and emission spectra of the samples were measured

at room temperature by a Perkin Elmer LS 50B spectrophotometer equipped with a Xe

flash lamp. The reflection spectra were calibrated with the reflection of black felt

(reflection 3%) and white barium sulfate (BaSO4, reflection ~100%) in the wavelength

region of 230-700 nm. The excitation and emission slits were set at 2.5 nm. The emission

spectra were corrected by dividing the measured emission intensity by the ratio of the

observed spectrum of a calibrated W-lamp and its known spectrum from 300 to 900 nm.

Excitation spectra were automatically corrected for the variation in the lamp intensity by

a second photomultiplier and a beam-splitter. All the spectra were measured with a scan

speed of 100 nm/min.

7.3. Results and discussion

7.3.1. Synthesis and phase formation

Similar to YTbSi4N6C [24], we found that it is difficult to obtain high phase-pure

Y2Si4N6C compound when using β-Si3N4 powder (normally total oxygen content < 1

wt%) as a starting material or when applying a prolonged firing time in an open crucible

under nitrogen-hydrogen (10 vol%) atmosphere. On the contrary, Y2Si4N6C was readily

formed when using α-Si3N4 containing some more oxygen than β-Si3N4. Therefore, like

YTbSi4N6C, Y2Si4N6C seems also to be an oxygen-containing compound, which prefers

some oxygen incorporation to keep its structure stable.

125

Chapter 7

Unlike YTbSi4N6C and YCeSi4N6C [23], Y2Si4N6C is sensitive to air, which implies

that incorporation of a small amount of larger cation ions (i.e. Tb3+ and Ce3+) is favorable

for the chemical and structural stability.

7.3.2. X-ray powder diffraction data and structure of Y2Si4N6C

The X-ray diffraction pattern of most pure Y2Si4N6C compound shows that always

some traces of Y2Si3O3N4 and SiC present. The X-ray powder diffraction data of

Y2Si4N6C are listed in Table 7.1. Using Ho2Si4N6C as the starting model, the final

refinement gives a reasonable structure for Y2Si4N6C with good discrepancy R-factors

(Table 7.2). On the contrary, when we used MYSi4N7 and La2Si4N6C as initial model, the

refinement of Y2Si4N6C could not converge. Therefore, as we expected, the structure of

Y2Si4N6C is similar to that of YTbSi4N6C [24] which is isostructural with Re2Si4N6C (Re

= Ho, Tb) [31]. The refined structure parameters are summarized in Table 7.2. Bond

valence sum (BVS) calculations [36, 37] for Y, Si, N and C in Y2Si4N6C based on the

refined parameters (Table 7.2) support our refinement. As expected, the unit cell volume

of Y2Si4N6C (598.92 Å3) slightly shrinks compared to YTbSi4N6C (606.62 Å3) [24] and

Tb2Si4N6C (609.84 Å3) [31] due to the smaller Y3+ ions instead of the larger Tb3+ ions.

Fig. 7.1 shows the good agreement between the observed and calculated powder

diffraction patterns of Y2Si4N6C.

The structural characteristic of Y2Si4N6C is that its framework is built up by corner-

sharing SiN3C tetrahedra, of which the carbon atom bridges four SiN3C tetrahedra and

forms units [31]. Two different Y (I) and Y (II) sites are present, and both

of them are located in the channels (along [100] axis) formed by the six-ring silicon-

nitride-carbide network with average Y-N distances of 2.372 (Y(I)) and 2.398 (Y(II)) Å,

respectively. Y(I) is coordinated by five nitrogen atoms, while Y(II) ions are coordinated

by six nitrogen atoms at a distance between 2.1 and 2.6 Å. Correspondingly, the

coordination polyhedron volumes of Y(I)N

[2]3 4[C(SiN ) ]

5 and Y(II)N6 calculated by the program

IVTON [38] are 9.299 ± 0.031Å3 and 15.433 ± 0.051Å3, respectively, showing largely

different Y sites (Fig. 7.2). In addition, the nearest distance between Y and C is about

126

Chapter 7

Table 7.1. X-ray diffraction data for Y2Si4N6C. h k l 2 θexp(°) dexp(Å) I/Io(%) h k l 2 θexp(°) dexp(Å) I/Io(%) 0 0 2 0 2 0 0 1 2 0 2 1 -1 2 1 1 1 1 -1 1 3 0 3 1 1 2 1 1 0 2 -1 1 4 0 3 2 -2 1 1 0 0 4 -2 0 4 -2 2 2 -1 3 3 -2 2 1 1 1 3 0 4 2 2 1 1 -2 2 5 1 0 4 -1 0 6 1 3 3 2 1 2 -2 3 5 2 2 2 -3 2 4 0 4 4 -3 2 5 3 0 0 1 3 4 -2 4 5

17.099 17.801 19.299 19.761 23.280 24.419 24.841 28.619 29.059 29.980 31.500 32.101 32.581 34.681 35.140 35.561 35.880 36.480 39.242 40.358 41.240 44.521 46.460 46.760 47.360 47.680 49.261 50.439 50.919 51.158 53.260 53.539 54.780 55.700

5.1813 4.9787 4.5955 4.4889 3.8178 3.6422 3.5813 3.1165 3.0703 2.9780 2.8377 2.7860 2.7460 2.5844 2.5517 2.5225 2.5007 2.4610 2.2939 2.2330 2.1873 2.0334 1.9529 1.9411 1.9179 1.9058 1.8482 1.8078 1.7919 1.7841 1.7185 1.7102 1.6743 1.6489

8.0 3.3 16.0 3.5 2.1 3.5 3.3 2.0 6.0 33.8 2.5 100.0 2.8 5.8 39.9 6.5 38.2 2.8 2.9 8.5 1.7 9.6 1.6 2.8 2.9 6.4 7.5 2.2 2.1 1.1 10.4 4.4 16.7 3.4

0 6 1 -3 2 0 0 6 2 -3 1 7 1 4 4 3 2 1 0 6 3 1 3 5 1 6 2 1 0 6 2 4 3 -2 4 7 2 3 4 2 6 1 -4 2 1 -4 2 7 -4 0 8 1 5 5 1 6 4 -2 7 0 -3 5 7 1 7 3 -3 3 9 -4 1 9 -2 6 7 -5 1 6 -4 5 2 1 7 4 1 0 8 -1 4 9 -1 1 10 -5 1 8 2 5 5 -5 2 2

56.639 56.939 58.781 60.019 60.579 62.201 62.580 63.000 64.641 65.300 67.059 67.401 69.261 70.879 71.899 72.300 74.220 75.001 75.700 77.200 78.139 79.160 79.839 80.821 81.839 82.760 83.819 84.960 86.539 87.100 87.519 88.060 89.300 89.840

1.6237 1.6159 1.5696 1.5401 1.5272 1.4912 1.4831 1.4742 1.4407 1.4278 1.3945 1.3883 1.3555 1.3284 1.3121 1.3058 1.2767 1.2653 1.2554 1.2347 1.2222 1.2089 1.2003 1.1882 1.1760 1.1652 1.1532 1.1406 1.1238 1.1180 1.1137 1.1083 1.0961 1.0909

10.9 5.3 14.3 1.9 2.1 6.9 9.9 2.8 4.8 1.3 1.2 5.2 1.4 1.6 5.5 3.8 1.6 1.1 2.4 1.1 1.2 < 1 1.3 1.4 < 1 < 1 1.1 1.4 1.4 1.1 1.9 1.4 1.5 1.5

3.439Å, indicating no direct coordination between Y and C. Therefore, no direct

influence is expected of C on the rare-earth ions incorporated on the Y sites. However, a

higher covalency in the silicon-nitride-carbide network is expected compared with a

silicon-nitride network without carbon.

127

Chapter 7

Table 7.2. Structural parameters, refined atomic parameters and calculated bond-valence sums (BVS) for Y2Si4N6C. Formula Space group Lattice constants Formula units per cell RwpRp χ2

Y2Si4N6C P21/c (No. 14) Monoclinic a = 5.9073(2) Å b = 9.8560(2) Å c = 11.8364(3) Å β = 119.65(1) V = 598.92(4) Å3

Z = 4 0.089 0.062 9.6

Atom Wyck. S.O.F. x/a y/b z/c U [Å2] BVS*

Y1 4e 1 0.3313(7) 0.5667(7) 0.0940(3) 0.0231 3.068 (2%)

Y2 4e 1 0.6662(8) 0.4146(6) 0.4220(4) 0.0229 3.441 (15%)

Si1 4e 1 -0.0031(7) 0.4761(5) 0.2469(8) 0.0178 4.366 (9%)

Si2 4e 1 0.1633(8) 0.2114(8) 0.4243(8) 0.0160 3.582 (-10%)

Si3 4e 1 0.1982(2) 0.2142(8) 0.1779(7) 0.0218 4.595 (15%)

Si4 4e 1 0.6778(2) 0.2267(9) 0.1573(8) 0.0223 3.649 (-9%)

N1 4e 1 0.0356(5) 0.1941(10) 0.0259(9) 0.0134 3.097 (3%)

N2 4e 1 0.0563(3) 0.0496(8) 0.4041(7) 0.0185 3.255 (8%)

N3 4e 1 0.2378(3) 0.0194(8) 0.2239(7) 0.0078 2.927 (-2%)

N4 4e 1 0.4770(4) 0.2885(8) -0.0150(2) 0.0201 3.021 (1%)

N5 4e 1 0.5000(4) 0.2527(5) 0.2396(5) 0.0077 3.125 (4%)

N6 4e 1 0.6890(3) 0.0651(9) 0.1362(6) 0.0077 2.965 (-1%)

C1 4e 1 0.0310(6) 0.3058(5) 0.2475(8) 0.0249 4.311 (8%) * The data in parentheses represent a deviation from the ideal valence states.

In comparison with YTbSi4N6C, the bond valence sums for Y, Si and N are similar. In

contrast, the bond valence sum for C is largely different (3.992 vs. 4.311 for YTbSi4N6C

and Y2Si4N6C, respectively). It implies that the SiCN3 tetrahedra in YTbSi4N6C are more

regular than in Y2Si4N6C as reflected in the overbonding characteristic of C. It is also

reflected by the standard deviations of the Si-C distances within the SiCN3 tetrahedra

128

Chapter 7

10 20 30 40 50 60 70

SiCY2Si3O3N4

Inte

nsity

(cou

nts)

2θ (deg)

Y2Si4N6C

Fig. 7.1. Observed (crosses), calculated (solid line) X-ray powder diffraction patterns and

the difference profile of the Rietveld refinement of Y2Si4N6C.

Fig. 7.2. Coordination polyhedron of the Y atoms and the Y-N distances (Å) in Y2Si4N6C.

129

Chapter 7

which are 0.069 Å and 0.188 Å for YTbSi4N6C and Y2Si4N6C, respectively. Therefore, a

larger distortion at the C sites caused mainly by the short C-Si distances can be one of the

reasons for structural instability of Y2Si4N6C compared with the YTbSi4N6C compound.

7.3.3. Incorporation of Ce3+ in Y2Si4N6C

Fig. 7.3 shows the relationship between the unit cell volume of Y2-xCexSi4N6C and

the x value (x = 0 - 0.2). As expected, with the Ce concentration increasing, the unit cell

volume of Y2-xCexSi4N6C becomes larger due to the incorporation of the Ce3+ ions larger

than Y3+ [32]. The unit cell volume almost keeps constant above x = 0.1 indicating that

the maximum solubility of Ce3+ in Y2Si4N6C is around this x value, corresponding with 5

mol% with respect to Y3+. Indeed, we found that the amount of second phase was

significantly increased as Ce3+ amounts larger than 5 mol% were added. This effect can

be well understood from the different crystal structures of Y2Si4N6C and YCeSi4N6C [23].

Furthermore, this conclusion is also confirmed by the luminescence spectra (i.e. emission

spectra vs. x value) of Y2-xCexSi4N6C (see next section). The β angle, the a/c and b/c

ratios are almost constant (~ 0.497 and ~ 0.833 for a/c and b/c, respectively) versus the

Ce concentration.

0.00 0.05 0.10 0.15 0.20596

598

600

602

604

606

608

610

Uni

t cel

l vol

ume

(Å3 )

x

Fig. 7.3. The unit cell volume dependence of x in Y2-xCexSi4N6C.

130

Chapter 7

As described above, there are two different crystallographic Y sites in the Y2Si4N6C

lattice (Table 7.2). Considering the structure of YCeSi4N6C in which Y and Ce occupy

unequivalent crystallographic sites rather than sharing the two available sites (while Ce is

located at the large site [24]), therefore, the dopant Ce3+ ions most likely preferentially

occupy the largest Y site in Y2Si4N6C. Moreover, the total lattice energy calculated by

GULP (General Utility Lattice Program) [39] shows that Ce on Y(II) site has a lower

lattice energy (-311.68 eV) than Ce on Y(I) site (-311.66 eV), and significantly lower

than Ce equal distribution on the two sites (-310.65 eV) in Y1.98Ce0.02Si4N6C. This is also

expected that Ce incorporates on the larger Y site from size difference reasons between Y

and Ce [32]. The luminescence data obtained for Ce3+-doped Y2Si4N6C (see next section)

further demonstrate that the Ce3+ ions mainly occupy one Y site (i.e. the larger Y(II) site,

see Fig. 7.2).

7.3.4. Luminescence properties of Ce3+-doped Y2Si4N6C

Fig. 7.4 shows the diffuse reflection spectra of undoped and Ce3+-doped Y2Si4N6C

compounds. Clearly, a strong absorption band around 260 – 290 nm is readily assigned to

the valence to conduction band transitions of the Y2Si4N6C host lattice. The band gap of

undoped Y2Si4N6C derived from the reflection spectrum is about 280 nm which is larger

than that of MYSi4N7 (~ 250 nm for M = Sr and 260 nm for M = Ba) [15, 16]. This is fair

agreement with its observed dark-grey daylight color. In contrast, Ce3+-doped Y2Si4N6C

shows a strong blue-green color in daylight. In the reflection spectrum of Ce3+-doped

Y2Si4N6C, there are two obvious absorption bands centered at about 380 nm and 427 nm

which are definitely related to the Ce3+ absorption because no absorption band in this

range can be found for undoped Y2Si4N6C. In addition, a weak absorption shoulder at

about 495 nm can possibly also be assigned to the Ce3+ ions, because with increasing

Ce3+ concentration this absorption shoulder becomes stronger (Fig. 7.4).

The excitation spectrum of Y2Si4N6C:Ce3+ shows three bands peaking at about 284,

388 and 425 nm corresponding with the bands observed in the diffuse reflection spectra

(Fig. 7.5). The low intensity band with its maximum at about 284 nm corresponds to the

host lattice excitation possibly overlapping with Ce3+ excitation (see Fig.4). The two

strong excitation bands at about 388 and 425 nm arise from the transition from the 4f

131

Chapter 7

200 300 400 500 600 700

0

10

20

30

40

50

60

Ref

lect

ion

(%)

Wavelength (nm)

Y2Si4N6C Y1.98Ce0.02Si4N6C Y1.9Ce0.1Si4N6C

Fig. 7.4. Diffuse reflection spectra of undoped and Ce3+-doped Y2Si4N6C.

200 300 400 500 600 700 800

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4 x = 0.02 x = 0.10 x = 0.20

Inte

nsity

(a.u

.)

Wavelength (nm)

Fig. 7.5. Excitation (left) and emission (right) spectra of Y2-xCexSi4N6C (λexc = 425 nm

and λem = 540 nm for x = 0.02; λexc = 425 nm and λem = 560 nm for x = 0.10, 0.20).

132

Chapter 7

ground state to the 5d levels of the excited Ce3+ ions. The long-wavelength excitation

band at about 425 nm is rather unusual, as general, the Ce3+ ions can be efficiently

excited in the range of 250 - 400 nm, i.e. the 5d excitation band of Ce3+ normally is at

higher energy [40]. Except for YAG:Ce3+ [40, 41] and MS:Ce3+ (M = Ca, Sr, Ba) based

materials [42, 43], so far no other Ce3+-doped materials have been found which can be

efficiently excited at wavelengths above 400 nm. Y2Si4N6C:Ce3+ is another example,

showing high promise for white-light LED applications. In contrast to YAG:Ce3+ (low

energy 5d band due to strong crystal-field splitting effect) and MS:Ce3+ (low energy 5d

band due to strong nephelauxetic effect) [40], in Y2Si4N6C:Ce3+, these two effects are

believed to take place simultaneously. On one hand, the carbon atoms will increase the

covalency of the silicon-nitride-carbide network and thus makes the nephelauxetic effect

stronger; On the other hand, a large crystal-field splitting is induced by N ions with a -3

formal charge higher than -2 for O (YAG) or S (CaS).

Ce3+-doped Y2Si4N6C shows a green emission with maxima in the range of 530 – 560

nm depending on the Ce3+ concentration (Fig. 7.5). The broad emission band in the range

of 450 – 800 nm is ascribed to the transition from the lowest energy crystal field splitting

component of the 5d level to the 4f ground state of Ce3+ [40]. The absence of a clear

doublet characteristic of the Ce3+ emission (i.e. the transition from 5d states to 2F5/2 and 2F7/2 4f ground state levels) indicates that the Ce3+ ions experience a strong crystal-field

effect in Y2Si4N6C [44]. By varying the excitation wavelength the same Ce3+ emission

band was found implying that the Ce3+ ions probably occupy a single Y site, most likely

the large Y(II) site. As described above, this assignment is reasonable because Y(II)

possesses a larger coordination number of 6 together with a larger coordination

polyhedron volume of 15.433 ± 0.051Å3 which is favorable for large Ce3+ ions. On an

energy scale, the Ce3+ emission band of 1 mol% Ce3+-doped Y2Si4N6C is decomposed

into two well-separated Gaussian components with maxima at about 17240 and 19084

cm-1 (corresponding to 580 and 524 nm, respectively. see Fig. 7.6) which roughly match

with the energy difference between the 2F5/2 and 2F7/2 ground state levels of Ce3+ (normal

∆E (2F5/2 - 2F7/2) ≈ 2000 – 2200 cm-1 [40]). Additionally, a long tail extending to lower

energy (higher wavelength) in the emission spectra also suggests that probably a very

small amount of Ce3+ enters into the Y(I) site (see Fig. 7.6), for which the shift to longer

133

Chapter 7

12000 14000 16000 18000 20000 22000 24000

0.0

0.2

0.4

0.6

0.8

1.0

800 750 700 650 600 550 500 450

Emis

sion

inte

nsity

(a.u

.)

Wavenumber (cm-1)

524

580

Wavelength (nm)

Fig. 7.6. The deconvoluted emission spectrum of Y1.98Ce0.02Si4N6C as a sum of two

Gaussian bands (λexc = 425 nm).

wavelength is expected.

With increasing Ce3+ concentrations, a red-shift of the emission band is obviously

observed changing its maximum from 535 to 560 nm (Fig. 7.6). This shift of the emission

band terminates at Ce concentrations surpassing 5 mol% in agreement with our previous

conclusion that the maximum solubility of Ce3+ in Y2Si4N6C is about 5 mol% (x = 0.1).

Whereas, the 5d subbands do not shift (as observed in the excitation spectra), indicating

that the center of gravity and the crystal-field splitting are negligibly dependent on the Ce

concentration. Because the change of the Stokes shift is very small ~ 200 cm-1 (i.e., ~

4400 and 4600 cm-1 for x = 0.02 and 0.1 in Y2-xCexSi4N6C, respectively), thus Ce3+

reabsorption process is mainly responsible for this red-shift of the emission band.

Evidently, as the intensity of the excitation band peaking at about 388 nm decreases,

another excitation band at about 425 nm becomes intense for higher Ce concentration.

Correspondingly, the emission band shifts to longer wavelengths at the expense of

decreasing the emission intensity (Fig. 7.5).

134

Chapter 7

7.4. Conclusions

Y2Si4N6C, synthesized by a solid state reaction, is the monoclinic cell with the space

group P21/c, a = 5.9339(1) Å, b = 9.8925(2) Å, c = 11.8870(3) Å, β = 119.62(1)° and Z =

4, which is isostructural with Ln2Si4N6C (Ln = Ho, Tb). Under excitation in the UV-blue

to visible range of 370 – 450 nm, Y2Si4N6C:Ce3+ shows an efficient green emission in the

range of 530 – 560 nm. The emission and, in particular, the excitation bands are at low

energy due to specific characteristics of the nitride-carbide host lattice (nephelauxetic

effect and crystal field effect). The efficient absorption of the blue-light from GaInN-

based LEDs and subsequent conversion of it into green-light demonstrates the high

potential of Y2Si4N6C:Ce3+ for white-light LED applications.

References:

1. S. Nakamura, Appl. Phys. Lett., 1994, 64, 1687.

2. S. Aanegola, J. Petroski and E. Radkov, SPIE, 2003, 10, 16.

3. Y. Narukawa, Optics & Photonics News, 2004, 4, 25.

4. L.S. Rohwer, A.M. Srivastava, The Electrochemical Society Interface, 2003, 36.

5. J.Y. Taso, Ed., in Light Emitting Diodes (LEDs) for General Illumination Update

2002, Optoelectronics Industry Development Association, Washington, DC (2002).

6. P. Schlotter, R. Schmidt, J. Schneider, Appl. Phys, 1997, A64, 417.

7. P. Schlotter, J. Baur, Ch. Hielscher, M. Kunzer, H. Obloh, R. Schmidt, J. Schneider,

Mater. Sci. Eng., 1999, B59, 390.

8. S. Nakamura, Appl. Phys. Lett., 1994, 64, 1687.

9. R. Mueller-Mach, G.O. Mueller, M.R. Krames, J. IEEE, 2002, 8, 339.

10. H.T. Hintzen and Y.Q. Li, Encyclopedia of Materials: Science and Technology,

2004, 1.

11. J.W.H. van Krevel, Ph.D. Thesis, Eindhoven University of Technology, 2000

12. J.W.H. van Krevel, H.T. Hintzen, R. Metselaar, and A. Meijerink, J. Alloys

135

Chapter 7

Comp., 1998, 268, 272

13. J.W.H. van Krevel, H.T. Hintzen, R. Metselaar, Mater. Res. Bull., 2000, 35, 747.

14. J.W.H. van Krevel, J.W.T. van Rutten, H. Mandal, H.T. Hintzen, and R. Metselaar,

J. Solid State Chem., 2002, 165, 19.

15. Y.Q. Li, G. de With and H.T. Hintzen, J. Alloys Comp., 2004, 385, 1.

16. Y.Q. Li, C.M. Fang, G. de With and H.T. Hintzen, J. Solid State Chem., 2004, 177,

4687.

17. H.T. Hintzen, J.W.H. van Krevel and G. Botty, European Patent EP-1104799

A1,1999.

18. H.A. Höppe, H. Lutz, P. Morys, W. Schnick and A. Seilmeier, J. Phys. Chem.

Solids, 2000, 61, 2001.

19. Y.Q. Li, G. de With and H.T. Hintzen, J. Lumin., in press.

20. Y.Q. Li, A.C.A. Delsing, G. de With and H.T. Hintzen, Chem. Mater., 2005, 17,

3242.

21. R. J. Xie, N. Hirosaki, K. Sakuma, Y. Yamamoto and M. Mitomo, Appl. Phys.

Lett., 2004, 84, 5404.

22. K. Sakuma, K. Omichi, N. Kimura, M. Ohashi, D. Tanaka, N. Hirosaki, Y.

Yamamoto, R.J. Xie, T. Suehiro, Optics Letters, 2004, 17, 2001.

23. H.T. Hintzen, K.V. Ramanujachary, Y.Q. Li, C.M. Fang, A.C.A Delsing, to be

published.

24. Y.Q. Li, G. de With and H.T. Hintzen, to be published (Chapter 6).

25. K. Liddell, D.P. Thompson, J. Mater. Chem., 2001, 11, 507.

26. H. Huppertz and W. Schnick, Angew. Chem. Int. Ed. Engl. 1996, 108, 2115.

27. H. Huppertz and W. Schnick, Z. Anorg. Allg. Chem., 1997, 623, 212.

28. H. Huppertz and W. Schnick, Acta Cryst. 1997, C53, 1751.

29. K. Liddell, D.P. Thompson, T. Brauniger, R.K. Harris, J. Eur. Ceram. Soc., 2005,

25, 37.

30. K. Liddell, D.P. Thompson, S.J. Teat, J. Eur. Ceram. Soc., 2005, 25, 49.

31. H.A. Höppe, G. Kotzyba, R. Pöttgen and W. Schnick, J. Mater. Chem., 2001,

11, 3300.

32. R.D. Shannon, Acta Cryst., 1976, A 32, 751.

136

Chapter 7

33. H.M. Rietveld, J. Appl. Crystallogr., 1969, 2, 65.

34. A.C. Larson and R.B. Von Dreele, Report LAUR 86-748, Los Alamos National

Laboratory, Los Alamos, NM, 2000.

35. B. H. Toby, J. Appl. Cryst. 2001, 34, 210.

36. I.D. Brown, Structure and Bonding in Crystals, vol 2, ed M. O’Keefe and

A. Navrotsky, New York: Academic, PP 1- 30, 1981.

37. N.E. Brese and M. O’Keefe, Acta Crystallogr., 1991, B 47, 192.

38. T. Balic Zunic, I. Vickovic, J. Appl. Cryst. 1996, 29, 305.

39. J.D. Gale, JCS Faraday Trans., 1997, 93, 629.

40. G. Blasse and B.C. Grabmaier, Luminescent materials, Springer-Verlag, Berlin,

1994.

41. G. Blasse and A. Bril, Appl. Phys. Lett., 1967, 11, 53.

42. B. Huttl, U. Troppenz, K.O. Velthaus, C.R. Ronda, R.H. Mauch, J. Appl. Phys., 1995,

78, 7282.

43. L. Eichenauer, B. Jarofke, H.C. Mertins, J. Dreyhsig, W. Busse, H.E. Gumlich,

P. Benalloul, C. Barthou, J. Benoit, C. Fouassier and A. Garcia, Phys. Stat. Sol.,

(a), 1996, 153, 515.

44. T.R.N. Kutty, Mater. Res. Bull., 1990, 25, 343.

137

Chapter 8

Luminescence properties of Eu2+-doped MAl2-xSixO4-xNx (M = Ca, Sr, Ba)

conversion phosphor for white-LED applications

ABSTRACT

Undoped and Eu-doped MAl2-xSixO4-xNx (M = Ca, Sr, Ba) were synthesized by a solid-

state reaction method at 1300-1400 °C under nitrogen-hydrogen atmosphere. The

solubility of (SiN)+ in MAl2O4 was determined. Nitrogen can be incorporated into

MAl2O4 by replacement of (AlO)+ by (SiN)+ pair, whose amount of solubility depends on

the M cation. The solubility of (SiN)+ is very low in CaAl2O4 and SrAl2O4 (x ≈ 0.025 and

0.045, respectively), whereas, a large amount of (SiN)+ can be incorporated into BaAl2O4

(x ≈ 0.6). Incorporation of (SiN)+ hardly modifies the luminescence properties of Eu2+-

doped MAl2O4 (M = Ca, Sr) because of limited solubility of (SiN)+, showing the blue and

green emission at almost constant wavelength of 440 and 515 nm, respectively. Eu2+-

doped BaAl2-xSixO4-xNx exhibits a broad green emission with a maximum in the range of

500 - 526 nm depending on the concentration of (SiN)+ and Eu2+. In addition, both

excitation and emission bands of Eu2+ show a significant red-shift as nitrogen is

incorporated. BaAl2-xSixO4-xNx:Eu2+ can be efficiently excited in the range of 390 - 440

nm radiation, which makes this material attractive as conversion phosphor for white-LED

lighting applications.

Keywords: alkaline-earth aluminates; silicon-aluminium-oxynitride; europium; tridymite

structure; X-ray powder diffraction; luminescence properties; conversion phosphor;

white-light LEDs

138

Chapter 8

8.1. Introduction

As an important class of phosphors Eu2+-doped MAl2O4 (M = Ca, Sr, Ba) has been

widely used as persistent luminescent materials because of their high efficiency, chemical

stability and long-lasting high-luminance afterglow characteristics [1-3]. These phosphors

also have been proposed for plasma display panel (PDP) [4] and optoelectronic

applications [5]. Under ultraviolet and cathode-ray excitation these phosphors show a

strong blue (M = Ca), green (M = Sr) and blue-green (M = Ba) fluorescence.

In MAl2O4 (M = Ca, Sr, Ba), the three-dimensional framework is built up by a motif of

six-rings formed by corner-sharing AlO4 tetrahedra. The tetrahedral framework is

isostructural with the SiO2 polymorph having tridymite structure [6-8]. The various

MAl2O4 structures differ in the arrangement and the number of crystallographic sites of

the divalent cations within the channels formed by the AlO4 rings. CaAl2O4 has a

monoclinic structure in space group P21/n. In this structure there are three Ca sites: two

of them are six-fold and the third is nine-fold coordinated with the oxygen atoms in a

different channel. However, for the larger M cations, SrAl2O4 and BaAl2O4 crystallize in

a monoclinic and a hexagonal system with space group P21 and P63, respectively. Both

Sr and Ba ions occupy two crystallographic sites, which are located in the channels of the

AlO4 rings, each with nine-fold coordination with the oxygen anions.

Although extensive investigations have been done on Eu2+-doped alkaline earth

aluminates MAl2O4 (M = Ca, Sr, Ba) [9-15], in previous studies considerable attention

has been paid to improve the performance of these phosphors by partial substitution of

the M ion (for example, replacement of Ca with Sr and vice versa [10-11]) and the Al ion

(i.e. partial replacement of Al by B [12]) and/or by co-doping activator ions (for example,

Eu2+ with Dy3+) [15]. These effects can enhance the efficiency, optimize the emission

color range and extend the persistence time. However, it is only possible to tailor the

excitation and absorption bands in the UV range. As a consequence, for obtaining an

efficient emission these phosphor materials have to be excited in the UV region (e.g.

below 350 nm). Therefore, with respect to white light-emitting diode (LED) applications,

the excitation bands of Eu2+-doped MAl2O4 (M = Ca, Sr, Ba) phosphors do not match

with the UV-blue emission (~ 370 - 460 nm) from InGaN-based LEDs. For use as

wavelength conversion-phosphors for white-light LEDs, usually a strong absorption in

139

Chapter 8

the UV-blue range (i.e. 370 - 460 nm) and highly efficient conversion from absorbed blue

into green, yellow and red light is required [16]. Hence, in order to make MAl2O4:Eu2+

(M = Ca, Sr, Ba) phosphors meet these requirements, apart from the above-mentioned

routes other novel approaches have to be adopted. Recently, Eu2+-doped alkaline-earth-

silicon-nitride [17, 18] and oxynitride [19] have shown an unusual long wavelength

emission with excitation in the visible range (370 – 460 nm). Therefore, if silicon and

nitrogen atoms can be incorporated into MAl2O4, e.g. (AlO)+ replacement by (SiN)+

(which is an opposite routine to convert nitride into oxynitride, for example, (SiN)+

(AlO)+ replacement in Si3N4 SiAlON [20] and Y2Si3O3N3 Y2Si3-xAlxO3+xN4-x [21]),

it is expected that MAl2-xSixO4-xNx:Eu2+ will extend the excitation bands into the visible

range and emit at longer wavelengths, i.e. green and yellow emission, due to the oxide

conversion into oxynitride lattice. Modification of the framework of MAl2O4:Eu2+ (M =

Ca, Sr, Ba) has already proved an efficient approach to improve its luminescence for

white-light LED applications [22].

Therefore, in this work, we synthesized undoped and Eu2+-doped MAl2-xSixO4-xNx (0 ≤

x < 2, M = Ca, Sr, Ba) materials by a solid state reaction and investigated the existence

region of MAl2-xSixO4-xNx compounds with stuffed tridymite structure. The effect of the

substitution of (SiN)+ for (AlO)+ on the phase formation and crystal structure was studied

by X-ray powder diffraction combined with the Rietveld refinement. Finally, the

luminescence of Eu2+-doped MAl2-xSixO4-xNx (M = Ca, Sr, Ba) and the dependence of

luminescence properties on Eu2+ concentration in BaAl2-xSixO4-xNx were also investigated.

8.2. Experimental

8.2.1. Starting materials

MCO3 (M = Ca, Sr, Ba) (Merck, > 99.0%), SiO2 (Degussa Aerosil OX50), γ-Al2O3

(AKPG, > 99.995), α-Si3N4 (SKW Trostberg, α content 23.3%, O content 0.7 wt%) and

Eu2O3 (Rhône-Poulenc, 99.99%) were employed as the raw materials. Oxygen presence

in the Si3N4 starting powder was not considered in the synthesis procedures.

140

Chapter 8

8.2.2. Synthesis of undoped and Eu2+-doped MAl2-xSixO4-xNx (M = Ca, Sr, Ba)

Si3N4 was used as the source of (SiN)+ using the following reaction:

MCO3 + (2-x)/2 Al2O3 + x/4 Si3N4 + x/4 SiO2 MAl2-xSixO4-xNx + CO2 (1)

The raw materials were homogeneously wet-mixed in the appropriate amounts by a

planetary ball mill for 4 – 5 hours in isopropanol with agate balls in an agate container.

After mixing the mixture was dried in a stove and ground in an agate mortar.

Subsequently, the powders were fired in Mo or alumina crucibles at 1300-1400 °C for 8 -

12 h in a reducing atmosphere of N2 - H2 (10%) in a horizontal tube furnace for two times

with an intermediate grinding in between the firing steps. The same processes were

adopted for obtaining Eu-doped materials.

8.2.3. Characterization

The obtained samples were analyzed by X-ray powder diffraction on a Rigaku

D/Max-γB diffractometer operating at 40 kV, 30 mA with Bragg-Brentano geometry (flat

graphite monochromator, scintillation counter) using CuKα radiation. Phase formation

was checked by a routine scan (2 °/min). The lattice parameters were determined in the

2θ range of 10-90 ° with step scan mode using silicon powder as an internal standard with

a step size of 0.01° 2θ and a counting time of 6 s per step. In order to correlate the

changes of the local structures with the luminescence properties, the structure of BaAl2-

xSixO4-xNx was refined by the Rietveld method [23] using structural parameters of

BaAl2O4 [8] as the starting model, assuming both Si4+ and N3- random distributing over

the Al3+ and O2- sites, respectively, in BaAl2O4. For the Rietveld refinement XRD data

were recorded with step scan mode within a 2θ range of 10-120° with a step size of 0.01°

2θ and a counting time of 15 s per step. Rietveld refinement was performed using the

program GSAS [24, 25].

The photoluminescence spectra were determined at room temperature on the powder

samples by a Perkin-Elmer LS-50B luminescence spectrometer with a Xenon discharge

lamp as excitation source. The radiation was detected by a red sensitive photomultiplier

R928. The spectra were obtained in the range of 200 – 700 nm with a scanning speed of

141

Chapter 8

100 nm/min and excitation and emission slit widths of 2.5 nm. Excitation spectra were

automatically corrected for the variation in the lamp intensity by a second photomultiplier

and a beam-splitter; and all the emission spectra were corrected by taking into account

the combined effect of the spectral response of the R928 detector and the monochromator

by using the measured spectra of a calibrated W-lamp as the light source. Diffuse

reflectance spectra were recorded in the range of 230 – 700 nm with BaSO4 white powder

(~ 100%) and black felt (3%) as the references.

8.3. Results and discussion

8.3.1. Effect of (SiN)+ substitution for (AlO)+ in MAl2O4 (M = Ca, Sr, Ba) on phase

formation and structure

When nitrogen is incorporated in MAl2O4, (AlO)+ is expected to be replaced by the

(SiN)+ pair to form hybrid (Al,Si)-(O,N)4 tetrahedra in the framework. As a proof, the

lattice parameters are expected to decrease corresponding to the unit cell volume

shrinkage because of shorter Si-N[2] distances (~ 1.65 - 1.75 Å [26], N[2] denotes nitrogen

bridging two silicon atoms) as compared to the Al-O[2] distances (~ 1.70 - 1.78 Å [27 -

29], O[2] denotes the oxygen bridging two aluminum atoms) in MAl2O4. With the ionic

radius of M decreasing from Ba to Ca, it is found that the incorporation of nitrogen

according to the reaction (1) becomes more difficult. As a consequence the maximum

solubility of (SiN)+ in MAl2O4 significantly decreases from the Ba to Sr and Ca

compounds. In the case of CaAl2O4 and SrAl2O4, the solubility of (SiN)+ is almost

negligible. The obtained lattice parameters of MAl2-xSixO4-xNx as a function of x

demonstrate that the maximum solubility of (SiN)+ in CaAl2O4 and SrAl2O4 lattice is

about x ≈ 0.025 (i.e. 1.25 mol%) and x ≈ 0.045 (i.e. 2.25 mol%), respectively (Fig. 8.1).

At high temperatures the two systems show complex solid-state reactions, it seems

that the tendency of preferential reaction taking place among the starting materials

yielding aluminium silicate phases becomes the dominating factor compared to that of

forming the solid solutions of MAl2-xSixO4-xNx (M = Ca, Sr), especially for larger

amounts of SiO2 related to more N incorporation (equation 1). Clearly, as expected the

lattice parameters decrease with increasing x up till the solubility limit (i.e. x ≈ 0.025 and

142

Chapter 8

0.045 for Ca and Sr, respectively). When the x value surpasses the maximum solubility

value, the lattice parameters nearly keep constant (Fig. 8.1) while secondary phases

appear. From the X-ray diffraction patterns a secondary phase of Ca2Al2SiO7 or

Sr2Al2SiO7 can be identified for the CaAl2-xSixO4-xNx and SrAl2-xSixO4-xNx, respectively.

0.00 0.02 0.04 0.06 0.08 0.10 0.121064

1065

1066

1067

1068

1069

1070

U

nit c

ell v

olum

e (Å

3 )

X

(a)

0.00 0.02 0.04 0.06 0.08 0.10383.0

383.1

383.2

383.3

383.4

383.5

383.6

Uni

t cel

l vol

ume

(Å3 )

X

(b)

Fig. 8.1. Relationship between the unit cell volume and x values of (a) CaAl2-xSixO4-xNx

and (b) SrAl2-xSixO4-xNx

As two extreme examples, Fig. 8.2 displays the X-ray powder diffraction patterns for

MAl2-xSixO4-xNx (M = Ca, Sr) with x = 0.1. It can be clearly seen that apart from CaAl2O4

143

Chapter 8

10 15 20 25 30 35 40 45 50

C aA l2 -xS ixO 4 -xN x (x = 0 .1 )

C a 2A l2S iO 7

Inte

nsity

(Cou

nts)

2 θ (de g .)

(a )

***C aA l2O 4

10 15 20 25 30 35 40 45 50

Sr2A l2S iO 7

Inte

nsity

(Cou

nts)

2θ (deg .)

S rA l2 -xS ixO 4 -xN x (x = 0 .1 ) (b )

SrA l2O 4

Fig. 8.2. X-ray powder diffraction pattern of MAl2-xSixO4-xNx (x = 0.1). The bars below

the diffraction patterns indicate the positions of Bragg reflections for MAl2O4 and

MAl2SiO7. (a) M = Ca, (b) M = Sr. * indicates unknown phase.

144

Chapter 8

and SrAl2O4 a large amount of secondary phases like Ca2Al2SiO7 and Sr2Al2SiO7 (or their

nitrogen containing solid solutions) are present in these two cases. Even for firing at 1600

˚C the solubility of (SiN)+ in CaAl2O4 and SrAl2O4 is limited.

In contrast, the lattice parameters and in particular the unit cell volume of BaAl2-

xSixO4-xNx significantly decrease with increasing x value up to 0.6 indicating that (SiN)+

is effectively incorporated into BaAl2O4 lattice with a high ratio (Fig. 8.3).

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.98.748.768.788.808.828.848.868.88

10.36

10.38

10.40

10.42

10.44

10.46

816

818

820

822

824

826

828

830

832

834

836

a c

a, c

(Å)

x in BaAl2-xSixO4-xNx

V

Uni

t cel

l vol

ume

(Å3 )

Fig. 8.3. Relationship between the lattice parameters of BaAl2-xSixO4-xNx and x.

For x values larger than 0.6, the unit cell volume remains almost constant and a distinct

secondary phase of Ba2SiO4 is found, thus the maximum solubility of (SiN)+ in BaAl2O4

is about x = 0.6 (Fig. 8.3). Consequently, the observed XRD pattern of BaAl2-xSixO4-xNx

perfectly matches with the calculated pattern based on BaAl2O4 tridymite structure [8]

(Fig. 8.4). Although, as expected, both the a and c axes decrease with increasing x (i.e.

(SiN)+ content) (Fig. 8.3), the c/a ratio of BaAl2-xSixO4-xNx is almost constant (~ 0.843).

The mean Al-O distance is similar in the three compounds, i.e.1.755 Å, 1.752 Å and

1.757 Å for CaAl2O4, SrAl2O4 and BaAl2O4, respectively [6-8]. But a larger difference

between the shortest (1.665 Å) and the longest (1.808 Å) Al-O bonds is observed for the

BaAl2O4 lattice. As a result, there are two types of tetrahedral (AlO4) units with high

145

Chapter 8

distortion in BaAl2O4. These largely distorted (AlO4) tetrahedra are probably very well

compatible with incorporation of a Si-N pair without changing its structure too much. In

contrast, in CaAl2O4 and SrAl2O4 all the tetrahedral (AlO4) units are very regular (i.e. the

Al-O distances are very similar with only a small deviation for each compound [6, 7]). In

addition, with the cation size increasing from Ca to Ba, it is expected that the lattice

becomes softer in the sequence Ca - Sr - Ba. The combination of the above mentioned

two factors may explain the larger solubility of (SiN)+ in the BaAl2O4 lattice.

When (AlO)+ is replaced by (SiN)+ in BaAl2O4, the average (Si,Al)-(O,N) distances,

obtained by the Rietveld refinement, decrease for larger (SiN)+ amounts corresponding to

an overall shrinkage of the lattice. For example, 1.7534 Å, 1.7529 Å and 1.7431 Å,

respectively, for x = 0, 0.1, 0.3 in BaAl2-xSixO4-xNx. At the same time, however, the

average Ba-(O, N) distances slightly increase (i.e. 2.918 Å, 2.924 Å for x = 0, 0.3,

respectively), indicating that Ba needs more space due to coordination with N (larger than

O). It is worth noting that, for a better understanding the site preferences of Si (on the

four available Al sites) and N (on the six available O sites) in BaAl2O4, neutron

diffraction experiments need to be performed due to the similar scattering factors of N3-

/O2- and Al3+/Si4+ for X-ray powder diffraction.

8.3.2. Luminescence properties of Eu-doped MAl2-xSixO4-xNx (M = Ca, Sr, Ba)

The luminescence properties of Eu-doped MAl2-xSixO4-xNx strongly depend on the

types of the cation M similar to the case of MAl2O4:Eu2+ (M = Ca, Sr, Ba) [9-15]. While

the position of the Eu2+ excitation and emission bands is nearly independent of x for M =

Ca and Sr, it strongly depends on x for M = Ba. Overview results of the obtained

luminescence properties (i.e., excitation, emission and the Stokes shift) are listed in Table

8.1.

8.3.2.1. MAl2-xSixO4-xNx:Eu2+ (M = Ca, Sr)

For CaAl2-xSixO4-xNx and SrAl2-xSixO4-xNx, as described above, the solubility of (SiN)+

is very low (Table 8.1), hence it is expected that the excitation and emission spectra of

the Eu-doped compounds have no significant change compared with Eu-doped MAl2O4

(M = Ca, Sr). Indeed, both the excitation and emission behaviour are so similar that the

146

Chapter 8

10 20 30 40 50 60 70 80 90 100 110 120

0

4

8

12

16In

tens

ity (C

ount

s x

104 )

2θ (deg.)

Fig. 8.4. The X-ray diffraction pattern for BaAl2-xSixO4-xNx (x = 0.3). Plus (+) marks

represent the observed intensities, and the solid line is the calculated pattern. A difference

(obs.– calc.) plot is shown in the bottom. The bars above the difference profile indicate

the positions of Bragg reflections for BaAl2O4 with the tridymite structure.

maximal shift of Eu2+ emission bands is less than 10 nm with increasing x (Fig. 8.5),

reflecting the fact that the solubility of (SiN)+ in MAl2-xSixO4-xNx (M = Ca, Sr) is

negligible. Therefore, such limited (SiN)+ incorporation can not efficiently modify the

local coordination of the Eu2+ activator.

8.3.2.2. BaAl2-xSixO4-xNx:Eu2+

With the content of incorporated (SiN)+ increasing, an additional excitation band

appears for BaAl2-xSixO4-xNx:Eu2+ (10 mol%), peaking at 425 – 440 nm for x values

above 0.3 (Fig. 8.6). Correspondingly, the broad emission band shifts to a longer

147

Chapter 8

Table 8.1. Overview of structural parameters of undoped and luminescence data of 10 mol%Eu-doped MAl2-xSixO4-xNx (M =

Ca, Sr, Ba).

MAl2-xSixO4-xNx

M = Ca M = Sr M = Ba

Maximum solubility of (SiN)+ x = 0.025 x = 0.045 x = 0.6 Structural parameters Monoclinic P21/n Monoclinic P21 Hexagonal P63 x x = 0 x = 0.02 x = 0 x = 0.02 x = 0 x = 0.3

a (Ǻ) b (Ǻ) c (Ǻ)

β (o) V (Å3)

8.6808(3) 8.0928(4) 15.1950(8) 90.26(1) 1067.47(8)

8.6714(4) 8.0923(7) 15.1979(3) 90.28(1) 1066.45(6)

8.4435(8) 8.8184(9) 5.1575(7) 93.40(1) 383.35(10)

8.4384(10) 8.8275(8) 5.1527(5) 93.32(2) 383.18(9)

10.4432(2)

8.79167(4)

830.371(7)

10.4224(2)

8.7992(3)

827.778(11) Excitation band (nm)

260, 329, 380 260,339, 380 260,340, 386, 420

260, 340, 386, 420

280, 340, 387 280, 340, 400, 440

Emission band (nm)

438 443 514 519 498 526

Stokes shift (cm-1)1

3500 3600 6500 6600 5800 3700

Crystal field splitting (cm-1) 2

13360 13600 14000 14000 10000 13000

1. Stokes shift calculated from the energy difference between the lowest 5d excitation band and emission band of Eu2+. 2. Crystal-field splitting estimated from the energy difference between highest and lowest observed 5d excitation levels of Eu2+.

148

Chapter 8

200 250 300 350 400 450 500 550 600 650

x= 0.10

x= 0.05

x= 0.02

Inte

nsity

(a.

u.)

Wavelength (nm)

x = 0

(a)

200 250 300 350 400 450 500 550 600 650 700 750

x= 0.10

x= 0.05

x= 0.02

Inte

nsity

(a.u

.)

Wavelength (nm)

x= 0

(b)

Fig. 8.5. Excitation (left) and emission (right) spectra of MAl2-xSixO4-xNx:Eu (10%) with

various x, (a) M = Ca (λexc = 340 nm; λem = 450 nm); (b) M = Sr (λexc = 387 nm; λem =

520 nm).

149

Chapter 8

200 250 300 350 400 450 500 550 600 650 700

x= 0.8

x= 0.6

x= 0.5

x= 0.3

Inte

nsity

(a.u

.)

Wavelength (nm)

x= 0

Fig. 8.6. Excitation (left) and emission (right) spectra of BaAl2-xSixO4-xNx: Eu (10%) with

various x (λexc = 390 nm, λem = 500 nm for x = 0; and λexc = 440 nm, λem = 530 nm for x

= 0.3 – 0.8).

wavelength from 498 to 527 nm until up to x = 0.6 (Fig. 8.6), which is also consistent

with our observation of a high solubility of (SiN)+ in BaAl2O4 (Table 8.1 and Fig. 8.3).

Since the BaAl2O4 lattice becomes more rigid when more covalent nitrogen is introduced

into the three-dimensional (Al, Si)(O, N)4 framework, it is evident that the Stokes shift

shows a decrease in the x range from 0 to 0.3 (Table 8.1). Therefore, we then can readily

attribute this red-shift of the Eu2+ emission band to a concomitant shift of the excitation

band due to an increase of the crystal field splitting of the 5d state of Eu2+ ions as a

consequence of the replacement of O2- by N3-. Also an increase of the degree of covalent

bonding contributes to this shift (Table 8.1). For x > 0.3, the excitation bands show a

slight blue-shift, especially for x > 0.5 (Fig. 8.6), also in agreement with the above

mentioned structural changes showing that the EuBa-O/N distances become larger as the

150

Chapter 8

amount of (SiN)+ increases counteracting the effect of the replacement of O2- by N3-.

Besides the replacement of (AlO)+ by (SiN)+, as usual the Eu2+ concentration also

shows a significant influence on the structure and luminescence properties of BaAl2-

xSixO4-xNx:Eu2+. Fig. 8.7 shows the lattice parameters changes of BaAl2-xSixO4-xNx (x =

0.3) with the Eu concentration.

0.00 0.02 0.04 0.06 0.08 0.108.768.778.788.798.808.818.828.83

10.38

10.39

10.40

10.41

10.42

10.43

10.44

10.45

826

827

828

829

830

831

832

a c

a, c

(Å)

x in Ba1-yEuyAl1.7Si0.3O3.7N0.3

V

Uni

t cel

l vol

ume

(Å3 )

Fig. 8.7. Relationship between the lattice parameters of Ba1-yEuyAl1.7Si0.3O3.7N0.3 and the

Eu concentration.

As expected, the lattice parameters decrease with increasing the Eu concentration because

the ionic radius of Eu2+ (1.30 Å for CN = 9) is much smaller than that of Ba2+ (1.47 Å for

CN = 9) [30]. The incorporation of (SiN)+ just makes a very small red-shift (~ 7 nm) of

the host lattice absorption edge (Fig. 8.8). The broad absorption bands superimposed on

the absorption curve of the host lattice can be assigned to the Eu2+ ions in the spectral

range of 300 – 500 nm. With the Eu content increasing from 1 to 10 mol%, the absorption

edge of Eu2+ extends from 400 to 460 nm, meanwhile its absorption intensity becomes

151

Chapter 8

intense (Fig. 8.8). Clearly, the principal excitation band shifts to long-wavelength (i.e.

400 - 440 nm) at high Eu concentration (Fig. 8.9) in agreement with the reflection spectra

(Fig. 8.8). This can be understood from shrinkage of the lattice (Fig. 8.7), which induces

larger crystal field splitting. Excitation in the range of 400 – 440 nm yields a green

emission with maximum at about 500 - 526 nm depending on the Eu content. In addition,

the red-shift of the emission band is attributed to a larger crystal-field splitting (i.e., 8600

cm-1 vs. 12000 cm-1 for 1% and 10% Eu, respectively) originated from shorting the BaEu -

O/N bond and the Stokes shift (e.g., 2800 cm-1 vs. 3700 cm-1 for 1% and 10% Eu,

respectively). The integrated emission intensity increases with increasing the Eu content

for long excitation wavelength (440 nm), but decreases for short excitation wavelength

(390 nm) (see inset in Fig. 8.9).

200 300 400 500 600 700

0

20

40

60

80

100

Ref

lect

ion

(%)

Wavelength (nm)

BaAl2O4

BaAl2-xSixO4-xNx

x=0.3, y=0.01 x=0.3, y=0.05 x=0.3, y=0.10

Fig. 8.8. Reflection spectra of BaAl2-xSixO4-xNx (x = 0, 0.3) and Ba1-yEuyAl2-xSixO4-xNx:

Eu2+ (x = 0.3, y = 0.01, 0.05, 0.10 corresponding to 1%, 5% and 10 mol% Eu).

152

Chapter 8

As the substitution of (SiN)+ for (AlO)+ is favorable to enhance the excitation or

absorption band in the UV-blue to visible range (i.e. 400 – 450 nm), Eu-doped BaAl2-

xSixO4-xNx is an attractive conversion phosphor for InGaN-based LED lighting

applications.

200 300 400 500 600 700 800

0

50

100

150

200

0.00 0.02 0.04 0.06 0.08 0.10150

200

250

300

350

400

450

500

Inte

nsity

(a.u

.)

Wavelength (nm)

y = 0.01 y = 0.05 y = 0.10

λexc = 390 nm λexc = 440 nm

Emis

sion

inte

nsity

(a.u

.)

x

Fig. 8.9. Excitation (left) and emission (right) spectra dependence of y in Ba1-

yEuyAl1.7Si0.3O3.7N0.3 (λexc = 390 nm, λem = 500 nm for x = 0.01; λexc = 440 nm, λem = 515

nm for x = 0.05; and λexc = 440 nm, λem = 530 nm for x = 0.10). Inset shows the

integrated emission intensity as a function of the Eu concentration (λexc = 390 nm and

λexc = 440 nm).

153

Chapter 8

8.4. Conclusions

The maximum solubility of (SiN)+ in MAl2-xSixO4-xNx with tridymite structure

significantly decreases from Ba to Sr and Ca. In CaAl2O4 and SrAl2O4, the solubility of

(SiN)+ is very limited (i.e. x ≈ 0.025 (1.25%) and x ≈ 0.045 (2.25%) respectively),

whereas the maximum solubility of (SiN)+ in the BaAl2O4 lattice is about x ≈ 0.6 (i.e.

30%). As a consequence the Eu2+ emission is found at 440 and 515 nm for Eu-doped

MAl2-xSixO4-xNx (M = Ca, Sr), similar to the compounds without incorporation of

nitrogen. BaAl2-xSixO4-xNx:Eu2+ exhibits a long-wavelength excitation band peaking at

about 440 nm corresponding to a green emission at about 500 -526 nm (x ≈ 0.3). This

red-shift of both excitation and emission bands due to the incorporation of nitrogen can

be understood from increased covalency and crystal field splitting. Therefore, the

luminescence properties of BaAl2-xSixO4-xNx:Eu2+ can be modified by adjusting the

amount of (SiN)+ and the Eu concentration. Consequently, BaAl2-xSixO4-xNx:Eu2+ shows

very promise as a green-emitting phosphor for white-light LED applications.

References:

1. G. Blasse and B.C. Grabmaier, Luminescent materials, Springer-Verlag, Berlin, 1994.

2. F.C. Palilla, A.K. Levine, M.R. Tomkus, J. Electrochem. Soc., 1968, 115, 642.

3. V. Abbruscato, J. Electrochem. Soc., 1971, 118, 930.

4. S.Tanaka, I.Ozaki, T. Kunimoto, K. Ohmi and H. Kobayashi, J. Lumin., 2000, 87-89,

1250.

5. H. Yamamoto and T. Matsuzawa, J. Lumin., 1997, 72-74, 287.

6. W. Horkner and H.K. Müller-Buschbaum, J. Inorg. Nucl. Chem., 1976, 38, 983.

7. A.R. Schulze and H.K. Müller-Buschbaum, Z. Anorg. Allg. Chem, 1981, 475, 205.

154

Chapter 8

8. W. Horkner and H.K. Müller-Buschbaum, Z. Anorg. Allg. Chem, 1979, 451, 40.

9. G. Blasse, and A. Bril, Philips Res. Rep., 1968, 23, 201.

10. J. Holas, H. Jungner, M. Lastusaari and J. Niittykoski, J. Alloy and Compd., 2001,

323-324, 326.

11. T. Aitasalo, J. Holas, H. Jungner, J.C. Krupa, M. Lahtinen, M. Lastusaari,

J. Legendziewica, J. Niittykoski, and J. Valkonen, Radiat. Eff. Defects Solids,

2003, 158, 309.

12. J. Niittykoski, T. Aitasalo, J. Holas, H. Jungner, M. Lastusaari, M. Parkkinen,

M.Tukia, J. Alloy and Compd., 2004, 374, 108.

13. S.H. Ju, U.S. Oh, J.C. Choi, H.L. Park, T.W. Kin and C.D. Kim, Mater. Res. Bull.,

2000, 35, 1831.

14. D. Ravichandran, S.T. Johnson, S. Erdei, R. Roy and W.B. White, Displays, 1999, 19,

197.

15. T. Matsuzawa, Y.Aoki, N. Takeuchi and Y. Murayama, J. Electrochem. Soc., 1996,

143, 2670.

16. J.Y. Taso, Ed., in Light Emitting Diodes (LEDs) for General Illumination

Update2002, Optoelectronics Industry Development Association, Washington,

DC (2002).

17. J.W.H. van Krevel, Ph.D. Thesis, Eindhoven University of Technology, 2000.

18. H.A. Höppe, H. Lutz, P. Morys, W. Schnick and A. Seilmeier, J. Phys. Chem.

Solids, 2000, 61, 2001

19. J.W.H. van Krevel, J.W.T. van Rutten, H. Mandal, H.T. Hintzen, and R. Metselaar,

J. Solid State Chem., 2002, 165, 19.

20. S. Hampshire, H.K. Park and D.P. Thompson, Nature, 1978, 274, 880.

21. J.W.H. van Krevel, H.T. Hintzen and R. Metselaar, Mater. Res. Bull., 2000, 5, 35.

22. H.T. Hintzen and Y.Q. Li, WO 2004/029177 A1, 2004.

23. H.M. Rietveld, J. Appl. Crystallogr., 1969, 2, 65.

24. A.C. Larson and R.B. Von Dreele, Report LAUR 86-748, Los Alamos National

Laboratory, Los Alamos, NM, 2000.

25. B. H. Toby, J. Appl. Cryst. 2001, 34, 210.

26. W. Schnick and H. Huppertz, Chem. Eur., J., 1997, 3, 679.

155

Chapter 8

27. K.H. Jack, J. Mater. Sci., 1976, 11, 1135.

28. K.H. Jack, Progress in Nitrogen Ceramics, ed. F.L. Riley, Martinus Nijhoff

Publishers, 1983.

29. W. Schnick, Inter. J. Inorg. Mater., 2001, 3, 1267.

30. R.D. Shannon, Acta Cryst., 1976, A 32, 751.

156

Chapter 9

Luminescence properties of Eu2+-activated alkaline earth silicon

oxynitride MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba):

a promising class of novel LED conversion phosphors

ABSTRACT

The luminescence properties of Eu2+-activated alkaline-earth silicon-oxynitrides have

been studied. In the BaO-SiO2-Si3N4 system, a new BaSi2O2N2 compound was obtained

having the monoclinic structure with lattice parameters a = 14.070(4) Å, b = 7.276(2) Å,

c = 13.181(3) Å, β = 107.74(6)°. All MSi2O2-δN2+2/3δ:Eu2+ (M = Ca, Sr, Ba) materials can

be efficiently excited in the UV to visible region (370 - 460 nm), making them attractive

as conversion phosphors for LED applications. A blue-green emission at 490-500 nm is

observed for BaSi2O2N2:Eu2+, a yellow emission at 560 nm for CaSi2O2-δN2+2/3δ:Eu2+ (δ ≈

0) and a green - yellow emission peaking from 530 to 570 nm for SrSi2O2-δN2+2/3δ:Eu2+ (δ

≈ 1) , the position depending on the exact value of δ. BaSi2O2N2:Eu2+ is the most

promising conversion phosphor for white-light LEDs due to its high conversion

efficiency for blue light from InGaN-based LEDs related to its very small Stokes shift.

Keywords: luminescence, alkaline-earth-silicon-oxynitride, europium, X-ray powder

diffraction, monoclinic, conversion phosphor, quantum efficiency, white-LEDs.

157

Chapter 9

9.1. Introduction

Since the invention of blue emitting InGaN-based white-light-diodes (LED), the

efficiency of white-light LEDs has been improved significantly. So far, the efficiency of

white-light LEDs has already surpassed that of incandescent lamps and is competitive

with fluorescent lamps. White-light LEDs shows high potential for replacement of

conventional lighting like incandescent and fluorescent lamps, the advantages being its

long life-time, saving energy consumption and its environmental-friendly characteristics

[1-5]. White-light LEDs can be realized by combining a InGaN-based diode with

phosphor materials, like YAG:Ce3+, from which white light is then produced by additive

mixing of yellow light emitted by the phosphor with blue light from the LED. Therefore,

the phosphor materials play an important role in white-light LEDs. However, with respect

to the presently used phosphors in white-light LED systems, most of them do not meet

the optimum requirements of white-light LEDs. For example, YAG:Ce3+ shows a high

thermal quenching and a poor color rendition which can be improved by sulfide-based

phosphors (i.e. red: SrS:Eu2+ and CaS:Eu2+; green: SrGa2S4:Eu2+). However, these sulfide

materials suffer from low chemical stability in LEDs environment. Ideally the conversion

phosphors for white-light LEDs must combine a high quantum efficiency and absorption

for UV-blue radiation with the ability to withstand the high temperature generated by the

LED without degrading and quenching the luminescence, and moreover should be

chemically stable. Thus, novel phosphor materials with improved properties are greatly in

demand.

Recently, some nitride-based phosphor materials have been invented with

unconventional properties for use in white-light LEDs [6-16]. Among these phosphors,

Eu2+-activated M2Si5N8 (M = Ca, Sr, Ba) is a new family of divalent europium doped

red- or red-orange-emitting alkaline-earth silicon nitride materials, which has proved to

be excellent phosphor materials for white-light LED application [10,17,18]. However,

very few efficient new phosphors with yellow and green emission have actually been

158

Chapter 9

found for white-light LEDs. As expected, the performances of white-light LEDs, such as

the color rendition index (CRI), color temperature and color range, can be significantly

improved by combination of the above mentioned red emitting phosphors and a green

emitting phosphor together with the blue light source from a InGaN-chip.

In contrast to the recently found alkaline-earth silicon nitride compounds M2Si5N8

(M = Ca, Sr, Ba), several alkaline-earth silicon oxynitride compounds, i.e. CaSi2O2N2 and

SrSi2O2N2, were reported earlier in the CaO-Si3N4-AlN and Sr-Si-O-N systems,

respectively [19-21]. Just recently, a single crystal structure determination was published

for CaSi2O2N2 [22]. However, the luminescence properties of rare-earth doped CaSi2O2N2

and SrSi2O2N2 have not been reported yet. Additionally, further extending to the

Ba-Si-O-N system is also very interesting like in the case of the alkaline-earth silicates

[23-28] and alkaline-earth silicon nitrides [6, 10, 11]. In the present study, we therefore

focus on the preparation and luminescence properties of Eu2+-doped MSi2O2-δN2+2/3δ (M

= Ca, Sr, Ba) compounds aiming at exploring new oxynitride-based phosphors for use in

white-light LEDs.

9.2. Experimental

9.2.1. Preparation

All powder samples of undoped and Eu2+-doped MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba)

were synthesized by a high temperature solid-state reaction. The starting materials were

high-purity MCO3 (M = Ca, Sr, Ba) (Merck, > 99.0%), SiO2 (Aerosil OX 50, Degussa),

Si3N4 (SKW Trostberg, β content: 23.3%, O ~ 0.7%) and Eu2O3 (Rhône-Poulenc,

99.99%). The Eu2+ mole fractions with respect to the M2+ ion range from 1% (x = 0.01) to

10% (x = 0.1). The starting materials were weighed out in various amounts (keeping the

M/Si ratio constant to 0.5), and subsequently homogeneously wet-mixed by a planetary

ball mill for 4 - 5 hours in isopropanol. After mixing the slurry was dried and ground in

an agate mortar. Subsequently, the dried powder mixtures were fired in a molybdenum or

159

Chapter 9

alumina crucibles at 1100-1400 °C for 6 - 12 h under a reducing atmosphere of N2 - H2

(10%) in horizontal tube furnaces. After firing, the samples were cooled down to room

temperature in the furnace and were ground again with an agate mortar.

9.2.2. X-ray powder diffraction

All final products were checked by X-ray powder diffraction (Rigaku, D/MAX-B)

using Cu-Kα radiation at 40 kV and 30 mA with a graphite monochromator. For phase

identification a normal scan (2° /min) was performed. The crystallographic data were

collected on the powder samples using a step scan mode with a step size of 0.02˚ and a

counting time of 10 second per step in the range 2θ 10 to 90˚. In order to avoid the

preferred orientation of the obtained samples, the powder samples were mounted into a

flat plate holder by the side filling method.

The unit cell of MSi2O2-δN2+2/3δ was determined from the X-ray powder diffraction

patterns using indexing programs DICVOL04 [29] for M = Ca, Sr and McMaill [30] (an

indexing program for X-ray powder diffraction based on Monte Carlo and grid search)

for M = Ba based on the first 20 lines for the search of solutions. The possible space

groups are determined according to the systematic absences and the obtained unit cells

are further examined by fitting the full profile X-ray powder diffraction patterns using Le

Bail method [31] within the program GSAS [32, 33].

9.2.3. Optical measurements

The diffuse reflectance, emission and excitation spectra of the samples were obtained

at room temperature by a Perkin Elmer LS 50B spectrophotometer equipped with a Xe

flash lamp. The reflection spectra were calibrated with the reflection of black felt

(reflection 3%) and white barium sulfate (BaSO4, reflection ~100%) in the wavelength

region of 230-700 nm. The excitation and emission slits were set at 2.5 nm. The emission

spectra were corrected by dividing the measured emission intensity by the ratio of the

observed spectrum of a calibrated W-lamp and its known spectrum from 300 to 900 nm.

160

Chapter 9

Excitation spectra were automatically corrected for the variation in the lamp intensity by

a second photomultiplier and a beam-splitter. All the spectra were measured with a scan

speed of 100 nm/min. Further the quantum efficiency (400 nm, 460 nm) was determined

as compared to the standard materials.

9.3. Results and discussion

9.3.1 Phase identification

In the BaO-SiO2-Si3N4 system, we obtained a single-phase compound with an

approximate composition BaSi2O2N2, which crystallizes in the monoclinic crystal system

with the lattice parameters: a = 14.070(4) Å, b = 7.276(2) Å, c = 13.181(3) Å, β =

107.74(6)° (Table 9.1).

Table 9.1. Lattice parameters of MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba).

Formula

Crystal system

Space group

Lattice constants

a (Å)

b (Å)

c (Å)

β (°)

V (Å3)

Figure-of-Merit

M(20)

F(20)

CaSi2O2N2 (δ ≈ 0)

Monoclinic

P21/C

15.035(4)

15.450(1)

6.851(2)

95.26(3)

1584.53

10.5

15.7(0.0088, 144)

SrSi2ON8/3 (δ ≈ 1)

Monoclinic

P21/M

11.320(4)

14.107(6)

7.736(4)

91.87(3)

1234.67

10.8

14.9(0.0090, 150)

BaSi2O2N2 (δ = 0)

Monoclinic

P2/M

14.070(4)

7.276(2)

13.181(3)

107.74(6)

1285.23

10.3

15.4( 0.0095, 137)

161

Chapter 9

The X-ray powder diffraction data are given in Table 9.2. For the CaO-SiO2-Si3N4

system, CaSi2O2N2 was formed as a nearly single phase material with always some traces

of Ca2SiO4 and CaSiO3. This suggests that the composition of CaSi2O2-δN2+2/3δ probably

may be somewhat more nitrogen rich than CaSi2O2N2, i.e. δ ≥ 0. This is supported by the

fact that the CaSi2O2N2 compound reported in the literature was prepared from

CaO-Si3N4 mixtures [22]. A similar, but more profound, behaviour was found for the

SrO-SiO2-Si3N4 system, where we could only obtain almost single-phase material when

completely omitting the SiO2 starting material and just starting with only SrO and Si3N4.

The approximate composition of this strontium silicon oxynitride compound thus is

SrSi2ON8/3 (δ ≈ 1). Fig. 9.1 shows the observed and simulated powder diffraction pattern

of the most pure MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba) compounds.

Table 9.2. X-ray powder diffraction data for BaSi2O2N2. h k l dobs (Å) 2θobs (deg.) 2θcal (deg.) ∆2θ (deg.) I/Io (%) 0 1 0

-2 0 1 0 1 1 2 0 1

-3 0 1 -3 0 2 -2 1 3 0 2 1

-1 2 1 3 1 1 4 0 0

-2 2 1 -1 1 4 0 1 4 5 0 0 0 0 5 4 1 2

-2 2 4 3 2 2

-5 1 4 4 0 3

7.2725 6.8465 6.3261 5.2790 4.6402 4.3143 3.6144 3.4942 3.4543 3.4050 3.3558 3.2098 2.9937 2.8815 2.6797 2.5116 2.4827 2.4225 2.4044 2.3104 2.2963

12.1600 12.9200 13.9880 16.7800 19.1110 20.5700 24.6100 25.4700 25.7700 26.1500 26.5400 27.7700 29.8200 31.0100 33.4100 35.7200 36.1500 37.0800 37.3690 38.9500 39.2000

12.1546 12.9325 14.0572 16.7757 18.9342 20.5774 24.6085 25.4713 25.7773 26.1446 26.5853 27.7529 29.8079 31.0061 33.4056 35.7313 36.1625 37.0784 37.3496 38.9341 39.2018

-0.0062 -0.0240 -0.0808 -0.0072 0.1653 -0.0189 -0.0098 -0.0126 -0.0187 -0.0059 -0.0567 0.0058 0.0008 -0.0074 -0.0067 -0.0224 -0.0235 -0.0094 0.0083 0.0049 -0.0128

84.7 20.2

< 1 < 1 2.3 2.6

100.0 36.0

4.5 20.8

5.9 21.0 29.8 80.8 22.2 12.7

6.2 21.5 14.3

4.0 6.9

162

Chapter 9

1 2 4 -3 3 1 -3 3 0 0 3 3 2 3 2

-7 0 2 3 0 5 2 3 3

-5 3 1 -3 1 7 -1 4 0 -1 4 1 5 0 4

-2 4 0 0 1 7 3 0 6

-3 4 1 -7 0 6 -6 3 3 5 3 2

-4 4 2 -4 4 0 -4 3 6 1 2 7 1 4 4 3 0 7

-1 4 5 5 2 5 6 0 5

-1 5 0 -7 2 7 4 4 3

-6 4 0 -10 0 2 -1 5 3 -10 1 4 -2 3 8 2 3 7 9 1 2

-10 0 6 -4 0 10 1 5 4

2.2495 2.1494 2.1310 2.0966 2.0715 2.0095 1.9478 1.9206 1.8371 1.8148 1.8017 1.7945 1.7863 1.7534 1.7416 1.7046 1.6973 1.6889 1.6620 1.6386 1.6061 1.5997 1.5799 1.5504 1.5352 1.5107 1.4932 1.4696 1.4618 1.4461 1.4403 1.4258 1.4098 1.3999 1.3803 1.3711 1.3603 1.3435 1.3361 1.3263 1.3138 1.2976

40.0500 42.0000 42.3800 43.1100 43.6600 45.0800 46.5910 47.2890 49.5800 50.2300 50.6200 50.8400 51.0900 52.1200 52.5000 53.7300 53.9800 54.2700 55.2200 56.0800 57.3200 57.5690 58.3600 59.5800 60.2300 61.3100 62.1100 63.2200 63.6000 64.3700 64.6600 65.4010 66.2400 66.7690 67.8400 68.3610 68.9800 69.9690 70.4100 71.0100 71.7900 72.8300

39.9989 41.9154 42.3736 43.0737 43.5730 45.0693 46.5426 47.2928 49.6847 50.2511 50.6018 50.8410 51.0711 52.0567 52.5091 53.6943 54.0411 54.1507 55.2538 55.9491 57.3267 57.6168 58.3587 59.6896 60.1958 61.2289 62.0983 63.2934 63.5824 64.3456 64.6476 65.4031 66.2085 66.7334 67.7859 68.3952 68.9807 69.9599 70.4387 70.9743 71.7456 72.8261

0.0402 0.0737 -0.0044 0.0254 0.0762 -0.0000 0.0377 -0.0145 -0.1152 -0.0316 0.0077 -0.0115 0.0083 0.0528 -0.0195 0.0253 -0.0715 0.1089 -0.0441 0.1207 -0.0169 -0.0580 -0.0089 -0.1197 0.0242 0.0711 0.0017 -0.0833 0.0078 0.0145 0.0025 -0.0119 0.0218 0.0258 0.0445 -0.0439 -0.0103 -0.0005 -0.0382 0.0263 0.0350 -0.0055

28.0 7.0 9.7 5.7 1.6 10.6

4.3 1.4 6.5 1.4 5.0 13.1 26.6

3.6 7.7 4.9 2.2

< 1 4.6 13.7

4.9 1.6 5.4 2.3 3.9 < 1

3.3 7.7 5.3 4.3 10.2

1.1 < 1 < 1 3.3 1.9 3.5 < 1 < 1 2.8 2.0 2.3

163

Chapter 9

-6 4 6 3 3 7 3 2 8

-6 1 10 -8 4 4 -7 0 10 -6 4 7 2 2 9

-1 5 6 -11 2 2 -10 3 5 4 5 4

-5 5 6 -12 0 5 1 2 10

-12 1 5 2 3 9

-4 6 3 3 0 10

-12 0 7 -6 4 9 10 3 2

1.2901 1.2828 1.2708 1.2528 1.2454 1.2368 1.2298 1.2151 1.2086 1.1995 1.1911 1.1777 1.1692 1.1596 1.1541 1.1455 1.1384 1.1302 1.1222 1.1123 1.1035 1.1011

73.3190 73.8090 74.6210 75.8810 76.4100 77.0400 77.5600 78.6800 79.1890 79.9090 80.5900 81.7000 82.4200 83.2510 83.7400 84.5100 85.1600 85.9300 86.6900 87.6600 88.5400 88.7800

73.3167 73.7669 74.5857 75.8409 76.3822 77.0496 77.5854 78.6604 79.1837 79.9124 80.5568 81.6880 82.4023 83.2207 83.8131 84.5126 85.1509 85.9273 86.6400 87.6793 88.5323 88.7791

-0.0071 0.0328 0.0261 0.0309 0.0186 -0.0187 -0.0345 0.0106 -0.0037 -0.0123 0.0243 0.0032 0.0090 0.0216 -0.0818 -0.0112 0.0005 -0.0058 0.0416 -0.0277 -0.0007 -0.0074

< 1 < 1 4.8 < 1 < 1 5.3 1.8 2.6 1.1 < 1 2.7 2.1 < 1 1.4 2.6 < 1 < 1 < 1 2.3 1.6 1.3 1.1

10 20 30 40 50 60 70 80 90

10 15 20 25 30 35 40 45

Inte

nsity

(Cou

nts)

2θ (deg.)

CaSi2O2-δN2+2/3δ

Inte

nsity

(Cou

nts)

2θ (deg.)

(a)

164

Chapter 9

10 20 30 40 50 60 70 80 90

10 15 20 25 30 35 40 45

Inte

nsity

(Cou

nts)

2θ (deg.)

SrSi2O2-δN2+2/3δ

Inte

nsity

(Cou

nts)

2θ (deg.)

(b)

10 20 30 40 50 60 70 80 90

10 15 20 25 30 35 40 45

Inte

nsity

(Cou

nts)

2θ (deg.)

BaSi2O2N2

Inte

nsity

(Cou

nts)

2θ (deg.)

(c)

Fig. 9.1. The observed ( ) and simulated (solid line) X-ray powder diffraction pattern of

MSi2O2-δN2+2/3δ (a) M = Ca, (b) M = Sr, (c) M = Ba. The difference profile (observed –

calculated) is shown at the bottom. The bars below the profile indicate the positions of all

the reflections allowed for MSi2O2-δN2+2/3δ.

165

Chapter 9

All MSi2O2-δN2+2/3δ compounds crystallize in a monoclinic unit cell but with

different space groups and lattice parameters for M = Ca, Sr, Ba (Table 9.1) [34]. It is

evident that the structure of BaSi2O2N2 is different from that of MSi2O2-δN2+2/3δ (M = Ca,

Sr), showing resemblances for Ca and Sr. Although the powder diffraction patterns of

MSi2O2-δN2+2/3δ (M = Ca, Sr) are essentially close to those previously reported for

CaSi2O2N2 and SrSi2O2N2 (low- temperature form) [20, 21], we found that these data are

inexact related to missing peaks and wrong indexing. In addition, the XRD pattern

belonging to the MSi2O2-δN2+2/3δ (M = Ca, Sr) compound depends on the composition of

the starting mixture. For example, for the nitrogen-rich SrSi2O2-δN2+2/3δ (δ ≈ 1) samples,

the strongest peak is at about 25.35° 2θ with the smallest amount of second phases. In

reverse, for oxygen-rich SrSi2O2-δN2+2/3δ (δ ≈ 0) the strongest peak is located at 31.69° 2θ.

The XRD data of our CaSi2O2-δN2+2/3δ (δ ≈ 0) powder sample could not be successfully

refined with the structural parameters determined for a CaSi2O2N2 single crystal by

Höppe et al. [22]. Probably several modifications of CaSi2O2N2 exist depending on the

temperature (the powder was prepared at 1400 °C, while the single crystal was obtained

by raising the temperature up till 1900 °C [22]), similar to what is also found for

SrSi2O2-δN2+2/3δ [20].

9.3.2 Luminescence of Eu2+-doped MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba)

The daylight color of the undoped alkaline-earth silicon oxynitrides is grey-white.

Therefore, MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba) shows a high reflection in the visible range

(400 – 650 nm) and a sharp drop between 250 and 300 nm (Fig. 9.2) which corresponds

to the host lattice excitation. Accordingly, the estimated absorption edge of the undoped

materials are around 240 – 280 nm (4.4 – 5.2 eV) (Table 9.3). Fig. 9.2 also shows the

diffuse reflection spectra of the Eu doped compounds. Clearly, these refection spectra

illustrate that the absorption bands of Eu extend into the visible region. The onset of the

absorption bands for the compounds doped with 10% Eu is around 490 nm for M = Ca,

585 nm for M = Sr and 500 nm for M = Ba.

2+

166

Chapter 9

0

20

40

60

80

100

0

20

40

60

80

Ref

lect

ion

(%)

200 300 400 500 600 700

0

20

40

60

80

100

Wavelength (nm)

(a)

(b)

(c)

Fig. 9.2. Diffuse reflection spectra of undoped (dashed line) and 10% Eu-doped (solid

line) MSi2O2-δN2+2/3δ, (a) M = Ca, (b) M = Sr, (c) M = Ba.

Excitation and emission spectra for 10 mol% Eu2+-doped MSi2O2-δN2+2/3δ (M = Ca,

Sr, Ba) are depicted in Fig. 9.3. The excitation spectra of M0.9Eu0.1Si2O2-δN2+2/3δ (M = Ca,

Sr, Ba) are consistent with the corresponding reflection spectra (Fig. 9.2) and show a

number of broad bands corresponding to the crystal-field components of the 5d level in

the excited 4f 5d configuration of the Eu ion (see Table 9.3). From Table 9.3 it can be

clearly seen that the position of the excitation bands is very similar for M = Ca, Sr, Ba,

which suggests that the crystal field splitting and the center of gravity of Eu are not very

much influenced by the different crystal structures, but seems to be fixed by the Si2O2N2

network.

6 2+

2+

The emission spectra of Eu -doped MSi2+2O2-δN2+2/3δ (M = Ca, Sr, Ba) show a typical

broad band emission resulting from the 5d 4f transition of Eu , as shown in Fig. 9.3.

The position of the emission band differs with the type of M ions as generally found in

Eu-doped alkaline-earth silicates and aluminates [35-37]. Excitation into the UV-blue

2+

167

Chapter 9

200 300 400 500 600 700 800

0.0

0.2

0.4

0.6

0.8

1.0

200 300 400 500 600 700 800

0.0

0.2

0.4

0.6

0.8

1.0

200 300 400 500 600 700 800

0.0

0.2

0.4

0.6

0.8

1.0

λem = 560 nm λexc = 400 nm

(a)

(b)

(c)

λem = 570 nm λexc = 400 nm

Wavelength (nm)

λem = 500 nm λexc = 440 nm

Fig. 9.3. Excitation and emission spectra of M0.9Eu0.1Si2O2-δN2+2/3δ: (a) M = Ca, (b) M =

Sr, (c) M = Ba.

range (370 – 450 nm), M0.9Eu0.1Si2O2-δN2+2/3δ (M = Ca, Sr, Ba) yields efficient emission

in the blue-green to yellow spectrum region. BaSi2O2N2: Eu2+ shows a blue-green

emission with a very narrow emission band at about 499 nm (FWHM ~ 35 nm).

CaSi2O2-δN2+2/3δ Eu2+ shows a yellowish emission with a maximum at 560 nm. Similarly,

the emission spectrum of SrSi2O2-δN2+2/3δ:Eu2+ is composed of a broad emission band

ranging from 530 - 570 nm depending on the Eu concentration and the O/N ratio (the

emission band shows a red-shift with decreasing O/N ratio). As compared to the pure

nitride compounds M2Si5N8:Eu (M = Ca, Sr, Ba; λem > 600 nm ) [10] the Eu2+ emission in

MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba; λem < 570 nm ) is at significantly lower wavelengths,

suggesting that Eu is mainly coordinated to oxygen ions in MSi2O2-δN2+2/3δ. In

accordance with this, the crystal structure determination of CaSi2O2N2 points to O atoms

singly bonded to Si atoms, whereas the N atoms are triply bonded [22].

The variation in position of the emission bands, while the excitation bands are at nearly

168

Chapter 9

the same energies, points to strongly different Stokes shifts depending on the type of M

ion. Both CaSi2O2-δN2+2/3δ:Eu2+ and SrSi2O2-δN2+2/3δ:Eu2+ have a significantly larger

Stokes shift than BaSi2O2N2: Eu2+ (Table 9.3). The observation of the smallest Stokes

shift for MSi2O2-δN2+2/3δ:Eu2+ with the largest M ion (i.e. Ba) is consistent with our

previous findings for Eu2+- and Ce3+-doped MYSi4N7 (M = Sr, Ba) [13, 14]. In addition

to long-wavelength emission, this also results in lower quantum efficiency for Eu2+-doped

MSi2O2-δN2+2/3δ (M = Ca, Sr) as compared to BaSi2O2N2: Eu2+. Besides a high quantum

efficiency for UV-blue excitation (> 60%), the smaller Stokes shift of BaSi2O2N2:Eu2+ is

responsible for the narrow emission band and results in improved thermal quenching

behavior. Furthermore, having a somewhat larger crystal field splitting, the lowest energy

excitation band of BaSi2O2N2:Eu2+ at unusual long-wavelength (400 – 450 nm) is

expected.

With an excitation maximum in the range of 430 to 460 nm, Si2O2-δN2+2/3δ:Eu2+ (M =

Ca, Sr, Ba) can be efficiently excited in the blue region of the spectrum, which is very

attractive for application in white-light LEDs. The chromaticity points of

M0.9Eu0.1Si2O2-δN2+2/3δ with different cation M (Ca, Sr, Ba) are shown in the CIE (1931)

chromaticity diagram (Fig. 9.4). For comparison, YAG:Ce3+ and Sr2Si5N8:Eu2+ (excitation

at 460 nm) are also plotted in Fig. 9.4. Similar to YAG:Ce3+, MSi2O2-δN2+2/3δ:Eu2+ (M =

Ca or Sr) in combination with a blue light source (InGaN chip) can generate white-light;

while BaSi2O2N2:Eu2+ (blue-green) together with Sr2Si5N8:Eu2+ (orange-red) in

combination with a blue light source also can give white-light in the RGB

(Red-Green-Blue) model which moreover has a high color rendering index (CRI), an

extensive color range and color stability as compared to the former case [38].

9.3.3 Effect of the Eu2+ concentration on the luminescence of BaSi2O2N2: Eu2+

As usual, with varying amounts of Eu2+ incorporated in the host lattice, the local

surroundings around a substituted site will significantly change (i.e. bond length and

169

Chapter 9

Table 9.3. Excitation and emission bands, crystal field splitting and centre of gravity of the 5d level as well as the Stokes shift of

M0.9Eu0.1Si2O2-δN2+2/3δ and the absorption edge of undoped MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba).

M Excitation band

(nm)

Emission band

(nm)

Absorption edge*

(nm)

Crystal field splitting

(cm-1)

Centre of gravity

(cm-1)

Stokes shift

(cm-1)

Ca 259, 341, 395, 436 560 ~ 280 ~ 15700 ~ 29000 ~ 5100

Sr 260, 341, 387, 440 530 - 570 ~ 270 ~ 15700 ~ 29100 ~ 3900 - 5200

Ba 264, 327, 406, 460 499 ~ 240 ~ 16100 ~ 28700 ~ 1700

* undoped MSi2O2-δN2+2/3δ.

Chapter 9

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.90.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Y

X

Θ

Fig. 9.4. CIE chromaticity coordinates of M0.9Eu0.1Si2O2-δN2+2/3δ, ( , )

Ca0.9Eu0.1Si2O2-δN2+2/3δ; ( , ) Sr0.9Eu0.1Si2O2-δN2+2/3δ; ( , ) Ba0.9Eu0.1Si2O2N2.

Open symbols: λexc = 400 nm; Filled symbols: λexc = 460 nm. ( ) YAG:Ce3+ (λexc =

460 nm); ( ) Sr2Si5N8:Eu2+ (λexc = 460 nm); (Θ) Blue InGaN chip.

angle as well as point symmetry) which eventually makes it possible to tune the

luminescence properties. Similarly, it also can be realized by replacement of Ba by Ca

and/or Sr. As a typical example, Fig. 9.5 shows the relationship between the diffuse

reflection spectra and the Eu2+ concentration of BaSi2O2N2: Eu2+. Obviously, with the

Eu2+ concentration increasing from 1 to 10 mol%, as expected the onset of the Eu2+

absorption band extends at the long-wavelength side from 480 to 500 nm, meanwhile

the absorption intensity is enhanced in the visible range of 400 – 460 nm which

perfectly matches with the emission of the blue-InGaN based LEDs. Correspondingly,

the excitation band also shifts to longer wavelength due to the increased crystal field

splitting and covalency, as shown in an inset in Fig. 9.6 (both crystal field splitting

and center of gravity were derived from the excitation spectra), which results in a

red-shift of the emission band of Eu2+ from 490 to 500 nm (Fig. 9.6). Because the

estimated Stokes shift of Ba1-xEuxSi2O2N2 has not significantly increased from x =

0.01 to 0.1, this effect can be well explained by the replacement of the large Ba2+ ion

by the smaller Eu2+ ion [39] which results in the shrinkage of the BaEu-O/N bond as

we observed in other Eu2+-doped systems [13, 14]. In addition, the emission red-shift

can also be augmented by self absorption at higher Eu2+ concentration.

171

Chapter 9

200 300 400 500 600 700

0

20

40

60

80

100

Ref

lect

ion

(%)

Wavelength (nm)

x = 0.00 x = 0.01 x = 0.05 x = 0.10

Fig. 9.5. Diffuse reflection spectra of Ba1-xEuxSi2O2N2 (x = 0, 0.01, 0.05, 0.1).

400 450 500 550 600 650

0.0

0.2

0.4

0.6

0.8

1.0

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11

15.4

15.6

15.8

16.0

16.2

28.6

28.8

29.0

29.2

29.4

29.6

29.8

30.0

30.2

30.4

Emis

sion

inte

nsity

(a. u

.)

Wavelength (nm)

x = 0.01 x = 0.05 x = 0.10

Cry

stal

fiel

d sp

littin

g (x

103 c

m-1)

Eu2+ concentration

Cen

ter o

f gra

vity

(x10

3 cm

-1)

Fig. 9.6. Emission spectra of Ba1-xEuxSi2O2N2 with varying Eu2+ concentration (λexc =

440 nm). Inset shows the dependence of the crystal field splitting and center of

gravity of the 5d level on the Eu2+ concentration.

9.4. Conclusions

Eu2+-activated MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba) has been synthesized and

characterized using X-ray powder diffraction as well as reflectance, excitation and

emission spectroscopy. A new oxynitride compound BaSi2O2N2 was obtained in the

BaO-SiO2-Si3N4 system. BaSi2O2N2 crystallizes in a monoclinic unit cell with the

lattice parameters a = 14.070(4) Å, b = 7.276(2) Å, c = 13.181(3) Å, β = 107.74(6)°.

172

Chapter 9

For excitation with radiation in the UV-blue range, MSi2O2-δN2+2/3δ:Eu2+ exhibits

efficient blue-green emission at 490 – 500 nm for M = Ba, whereas yellow and

green-yellowish emission at 560 and 530 - 570 nm were found for M = Ca and M = Sr,

respectively. With an intense absorption and excitation band in the UV-blue spectral

region (370 – 460 nm), combined with a high quantum efficiency, MSi2O2N2: Eu2+

can be used as novel conversion phosphors for white-light LEDs.

References:

1. Nakamura S. Appl. Phys. Lett. 1994, 64, 1687.

2. Aanegola S.; Petroski J.; Radkov E. SPIE. 2003, 10, 16.

3. Narukawa Y. Optics & Photonics News. 2004, 4, 25.

4. Rohwer L.S.; Srivastava A.M. The Electrochemical Society Interface. 2003, 12, 36.

5. Taso J.Y., Ed., in Light Emitting Diodes (LEDs) for General Illumination Update

2002, Optoelectronics Industry Development Association: Washington, DC, 2002.

6. Hintzen H.T.; Li Y.Q. Encyclopedia of Materials: Science and Technology

2004, 1.

7. van Krevel J.W.H. Ph.D. Thesis, Eindhoven University of Technology, 2000.

8. van Krevel J.W.H.; Hintzen H.T.; Metselaar R.; Meijerink A. J. Alloys

Comp. 1998, 268, 272

9. van Krevel J.W.H.; Hintzen H.T.; Metselaar R. Mater. Res. Bull. 2000, 35, 747.

10. Hintzen H.T.; van Krevel J.W.H.; Botty G. European Patent EP1104799 A1, 1999.

11. Höppe H.A.; Morys H. Lutz, P.; Schnick W.; Seilmeier A. J. Phys. Chem.

Solids. 2000, 61, 2001

12. van Krevel J.W.H.; van Rutten J.W.T.; Mandal H.; Hintzen H.T.; Metselaar R.

J. Solid State Chem. 2002, 165, 19.

13. Li Y.Q.; De.With G.; Hintzen H.T. J. Alloys Com. 2004, 385, 1.

14. Li Y.Q.; Fang C.M.; De.With G.; Hintzen H.T. J. Solid State Chem. 2004, 177,

4687.

15. Yamada M.; Naitou T.; Izuno K.; Tamaki H.; Murazaki Y.; Kameshima M.;

Mukai T. Jpn. J. Appl. Phys. 2003, 42, L20.

16. Xie R.J.; Hirosaki N.; Sakuma K.; Yamamoto Y. ; Mitomo M. Appl. Phys.

Lett. 2004, 84, 5404.

173

Chapter 9

17. Tamaki H.; Kameshima M; Takashima S.; Yamada M.; Naitou T.; Sakai K.;

Murazaki Y.; European Patent EP 1433831 A1, 2003.

18. Juestel T.; Schmidt T.; Hoeppe H.; Schnick W.; Mayr W. PCT WO 2004/055910

A1.

19. Huang Z.K.; Sun W.Y.; Yan D.S. J. Mater. Sci. Lett. 1985, 4, 255.

20. Zhu W.H.; Wang P.L.; Sun W.Y.; Yan D.S. J. Mater. Sci. Lett. 1994, 13, 560.

21. Cao G. Z.; Huang Z.K.; Fu X.R.; Yan D.S. Int. J. High Tech. Ceram. 1986,

2, 115.

22. Höppe H.A.; Stadler Florian; Oeckler Oliver; Schnick Wolfgang.

Angew. Chem. Int. Ed. 2004, 43, 5540.

23. Poort S.H.M.; Janssen W.; Blasse G.. J. Alloys Comp. 1997, 260, 93.

24. Poort S.H.M.; Reijnhoudt H.M.; van der Kulp H.O.T.; Blasse G. J. Alloys Comp.

1996, 241, 75.

25. Poort S.H.M.; Meyerink A.; Blasse G.. J. Phys. Chem. Solids. 1997, 58, 1451.

26. Kim J.S.; Kang J.Y.; Jeon P.E.; Choi J.C.; Park H.L.; Kim T.W. Jpn. J. Appl.

Phys. 2004, 43, 989.

27. Kim J.S.; Jeon P.E.; Choi J.C.; Park H.L. Solid State Comm. 2005, 133, 187.

28. Kim J.S.; Park Y.H.; Kim S.M.; Choi J.C.; Park H.L. Solid State Comm. 2005,

133, 445.

29. Boultif A; Louer D., J. Appl. Cryst., 2004, 37, 724.

30. Le Bail A., http://www.cristal.org/McMaille/

31. Le Bail A ; Duroy H. and Fourquet J.L., Mater. Res. Bull., 1988, 23, 447

32. Larson A.C. and Von Dreele R.B., Report LAUR 86-748, Los Alamos National

Laboratory, Los Alamos, NM, 2000.

33. Toby B.H., J. Appl. Cryst. 2001, 34, 210.

34. Li Y.Q.; De With G.; Hintzen H.T. to be published.

35. Blasse G.; Grabmaier B.C. Luminescent materials, Springer-Verlag, Berlin:

1994.

36. Blasse G.; Wanmaker W.L.; Ter Vrugt J.W.; Bril A. Philps. Res. Repts. 1967, 23,

189.

37. Blasse G.; Bril A. Philips. Res. Repts. 1968, 23, 201.

38. Hintzen H.T.; Li Y.Q.; Delsing A.C.A., PCT WO 2004/029177 A1.

39. Shannon R.D. Acta Cryst. 1976, A32, 751.

174

Chapter 10

Luminescence of a new class UV- blue- emitting phosphors

MSi2O2-δN2+2/3δ:Ce3+ (M = Ca, Sr, Ba)

ABSTRACT

The luminescence properties of Ce3+, Na+-codoped MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba) are

reported. The undoped and Ce3+, Na+-codoped MSi2O2-δN2+2/3δ powders were prepared by

a solid-state reaction at temperatures between 1300 - 1400 oC under N2-H2 (10%)

atmosphere in the system MO-SiO2-Si3N4 (M = Ca, Sr, Ba). MSi2O2-δN2+2/3δ (M = Ca, Sr,

Ba) crystallizes in the monoclinic system with different crystal structures. For excitation

in the range of 300 - 360 nm, MSi2O2-δN2+2/3δ:Ce3+ shows typical broad emission bands

peaking at about 392, 473 and 396 nm for M = Ca, Sr and Ba, respectively. In particular,

CaSi2O2N2:Ce3+ shows an unusual short-wavelength emission (~ 392 nm) with a very

small Stokes shift of 2200 cm-1; BaSi2O2N2:Ce3+ shows an interesting white-light

emission in the visible range of 350 – 600 nm for excitation at 365 nm.

Keywords: alkaline-earth-silicon-oxynitride, cerium, sodium, X-ray powder diffraction,

indexing, optical properties, luminescence properties, phosphor.

175

Chapter 10

10.1. Introduction

The Ce3+ ion has a 4f1 electronic ground state configuration. The luminescence of the

Ce3+ ion originates from a transition from the lowest 5d level to the ground states which

is split by the spin-orbit coupling into two components, 2F5/2 and 2F7/2, separated by

~2000 cm-1 [1]. Since the position of the lowest 5d levels is strongly influenced by the

local coordination, the emission wavelengths of Ce3+ varies with different host lattices

from UV to the visible range corresponding to emission colors from blue to red [1]. In

oxide host lattices, the emission of Ce3+ generally is located in the UV to blue (300 – 500

nm) spectral range [1]. Exception is the yellow-emitting YAG:Ce3+ due to large crystal

field splitting [1, 2]. A large crystal field splitting can also be realized by N3- replacement

of O2-. In addition, nitride-based host lattices provide more covalent bonding (like in

sulfides) resulting in the 5d band shift to lower energy [1, 3-5]. Indeed, long-wavelength

emission in Ce3+-doped rare-earth-(oxy)nitride and alkaline-earth silicon nitride materials

is observed [4, 6].

In comparison with oxides, nitride and oxynitride-based materials can give some

surprises not only in structure (like an unusual motif) but also in physical characteristics

which are reflected by their unique mechanical, electrical, thermal and optical properties

[7-13]. Definitely, the nitrogen atom is believed to play a key role due to its high formal

charge and large covalent character in nitride-based materials [7-9].

In the system M-Si-O-N (M = Ca, Sr, Ba), alkaline-earth silicon oxynitride compounds

with composition MSi2O2N2 (M = Ca, Sr, Ba) are known [14-18]. This kind of

oxynitrides of interest for luminescent materials is that its composition is situated

between the oxide compound M2SiO4 and the pure nitride compound M2Si5N8 (M = Ca,

Sr, Ba). Eu2+-doped M2SiO4 phosphor materials are well-known green (M = Ca, Ba) and

yellow (M = Sr) emitting phosphors [19-22], while, M2Si5N8:Eu2+ (M = Ca, Sr, Ba) is a

new family of red-emitting phosphors showing excellent luminescence properties for

white-light LED applications [23, 24]. Recently, we have reported on Eu2+-doped

MSi2O2-δN2+2/3δ phosphor materials with yellow (M = Ca), green to yellow (M = Sr) and

blue-green (M = Ba) emission colors [17], which are also promising candidates for use as

conversion phosphors for white-light LED applications [25]. In contrast, the

176

Chapter 10

luminescence properties of Ce3+-activated alkaline-earth silicates in the BaO-SrO-SiO2

system have been reported in an earlier work which revealed that M2SiO4:Ce3+ and

MSiO3:Ce3+ (M = Sr, Ba) exhibited a peak emission wavelength at about 390 nm with

slight variations resulting from compositional changes [26]. Most recently, we have

reported the luminescence properties of Ce3+-activated M2Si5N8 (M = Ca, Sr, Ba) using

Li or Na as a charge compensator [6]. Especially, Sr2Si5N8:Ce3+ turns out to be a very

attractive green-emitting phosphor for use in white-light LEDs owing to its high

conversion efficiency in the UV blue range (370 – 450 nm). These peculiar behaviors

inspired us to extend our study to the Ce3+-doped MSi2O2-δN2+2/3δ system (M = Ca, Sr,

Ba). In this study, undoped and Ce3+-doped MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba) compounds

were synthesized by a solid-state reaction using Na+ as charge-compensator. Furthermore,

new X-ray powder diffraction data and the lattice parameters of MSi2O2-δN2+2/3δ (M = Ca,

Sr) are presented as we have found that previous studies on these compounds are

imprecise [15, 16]. Finally, the unconventional luminescence properties of MSi2O2N2:

Ce3+ (M = Ca, Sr, Ba) are reported.

10.2. Experimental 10.2.1. Synthesis

All powder samples of undoped and Ce3+, Na+-codoped MSi2O2-δN2+2/3δ (M = Ca, Sr,

Ba) were prepared by a solid-state reaction at high temperatures using Na+ as a charge

compensator. As found in our previous study [17], the approximate δ value to obtain

single-phase compounds is about 1 for M = Sr and close to 0 for M = Ca and Ba.

The starting materials were high-purity MCO3 (M = Ca, Sr, Ba) (Merck, > 99.0%),

SiO2 (Aerosil OX 50, Degussa), Si3N4 (SKW Trostberg, β content: 23.3%, O ~ 0.7%),

CeO2 (Rhône-Poulenc, 99.95%) and NaF (Merck, > 99.0%). The Ce concentrations in the

MSi2O2-δN2+2/3δ host lattices are 1 mol% for M = Ca and Sr and 1-3 mol% for M = Ba

with respective to the M ions (i.e. x = 0.01 or 0.03 in M1-2xCexNaxSi2O2-δN2+2/3δ).

Appropriate amounts of the starting materials were homogeneously wet-mixed using a

planetary ball mill for 4 -5 hours in isopropanol. After mixing the slurry was dried and

ground in an agate mortar. Subsequently, the dried powder mixtures were fired in

177

Chapter 10

molybdenum or alumina crucibles at 1300-1400 °C for 6 – 12 h under a reducing

atmosphere of N2 - H2 (10%) in a horizontal tube furnace. After firing, the samples were

cooled down to room temperature in the furnace and were ground again with an agate

mortar for further measurements.

10.2.2. X-ray powder diffraction

All obtained samples were checked by X-ray powder diffraction (Rigaku, D/MAX-B)

using Cu-Kα radiation at 40 kV and 30 mA with a graphite monochromator. The data

were collected on powder samples using a step scan mode with a step size of 0.02˚ and a

counting time of 10 second per step in the range 2θ 10 to 90˚. In order to avoid the

preferential particle orientation of the obtained samples, the powder samples were

mounted into a flat plate holder by the side filling method.

For M = Ca, Sr, the crystal system of MSi2O2-δN2+2/3δ was determined from the X-ray

powder patterns by the indexing program DICVOL04 [27] using the first 20 lines. The

possible space groups are determined according to the systematic absences.

10.2.3. Optical measurements

The diffuse reflection, excitation and emission spectra were measured at room

temperature by a Perkin Elmer LS 50B spectrophotometer equipped with a Xe flash lamp.

The reflection spectra were calibrated with the reflection of black felt (reflection 3%) and

white barium sulfate (BaSO4, reflection ~100%) in the wavelength region of 230-700 nm.

The excitation and emission slits were set at 2.5 nm. Excitation spectra were

automatically corrected for the variation in the lamp intensity by a second photomultiplier

and a beam-splitter. The emission spectra were corrected by dividing the measured

emission intensity by the ratio of the observed spectrum of a calibrated W-lamp and its

known spectrum from 300 to 900 nm. All the spectra were measured with a scan speed of

100 nm/min. The optical absorption edge is estimated by the wavelength value at which

the reflection intensity is half of the lowest and highest of the overall reflection intensity.

178

Chapter 10

10.3. Results and discussion

10.3.1 X-ray powder diffraction data of CaSi2O2N2 and SrSi2ON8/3

The X-ray powder diffraction data of CaSi2O2N2 and SrSi2ON8/3 compounds are given

in Table 10.1. These two compounds were successfully indexed on the monoclinic

systems with the lattice parameters: CaSi2O2N2, a = 15.036(4) Å, b = 15.450(1) Å, c =

6.851(2) Å, β = 95.26(2)°; SrSi2ON8/3, a = 11.320(4) Å, b = 14.107(6) Å, c = 7.736(4) Å,

β = 91.87(3)°. The calculated XRD data based on the crystal structure reported for

CaSi2O2N2 [18] do not fit as well our measured data in this work. XRD full-pattern

simulations support our proposed cell for polycrystalline CaSi2O2N2 [17]. This difference

is probably ascribed to the exact composition of the obtained compounds (i.e. O/N ratio

in CaSi2O2N2) arising from the different starting materials and synthetic approaches [18],

possibly resulting in two modifications as found in the case of SrSi2O2N2 [15]. A newly

found compound BaSi2O2N2 [17] also crystallizes in a monoclinic cell having different

structure with CaSi2O2N2 and SrSi2ON8/3, the obtained lattice parameters of MSi2O2-

δN2+2/3δ (M = Ca, Sr, Ba) are listed in Table 10.2.

10.3.2 Optical properties

10.3.2.1 Diffuse reflection

The observed day-light color is grey-white for undoped MSi2O2-δN2+2/3δ (M = Ca, Sr,

Ba) in agreement with the measured diffuse reflection spectra which show that only light

in the UV range (i.e. < 300 nm) is absorbed (Fig. 10.1). From the diffuse reflection

spectra the optical absorption edge of MSi2O2-δN2+2/3δ is estimated to be about 270 nm

(4.6 eV), 273 nm (4.55 eV) and 242 nm (5.13 eV) for M = Ca, Sr, Ba, respectively. The

drop in the reflection curve represents the host lattice absorption from the valence to

conduction band. BaSi2O2N2 show a much steeper drop starting from 275 nm (Fig. 10.1).

In the UV-blue to visible range, the reflection of CaSi2O2N2 and BaSi2O2N2 is higher (>

80%) than that of SrSi2ON8/3 (< 60%).

For all Ce3+-doped MSi2O2-δN2+2/3δ materials, only one obvious absorption band

centered at about 336, 355 and 308 nm for M = Ca, Sr, Ba, respectively, can be seen. In

179

Chapter 10

Table 10.1. X-ray powder diffraction data for MSi2O2-δN2+2/3δ (M = Ca, Sr) (a) M = Ca (δ = 0) h k l dobs (Å) 2θobs (deg.) 2θcal (deg.) ∆2θ (deg.) I/Io (%) 0 0 1 1 1 -1 2 1 -1 1 2 1 0 3 1 1 4 0 0 0 2 0 1 2 1 0 2 1 1 2 1 2 -2 1 2 2 4 2 1 3 1 -2 0 3 2 1 3 2 2 5 -1 1 4 -2 0 6 1 5 0 -2 0 5 2 0 1 3 1 0 3 2 5 2 2 7 0 0 7 1 0 3 3 1 4 -3 5 6 0 5 6 -1 5 1 -3 0 5 3 2 7 -2 3 8 0 7 3 -2 5 0 3 0 0 4 5 2 3 2 9 0 0 9 1 3 6 -3

6.87726 5.94881 4.99010 4.75165 4.11482 3.75130 3.42184 3.33621 3.26388 3.19291 3.11022 3.00756 2.92864 2.89511 2.84840 2.75605 2.66657 2.54901 2.41504 2.36117 2.29749 2.25539 2.22339 2.15431 2.12057 2.10172 2.08236 1.96743 1.94905 1.90811 1.88136 1.83713 1.82064 1.80177 1.77306 1.73674 1.70788 1.69325 1.67776 1.66546 1.64999

12.862 14.880 17.760 18.659 21.579 23.699 26.019 26.699 27.302 27.921 28.679 29.680 30.499 30.861 31.380 32.460 33.581 35.179 37.200 38.081 39.179 39.941 40.541 41.901 42.600 43.001 43.421 46.099 46.559 47.619 48.339 49.580 50.060 50.621 51.500 52.659 53.619 54.120 54.661 55.099 55.660

12.934 14.881 17.710 18.671 21.564 23.741 26.045 26.685 27.263 27.876 28.656 29.648 30.542 30.878 31.370 32.404 33.529 35.114 37.238 38.138 39.242 39.958 40.551 41.860 42.627 42.974 43.378 46.122 46.487 47.650 48.309 49.647 50.013 50.596 51.491 52.638 53.581 54.055 54.568 55.042 55.665

-0.072 -0.001 0.050 -0.012 0.015 -0.042 -0.026 0.014 0.039 0.045 0.023 0.032 -0.043 -0.017 0.010 0.056 0.052 0.065 -0.038 -0.057 -0.063 -0.017 -0.010 0.041 -0.027 0.027 0.043 -0.023 0.072 -0.031 0.030 -0.067 0.047 0.025 0.009 0.021 0.038 0.065 0.093 0.057 -0.005

33.8 3.2 4.5 7.8 16.6 8.6 89.0 64.9 8.3 17.6 15.8 21.7 21.3 10.6 15.9 9.8 91.3 62.1 15.3 100.0 24.1 20.9 10.9 5.9 5.4 6.3 5.0 5.4 8.8 18.4 15.6 10.5 6.2 15.5 2.5 6.9 13.7 2.6 10.6 0.9 3.6

180

Chapter 10

1 2 4 4 7 2 2 9 1 4 1 -4 4 6 -3 2 3 4 4 6 3 1 10 -1 0 5 4 5 3 -4 2 10 1 2 5 4 7 3 3 2 10 -2 2 6 4 0 11 1 0 0 5 1 7 -4 0 2 5 5 4 4 2 1 5 1 3 5 0 11 2 8 7 2 2 3 5 2 10 -3 1 8 4 1 12 1 0 7 -1 3 8 4 0 11 3 4 10 3 1 13 0 0 0 -4 2 13 -1 5 10 3 1 10 -4 1 10 4 1 8 -5 2 0 6 1 3 6

1.64078 1.63115 1.61955 1.60101 1.59492 1.55854 1.51279 1.50480 1.49603 1.47934 1.47177 1.44473 1.43247 1.39418 1.38147 1.37647 1.36586 1.35414 1.34564 1.33999 1.32004 1.30579 1.30117 1.28592 1.28205 1.27202 1.26853 1.26106 1.23414 1.22061 1.19663 1.19218 1.18657 1.17862 1.16090 1.15454 1.14831 1.13780 1.11924 1.11130 1.10174

56.000 56.360 56.800 57.519 57.759 59.240 61.220 61.580 61.981 62.759 63.119 64.441 65.060 67.079 67.779 68.059 68.661 69.340 69.841 70.179 71.400 72.301 72.599 73.600 73.859 74.540 74.780 75.300 77.241 78.259 80.140 80.500 80.960 81.621 83.140 83.701 84.259 85.220 86.981 87.760 88.720

55.938 56.394 56.806 57.499 57.784 59.276 61.150 61.606 61.985 62.730 63.048 64.434 65.047 67.065 67.852 68.001 68.587 69.281 69.803 70.174 71.472 72.267 72.639 73.667 73.854 74.520 74.780 75.339 77.251 78.282 80.131 80.496 80.989 81.609 83.139 83.633 84.212 85.183 86.968 87.760 88.678

0.062 -0.034 -0.006 0.020 -0.025 -0.036 0.070 -0.026 -0.004 0.029 0.071 0.007 0.013 0.014 -0.073 0.058 0.074 0.059 0.038 0.005 -0.072 0.034 -0.040 -0.067 0.005 0.020 0.000 -0.039 -0.010 -0.023 0.009 0.004 -0.029 0.012 0.001 0.068 0.047 0.037 0.013 0.000 0.042

17.2 6.2 2.4 3.7 4.1 29.0 11.2 8.4 2.8 5.6 3.6 9.4 2.2 3.2 5.6 1.6 10.4 3.6 5.2 1.8 3.0 2.6 2.1 14.5 5.4 6.2 4.4 2.7 2.4 1.7 2.3 2.9 2.1 3.4 5.4 1.8 1.8 1.7 2.3 3.4 3.8

181

Chapter 10

(b) M = Sr (δ = 1) h k l Dobs (Å) 2θobs (deg.) 2θcal (deg.) ∆2θ (deg.) I/Io (%) 0 2 0 2 0 0 2 1 -1 0 0 2 0 4 0 3 0 -1 3 1 -1 1 2 2 2 1 -2 1 4 -1 4 0 0 3 3 1 2 3 2 1 0 -3 3 4 1 2 5 1 0 6 0 3 3 2 1 3 3 3 5 1 5 2 1 1 6 2 0 0 4 1 0 4 2 7 1 2 0 4 1 7 -2 1 3 4 6 3 0 0 8 1 2 4 4 0 8 2 1 5 -4 0 0 5 2 9 1 3 1 5 1 5 -5 1 10 2

7.07617 5.66933 4.42284 3.88052 3.52827 3.44003 3.34100 3.22702 3.16228 3.10376 2.83073 2.73149 2.61518 2.52959 2.43023 2.39278 2.34935 2.31446 2.20578 2.15633 2.06163 1.97151 1.93412 1.89683 1.83852 1.81460 1.76989 1.75718 1.74967 1.71678 1.61326 1.60501 1.58386 1.54854 1.47845 1.40790 1.34971 1.31242

12.499 15.618 20.060 22.899 25.221 25.879 26.660 27.620 28.197 28.740 31.581 32.760 34.261 35.458 36.959 37.559 38.280 38.880 40.879 41.860 43.880 45.998 46.940 47.920 49.540 50.238 51.599 52.000 52.240 53.319 57.042 57.362 58.201 59.661 62.801 66.340 69.600 71.879

12.488 15.596 20.093 22.927 25.205 25.890 26.662 27.607 28.189 28.771 31.570 32.781 34.234 35.398 36.972 37.590 38.248 38.862 40.867 41.828 43.906 45.976 46.927 47.917 49.537 50.288 51.635 51.946 52.220 53.265 57.094 57.415 58.156 59.721 62.768 66.311 69.532 71.807

0.011 0.022 -0.033 -0.028 0.016 -0.011 -0.002 0.013 0.008 -0.031 0.011 -0.021 0.027 0.060 -0.013 -0.031 0.032 0.018 0.012 0.032 -0.026 0.022 0.013 0.003 0.003 -0.050 -0.036 0.054 0.020 0.054 -0.052 -0.053 0.045 -0.060 0.033 0.029 0.068 0.072

18.8 1.2 2.2 2.0 100.0 9.4 5.8 8.0 2.1 4.8 45.5 2.1 7.4 12.5 4.7 12.6 2.4 6.9 6.3 2.4 3.7 4.1 5.2 1.7 6.6 6.0 6.3 3.7 6.0 3.8 1.3 2.4 11.4 5.6 3.7 4.3 1.2 4.3

182

Chapter 10

Table 10.2. Lattice parameters of MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba).

Formula Crystal system Space group Lattice constants

a (Å) b (Å) c (Å) β (°) V (Å3)

Figure-of-Merit M(20)

F(20) Reference

CaSi2O2N2 (δ = 0) Monoclinic P21/c 15.035(4) 15.450(1) 6.851(2) 95.26(3) 1584.53 10.5 15.7(0.0088, 144) This work

SrSi2ON8/3 (δ ≈ 1) Monoclinic P21/m 11.320(4) 14.107(6) 7.736(4) 91.87(3) 1234.67 10.8 14.9(0.0090, 150) This work

BaSi2O2N2 (δ = 0) Monoclinic P2/m 14.070(4) 7.276(2) 13.181(3) 107.74(6) 1285.23 10.3 15.4( 0.0095, 137) [17]

addition, the absorption intensity of Ce3+ in CaSi2O2N2 and BaSi2O2N2 is stronger than

that in nitrogen-richer SrSi2ON8/3, possibly related to the amount of Ce incorporated.

Actually, more often we have found lower Ce3+ absorption in a nitrogen-richer

environment. Typical examples are M2Si5N8:Ce3+, Li+(Na+) and BaSi7N10:Ce3+, Li+(Na+),

where both the absorption and the luminescence intensity of the oxygen-poor materials

are lower than that of slightly oxygen-richer materials (i.e. the former using β-Si3N4

instead of α-Si3N4 as a starting material which is well known to contain more oxygen)

[28]. Similar observations were done for Eu2+-doped M2Si5N8 (M = Ca, Sr, Ba) and

BaSi7N10 materials [28].

10.3.2.2 Luminescence of MSi2O2-δN2+2/3δ:Ce3+

Fig. 10.2 shows the excitation and emission spectra of M0.98Ce0.01Na0.01Si2O2-δN2+2/3δ

(M = Ca, Sr, Ba). For all materials, the excitation band at short wavelength in the range

of 230 – 250 nm can be readily assigned to host lattice excitation as indicated by their

reflection spectra (see Fig. 10.1). Surprisingly, only one excitation band obvious for Ce3+

can be observed peaking at about 336, 366 and 308 nm for M = Ca, Sr and Ba,

respectively, in fair agreement with the obtained diffuse reflection spectra (Table 10.3).

183

Chapter 10

0

20

40

60

80

100

0

20

40

60

80

200 300 400 500 600 7000

20

40

60

80

100

CaSi2O2N2

(a)

(b)

SrSi2ON8/3

Ref

lect

ion

(%)

(c)

BaSi2O2N2

Wavelength (nm)

Fig. 10.1. Diffuse reflection spectra of undoped (solid line) and 1 mol% Ce3+, Na+-

codoped (dashed line) MSi2O2-δN2+2/3δ: (a) M = Ca, (b) M = Sr, (c) M = Ba.

In the case of SrSi2ON8/3:Ce3+, a weak shoulder at long wavelength around 410 nm can

be observed, probably originating from a second phase similar to what we found in

SrSi2ON8/3:Eu2+ [17].

In general, the 5d levels of Ce3+ can be split into at most five different crystal-field

components [1]. The above observation suggests that the excitation bands of Ce3+ in

MSi2O2-δN2+2/3δ are seriously overlapping (although bandwidth is small) and/or that some

of them may be located in the conduction band of the host lattice. M0.98Ce0.01Na0.01Si2O2-

δN2+2/3δ shows a typical broad emission band with maxima at about 392, 473 and 396 nm

for M = Ca, Sr and Ba, respectively (Fig. 10.2), located in the UV-blue spectral range

184

Chapter 10

200 250 300 350 400 450 500 550 600

0.0

0.2

0.4

0.6

0.8

1.0

Inte

nsity

(a.u

.)

Wavelength (nm)

λem= 380 nm λexc= 320 nm λexc= 337 nm

(a)

CaSi2O2N2

200 250 300 350 400 450 500 550 600 650 700

0.0

0.2

0.4

0.6

0.8

1.0

SrSi2ON8/3

λem= 470 nm λexc= 352 nm λexc= 364 nm

Inte

nsity

(a.u

.)

Wavelength (nm)

(b)

200 250 300 350 400 450 500 550 600 650

0.0

0.2

0.4

0.6

0.8

1.0

BaSi2O2N2

Inte

nsity

(a.u

.)

Wavelength (nm)

λem = 400 nm λexc = 308 nm λexc = 329 nm

(c)

Fig. 10.2. Excitation (solid line) and emission (dashed line) spectra of

M0.98Ce0.01Na0.01Si2O2-δN2+2/3δ: (a) M = Ca, (b) M = Sr, (c) M = Ba.

185

Chapter 10

Table 10.3. Optical properties of MSi2O2-δN2+2/3δ:Ce3+, Na+ (1 mol%) (M = Ca, Sr, Ba)

M δ

Absorption band (nm)

5d excitation band (nm)

Emission band (nm)

CIE coordinate (x, y)

Stokes shift (cm-1)

Ca 0 336 336 392 (0.165, 0.061) ~ 2200

Sr 1 355 366 473 (0.197, 0.263) ~ 6500

Ba 0 308 308 396 (0.161, 0.096) ~ 5000

(Table 10.3). The emission colour points are shown in Fig. 10.3. In the case of

CaSi2O2N2:Ce3+ and BaSi2O2N2:Ce3+, the broad emission band actually contains three

subbands (Fig. 10.2a and Fig. 10.2c) at about 368, 378 and 392 nm for the Ca-; and 369,

380 and 396 nm for the Ba-compound. Because the energy difference between these

subbands significantly deviates from the normal value between the ground states of 2F5/2

and 2F7/2 (~2000 cm-1) [1], these subbands suggest the presence of multi-emission centers

of the Ce3+ ions. This hypothesis is confirmed by the fact that the shape (for M = Ca)

and position (for M = Ba) of the emission subbands can be changed by varying the

excitation wavelength (Fig. 10.2). In agreement with this observation, for CaSi2O2N2 the

large number of six crystallographic Ca sites in a unit cell was reported [18], which could

also apply to BaSi2O2N2.

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.90.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Y

X

Fig. 10.3. Color coordinates deduced from the emission band of M0.98Ce0.01Na0.01Si2O2-

δN2+2/3δ, ( ) M = Ca, ( ) M = Sr, ( ) M = Ba (λexc = 337, 364, 308 nm for M = Ca, Sr

and Ba, respectively).

186

Chapter 10

As mentioned before, from the compositional point of view, MSi2O2-δN2+2/3δ

apparently lies between alkaline-earth silicates and alkaline-earth silicon nitrides.

Normally, a lower energy 5d Ce3+ excitation band together with a longer wavelength

emission band is expected due to a highly covalent bonding and a large crystal splitting in

the nitride or oxynitride compounds [4]. Therefore, the local coordination around M ions

in MSi2O2-δN2+2/3δ can be probed by the luminescent ions, such as Ce3+ and Eu2+. With

respect to the luminescence properties, both the excitation and emission spectra of

MSi2O2-δN2+2/3δ:Ce3+ are more alike to that of Ce3+-doped alkaline-earth silicates [26]

rather than that of M2Si5N8:Ce3+ [6]. For example, for MSiO3:Ce3+ (M = Sr, Ba) and

M2SiO4:Ce3+ (M = Sr, Ba), the main excitation band is around 300 - 335 nm and the

emission band is around 390 nm [26], while for M2Si5N8:Ce3+ the principle excitation

band is around 400 nm and emission band is found at about 470 – 560 nm depending on

the type of M [6]. For Eu2+-doped MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba) [17], the

luminescence properties are also close to those of Eu2+-doped alkaline-earth silicates [19-

22] while significantly different from M2Si5N8:Eu2+ (M = Ca, Sr, Ba) [23, 24]. Therefore

it can be concluded that the M ions in MSi2O2-δN2+2/3δ are dominantly coordinated by O

atoms, in agreement with the structure elucidation of CaSi2O2N2, which is a layer silicon

oxynitride in which Ca2+ ions are connected by six O atoms and one N atom in the range

of 2.28 – 2.79 Å [18]. In addition, based on the fact that these alkaline-earth silicates

consist of layer (i.e. SrSiO3), chain (i.e. BaSiO3) and isolated [SiO4] tetrahedra [29],

MSi2O2-δN2+2/3δ (M = Sr, Ba) is possibly also composed of layers of [Si(O,N)4] tetrahedral

groups similar to the reported CaSi2O2N2 structure [18].

The estimated Stokes shifts are about 2200, 6500 and 5000 cm-1 for M = Ca, Sr and Ba,

respectively (Table 10.3). These results are completely contrary to what we have found

for Eu-doped MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba), where BaSi2O2N2:Eu2+ has the smallest

Stokes shift ~ 1700 cm-1 while CaSi2O2N2:Eu2+ has a significantly larger Stokes shift (~

5100 cm-1) [17]. Exactly similar opposing trends for Ce3+ and Eu2+ were also found for

M2SiO4, i.e., 2900 (M = Sr) vs. 4800 cm-1 (M = Ba) for Ce3+ [26]; while 5500-6000 (M =

Sr) vs. 5000 cm-1 (M = Ba) for Eu2+ [19-21]. In general, with the ionic radius of M

increasing going from Ca, Sr to Ba the Stokes shift is expected to decrease for

isostructural compounds as we indeed have found in MYSi4N7:Ce3+ (M = Sr, Ba) [30, 31]

187

Chapter 10

with a homovalent Ce3+/Y3+substitution and M2Si5N8:Ce3+,Li+(Na+) (M = Sr, Ba) [6] with

heterovalent Ce3+/M2+substitution. The reason for the deviation of the Ca > Sr > Ba

sequence evidently is the fact that the MSi2O2-δN2+2/3δ compounds have different crystal

structures. In the case of different behaviors of the Ce3+ and Eu2+ ions, this is possibly

related to their different site preferences as influenced by charge compensation necessary

for Ce3+ in contrast to Eu2+ in MSi2O2-δN2+2/3δ, eventually resulting in significantly

different trends in luminescence properties.

Finally, it is worth noting that CaSi2O2N2:Ce3+ is a high potential UV-blue-emitting

phosphor material with high efficiency and low thermal quenching (i.e. a small Stokes

shift). With respective to Ce3+-doped BaSi2O2N2, first, when the Ce concentration

increases from 1 to 3 mol% both the excitation and emission bands show significant shift

to long-wavelength (Fig. 10.4).

200 250 300 350 400 450 500 550 600 650 7000.0

0.2

0.4

0.6

0.8

1.0

0

20

40

60

80

100

λem = 435 nm λexc = 330 nm

Rel

fect

ion

(%)

λ exc= 330 nmλ em= 450 nm

Inte

nsity

(a.u

.)

Wavelength (nm)

Fig. 10.4. Diffuse reflection, excitation and emission spectra of Ba0.94Ce0.03Na0.03Si2O2N2.

As no significant change in the Stokes shift (< 200 cm-1), this red-shift is mainly

attributed to the BaSi2O2N2 lattice shrinkage caused by the replacement of the large Ba2+

ion (1.35 Å, CN = 6) by the smaller Ce3+ (1.01 Å, CN = 6) and Na+ (1.02 Å, CN = 6) ions

[32]. Correspondingly, the BaCe-O(N) distances become shorter which leads to increase

in the crystal-field splitting. As a consequence, the lowest 5d level shifts to lower energy.

Second, an attractive feature of BaSi2O2N2:Ce3+ is that it shows white light for excitation

188

Chapter 10

under 365 nm, especially for high Ce concentrations. As far as we know, no such studies

have been reported. Just using a single activator within a host lattice to generate white-

light is rather unique, but it is indeed realized in BaSi2O2N2:Ce3+.

10.4. Conclusions

A new class of UV-blue-emitting phosphor materials MSi2O2-δN2+2/3δ:Ce3+ (M = Ca, Sr,

Ba) has been found. X-ray powder diffraction analysis showed that MSi2O2-δN2+2/3δ

crystallized in the monoclinic system with different crystal structures. Ce3+-doped

MSi2O2-δN2+2/3δ shows UV-blue emission with maxima at about 392, 473 and 396 nm for

M = Ca, Sr and Ba, respectively, under excitation in the UV range (300 – 360 nm).

Unexpectedly, CaSi2O2N2:Ce3+ emits light at very high energies for nitride-based

materials, ascribed to predominantly coordination with oxygen atoms combined with a

small Stokes shift due to a rigid lattice. For BaSi2O2N2:Ce3+, with increasing Ce

concentration both excitation and emission bands show a red-shift and itself can emit

white light when excited by 365 nm UV-light.

189

Chapter 10

References: 1. G. Blasse and B.C. Grabmaier, Luminescent materials, Springer-Verlag, Berlin, 1994.

2. G. Blasse and A. Bril, Appl. Phys. Lett., 1967, 11, 53.

3. B. Huttl, U. Troppenz, K.O. Velthaus, C.R. Ronda, R.H. Mauch, J. Appl. Phys., 1995,

78, 7282.

4. J.W.H. van Krevel, H.T. Hintzen, R. Metselaar and A. Meijerink, J. Alloys Comp.,

1998, 268, 272.

5. J.W.H. van Krevel, H.T. Hintzen, R. Metselaar, Mater. Res. Bull., 2000, 35, 747.

6. Y.Q. Li, G. de With and H.T. Hintzen, J. Lumin., in press.

7. D.H. Gregory, J. Chem. Soc., Dalton Trans, 1999, 259-270.

8. W. Schnick, Angew. Chem. Int. Ed., 1993, 32, 806.

9. R Marchand, F Tessier, A Le Sauze, N Diot, Inter. J. Inorg. Mater., 2001, 3, 1143.

10. R. Niewa, F.J. DiSalvo, Chem Mater., 1998, 10, 2733.

11. W. Schnick and H. Huppertz, Chem. Eur. J., 1997, 3, 679.

12. W. Schnick, International J. Inorg. Mater., 2001, 3, 1267.

13. G. Petzow, M. Herrmann, Silicon nitride ceramics. In: D.M.P. Mingos (ed.) Struct

Bond (High Performance Non-Oxide Ceramics II.), 2002, 102, 47.

14. Z.K. Huang, W.Y. Sun, D.S. Yan, J. Mater. Sci. Lett., 1985, 4, 255.

15. W.H. Zhu, P.L. Wang, W.Y. Sun, D.S. Yan, J. Mater. Sci. Lett., 1994, 13, 560.

16. G..Z. Cao, Z.K. Huang, X.R. Fu and D.S. Yan, Int. J. High Tech. Ceram., 1986,

2, 115.

17. Y.Q. Li, A.C.A. Delsing, G. de With and H.T. Hintzen, Chem. Mater., 2005, 17, 3242.

18. H.A. Höppe, Florian Stadler, O. Oeckler and W. Schnick, Angew. Chem. Int. Ed.,

2004, 43, 5540.

19. S.H.M. Poort, W. Janssen and G. Blasse, J. Alloys Comp., 1997, 260, 93.

20. S.H.M. Poort, H.M. Reijnhoudt, H.O.T. van der Kulp, G. Blasse, J. Alloys Comp.,

1996, 241, 75.

21. S.H.M. Poort, A. Meyerink, G. Blasse, J. Phys. Chem. Solids, 1997, 58, 1451.

22. J.S. Kim, P.E. Jeon, J.C. Choi, H.L. Park, Solid State Commun., 2005, 133, 187.

23. H.T. Hintzen, J.W.H. van Krevel and G. Botty, European Patent EP-1104799

A1,1999.

190

Chapter 10

24. H.A. Höppe, H. Lutz, P. Morys, W. Schnick and A. Seilmeier, J. Phys. Chem. Solids,

2000, 61, 2001.

25. H.T. Hintzen, A.C.A. Delsing and Y.Q. Li, PCT WO 2004/029177 A1.

26. P. V. Kelsey, Jr. and Jesse J. Brown, Jr, J. Electrochem. Soc., 1976, 123, 1384.

27. A. Boultif and D. Louer, J. Appl. Cryst., 2004, 37, 724.

28. Y.Q. Li, G. de With and H.T. Hintzen, unpublished results.

29. F. Liebau, Structural Chemistry of Silicates, Springer, Berlin, 1985.

30. Y.Q. Li, C.M. Fang, G. de With and H.T. Hintzen, J. Solid State Chem., 2004, 177,

4687.

31. Y.Q. Li, G. de With and H.T. Hintzen, J. Alloys Comp., 2004, 385, 1.

32. R.D. Shannon, Acta Cryst., 1976, A32, 751.

191

Chapter 11

Optical and magnetic properties of EuSi2O2N2

ABSTRACT

A new europium-silicon-oxynitride compound EuSi2O2N2 was obtained by a reaction of

Eu2O3, SiO2 and α-Si3N4 at 1300 °C under a nitrogen atmosphere. EuSi2O2N2 is indexed

on a monoclinic unit cell with a = 13.151(5) Å, b =17.311(5) Å, c = 7.956(2) Å, β =

104.12(4)° and V = 1756.56 Å3. EuSi2O2N2 shows a highly pure yellow color associated

with a very steep drop in the reflection spectrum with an optical absorption edge at about

512 nm (2.43 eV). On the other hand, EuSi2O2N2 can be efficiently excited in the visible

range 370 – 485 nm and shows a broad band emission peaking at about 568 nm

corresponding to the Eu2+ 4f65d1 4f7 transition. EuSi2O2N2 shows paramagnetic Curie

behaviour with an experimental magnetic moment of 7.89(3) µB in accordance with 7

unpaired spins of Eu2+. Additionally, no magnetic ordering can be observed down to 5K.

The divalent nature of the Eu ions in EuSi2O2N2 is evident from both luminescence and

magnetic properties.

Keywords: Europium-silicon-oxynitride, Divalent europium, X-ray powder diffraction,

Optical properties, Luminescence properties, Magnetic properties.

192

Chapter 11

11.1. Introduction

The valence state of europium has a great influence on both luminescence and

magnetic properties [1-3]. Europium can exist in trivalent and/or divalent state in

different host lattices strongly depending on the anion type and coordination; the

luminescence of europium displays sharp line emission and broad band emission for Eu3+

and Eu2+, respectively [1]. On the other hand, for magnetic properties, Eu3+ usually shows

a non-linear χ-1(T) characteristic with an effective magnetic moment of µeff = 3.4 – 3.51

while Eu2+ displays a linear χ-1(T) behavior with µeff ≈ 7.94 [2]. Apart from using a

reducing atmosphere, the divalent europium also can be obtained in reductive lattices,

such as silicate, silicon-nitride, sulfide, silicon-sulfide, boron-oxide, boron-nitride, boride

and cynamide [4 - 12]. Because Eu3+ is smaller than Eu2+, a large divalent lattice site also

has a high capability to reduce the valence of europium from Eu3+ into Eu2+ [13]. In

addition, Eu2+ ions can also exist in several glasses showing interesting luminescence

properties [14-17]. It is well-known that Eu2+ and Sr2+ have about the same ionic radius

and therefore exhibit very similar structural characteristics, as reflected in a number of

isotypic Eu2+- and Sr2+-compounds, such as Eu2Si5N8 [6] and Sr2Si5N8 [7], EuReSi4N7 [6,

8] and SrReSi4N7 (Re = Yb, Y) [8, 9]. To our knowledge no europium-silicon-oxynitride

compounds have been reported up to date. In contrast, a SrSi2O2N2 compound has already

been reported in the Sr-Si-O-N system with two modifications (i.e. a low- (1300 °C) and

a high-temperature (1600 °C) phase) which were indexed on an orthorhombic cell with

different lattice parameters [18].

Recently, the luminescence properties of Eu-doped MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba)

also have been investigated [19]. Therefore, it is interesting to check the existence of

EuSi2O2-δN2+2/3δ as SrSi2O2-δN2+2/3δ analog in the Eu-Si-O-N system. An important issue

is that europium-silicon-oxynitrides are possibly present as the second phase in Eu-doped

silicon nitride- and silicon-oxynitride-based phosphor materials which has influence on

the luminescence properties, especially at higher Eu concentrations. Accordingly, the

ability to distinguish those unknown Eu containing phases and investigate the

luminescence properties is very useful for better understanding and further optimization

of novel nitride- or oxynitride-based phosphor materials. In addition, Eu2SiO4 [20] and

193

Chapter 11

Eu2Si5N8 [21] have been found to be ferromagnets. EuSi2O2-δN2+2/3δ apparently lies in

between Eu2SiO4 and Eu2Si5N8 in composition, so it is also worthwhile to verify whether

or not EuSi2O2-δN2+2/3δ has ferromagnetic properties. In the present study, we report the

synthesis, X-ray powder diffraction data as well as the optical and magnetic properties of

a new compound, EuSi2O2-δN2+2/3δ, and compare them with those of Eu2SiO4 and

Eu2Si5N8.

11.2. Experimental

11.2.1. Preparation

EuSi2O2-δN2+2/3δ samples were prepared from Eu2O3, SiO2 and α-Si3N4 by a solid-

state reaction at high temperature. The starting materials were Eu2O3 (Rhône-Poulenc,

99.99%), SiO2 (Aerosil OX 50, Degussa) and α-Si3N4 (HC Stark, LC-12N, O content

~1.7%). Keeping the Eu/Si ratio at 0.5 while varying the O/N ratio in the range of 0.5 –

1.25, appropriate amounts of the starting materials were homogeneously wet-mixed by a

planetary ball mill for 4 -5 hours in isopropanol. After mixing the slurry was dried and

ground in an agate mortar. Subsequently, the dried powder mixtures were fired in a

molybdenum or an alumina crucible at 1300 °C for 6 – 12 h under nitrogen atmosphere in

a horizontal tube furnace. After firing, the samples were cooled down to room

temperature in the furnace and were ground again with an agate mortar for further

measurements. Eu2Si5N8 was prepared from nitrided Eu metal (Csre, 99.9%, lumps) and

α-Si3N4 powder (Permascand, P95H, α content 93.2%; Oxygen content: ~1.5%) and the

powder mixture was fired twice at 1300 – 1400 oC for 12 and 16 h, respectively, under

flowing 90%N2-10%H2 atmosphere [22].

11.2.2. X-ray powder diffraction

The obtained samples were identified by X-ray powder diffraction (Rigaku, D/MAX-B)

using Cu-Kα radiation at 40 kV and 30 mA with a graphite monochromator. The phase

formation is checked by the normal scan (2°/min) in the range of 10 – 90° 2θ. For further

indexing, the data were collected on powder samples mounted on an aluminum flat

holder using a step scan mode with a step size of 0.02˚ and a counting time of 20 second

194

Chapter 11

per step in the range 2θ 10 to 90˚. The XRD powder pattern of the single phase

compound was indexed using the indexing program DICVOL04 [23-25]. The possible

space groups are determined according to the systematic absences and the obtained unit

cells are further examined by fitting the full profile X-ray powder diffraction patterns

using Le Bail method [26] within the program GSAS [27-28].

11.2.3. Optical measurements

The diffuse reflectance, emission and excitation spectra of the samples were obtained

at room temperature by a Perkin Elmer LS 50B spectrophotometer equipped with a Xe

flash lamp. The reflection spectra were calibrated with the reflection of black felt

(reflection 3%) and white barium sulfate (BaSO4, reflection ~100%) in the wavelength

region of 230-700 nm. The excitation and emission slits were set at 2.5 nm. The emission

spectra were corrected by dividing the measured emission intensity by the ratio of the

observed spectrum of a calibrated W-lamp and its known spectrum from 300 to 900 nm.

Excitation spectra were automatically corrected for the variation in the lamp intensity by

a second photomultiplier and a beam-splitter. All the spectra were measured with a scan

speed of 100 nm/min. The optical absorption edge in this study was determined by the

wavelength value at which the reflection intensity is half of the lowest and highest of the

overall reflection intensity.

11.2.4 Magnetic measurements

The magnetic properties were measured with a dc extraction magnetometer and an ac

susceptometer (Quantum Design) in the temperature range 5 to 300 K. The dc

measurements were carried out in an applied magnetic field of 10000 Oe while the ac

susceptibility was recorded at a frequency of 1 kHz and at an amplitude field of 10 Oe in

zero dc field.

11.3. Results and discussion

11.3.1 Phase formation

A nearly single-phase compound with a small amount of europium silicates-like

secondary phase was obtained from the starting materials Eu2O3, α-Si3N4 and SiO2 in the

195

Chapter 11

molar ratios of 6:7:3. For higher nitrogen amounts (i.e. lower O/N ratio) significant

amounts of Eu2Si5N8 as a second phase appeared, whereas for lower nitrogen contents (i.e.

higher O/N ratio) a strong unknown phase became dominant in combination with a

EuSiO3-like phase. When preparing Eu2+ doped materials starting with Eu2O3, the driving

force for the conversion of Eu3+ into Eu2+ normally is considered to be the low partial

oxygen pressure in the firing atmosphere in combination with the divalent lattice site on

which Eu is incorporated. For the ratio Eu2O3:α-Si3N4:SiO2 = 6:7:3 the composition of

the resulting europium-silicon-oxynitride would then become EuSi2O2-δN2+2/3δ with δ ≈

0.5. However, in our case Eu can also be reduced by reaction of Eu2O3 with Si3N4

according to the reaction equation 6 Eu2O3 + Si3N4 12 EuO + 3 SiO2 + 2 N2 [5]. Such

a consumption of nitrogen as a consequence of reduction of Eu was indeed shown to

happen during the preparation of Eu2+-doped Sialon glasses from Eu2O3 [14]. Taking into

account the conversion of Si3N4 into SiO2 due to reaction with Eu2O3, the most probable

composition of the resulting europium-silicon-oxynitride is EuSi2O2-δN2+2/3δ with δ ≈ 0.

This approximate EuSi2O2N2 composition is supported by the weight increase we have

measured for oxidation of the compound at 1250 °C for 10 h in air (~ 9.5 wt%, as

compared to 10.4 wt% expected for δ = 0 and 11.9 wt% for δ = 0.5). Alkaline-earth

silicon-oxynitrides with the same composition MSi2O2N2 were found for M = Ca and Ba,

but surprisingly not for M = Sr which only yielded a single-phase material for the

composition SrSi2O2-δN2+2/3δ with δ ≈ 1 [19].

11.3.2. X-ray powder diffraction of EuSi2O2N2

EuSi2O2N2 was successfully indexed on a monoclinic unit cell with a = 13.151(5) Å, b

=17.311(5) Å, c = 7.956(2) Å, β = 104.12(4)° and V = 1756.56 Å3 with the figure-of-

merits M(20) = 10.9 and F(20) = 16.5(0.0072, 169). The space group is proposed to be

P21/a due to the observed extinction conditions 0k0 with k = 2n. The X-ray powder

diffraction data of EuSi2O2N2 are given in Table 11.1. The indexed data were further

examined by the simulation of the full-profile of X-ray powder diffraction pattern using

Le Bail method [26] which yielded reasonable R-factors (Rwp = 0.095, Rp = 0.059, χ2 =

7.3). Fig. 11.1(a) shows the observed and fitted X-ray powder diffraction pattern of

196

Chapter 11

Table 11.1. X-ray powder diffraction data of EuSi2O2N2.

h k l dobs (Å) 2θobs (deg.) 2θcal (deg.) ∆2θ (deg.) I/Io (%) 0 1 1 1 1 1 2 0 1 1 1 -2 0 2 2 2 4 -1 3 1 1 0 5 1 2 4 1 2 1 2 0 4 2 2 2 2 2 5 1 1 5 -2 0 0 3 5 1 0 2 4 2 1 1 3 1 7 -1 1 2 3 2 0 3 4 6 -1 2 3 3 2 8 -1 1 1 -4 1 2 -4 5 5 1 1 9 -1 6 2 -3 3 4 3 3 4 -4 1 3 4 4 3 3 6 2 2 1 8 3 0 6 4 3 0 -5 1 7 -4 1 7 4 0 8 4 1 4 5 0 6 5 3 0 5 1 0 -6

7.10942 5.69019 4.43609 3.88998 3.53646 3.44754 3.34104 3.17069 3.10368 2.96263 2.88974 2.83433 2.73297 2.61667 2.58015 2.53084 2.46626 2.39382 2.35166 2.31664 2.20878 2.15721 2.06337 2.02050 1.97304 1.93567 1.89752 1.86458 1.83987 1.81382 1.77174 1.75588 1.71850 1.66592 1.61223 1.60554 1.58434 1.54855 1.47799 1.44112 1.40864 1.36202 1.35002 1.31364

12.440 15.560 19.999 22.842 25.161 25.821 26.659 28.120 28.740 30.140 30.919 31.539 32.741 34.240 34.740 35.439 36.399 37.541 38.240 38.841 40.820 41.841 43.840 44.820 45.959 46.899 47.900 48.801 49.500 50.260 51.540 52.040 53.260 55.081 57.080 57.340 58.180 59.659 62.821 64.620 66.299 68.880 69.580 71.800

12.423 15.530 19.982 22.832 25.137 25.833 26.668 28.141 28.773 30.172 30.929 31.500 32.765 34.243 34.738 35.445 36.379 37.549 38.247 39.029 40.777 41.821 43.816 44.773 45.989 46.930 47.922 48.817 49.489 50.235 51.699 51.985 53.227 55.099 57.115 57.342 58.098 59.713 62.800 64.630 66.244 68.885 69.580 71.818

0.017 0.030 0.017 0.010 0.024 -0.012 -0.009 -0.021 -0.033 -0.032 -0.010 0.039 -0.024 -0.003 0.002 -0.006 0.020 -0.008 -0.007 -0.188 0.043 0.020 0.024 0.047 -0.030 -0.031 -0.022 -0.016 0.011 0.025 -0.159 0.055 0.033 -0.018 -0.035 -0.002 0.082 -0.054 0.021 -0.010 0.055 -0.005 0.000 -0.018

66.1 5.3 7.9 3.6 77.2 62.0 35.0 29.8 29.6 17.6 15.7 100.0 15.6 16.8 8.3 26.0 5.9 26.4 5.6 15.7 17.9 9.1 13.5 7.6 11.5 14.2 10.4 6.0 14.0 19.4 16.7 15.9 14.1 7.9 6.9 8.4 21.1 13.6 10.2 3.4 5.1 4.0 5.4 6.7

197

Chapter 11

2 10 -4 0 0 6 2 4 -6 5 6 4 4 8 4 0 0 1 0 9 5 1 6 -6 0 6 6 4 7 -6 0 15 0 2 9 5 1 15 1 3 4 6

1.30547 1.28709 1.26821 1.24891 1.23142 1.21201 1.20437 1.19437 1.17574 1.15746 1.15385 1.14431 1.13221 1.11655

72.320 73.520 74.800 76.160 77.441 78.920 79.520 80.320 81.861 83.440 83.760 84.620 85.739 87.240

72.330 73.507 74.810 76.175 77.461 78.911 79.544 80.269 81.902 83.446 83.801 84.654 85.710 87.237

-0.010 0.013 -0.010 -0.015 -0.020 0.009 -0.024 0.051 -0.041 -0.006 -0.041 -0.034 0.029 0.003

6.0 2.7 1.2 < 1 2.0 3.2 4.6 3.1 2.6 4.9 2.5 3.2 3.4 1.4

EuSi2O2N2. Although the X-ray diffraction pattern of EuSi2O2N2 (Fig. 11.1(a)) strongly

resembles that of SrSi2O2-δN2+2/3δ (δ ≈ 1) (Fig. 11.1(b)) with respect to the XRD lines, the

space group and the lattice parameters of the monoclinic cell are completely different

(Table 11.2). Therefore, in contrast to common practice that Eu2+-compounds correspond

with analogous Sr2+-compounds, MSi2O2-δN2+2/3δ turns out to be different for M = Eu2+

versus M = Sr2+ with respect to exact chemical composition and details of the crystal

structure; the typical example is M2SnS4 (M = Eu, Sr) [30].

Table 11.2. Lattice parameters of MSi2O2-δN2+2/3δ (M = Eu, Sr).

Formula Crystal system Space group Lattice constants

a (Å) b (Å) c (Å) β (°) V (Å3)

Figure-of-Merit M(20)

F(20) References

EuSi2O2N2 (δ = 0) Monoclinic P21/a 13.151(5) 17.311(5) 7.956(2) 104.12(4) 1756.56 10.9 16.5(0.0072, 169) This work

SrSi2ON8/3 (δ ≈ 1) Monoclinic P21/m 11.320(4) 14.107(6) 7.736(4) 91.87(3) 1234.67 10.8 14.9(0.0090, 150) [19], [29]

198

Chapter 11

10 20 30 40 50 60 70 80 90

10 15 20 25 30 35 40 45

Inte

nsity

(Cou

nts)

2θ (deg.)

EuSi2O 2N2

Inte

nsity

(Cou

nts)

2θ (deg.)

(a)

10 20 30 40 50 60 70 80 90

10 15 20 25 30 35 40 45

Inte

nsity

(Cou

nts)

2θ (deg.)

SrSi2O2-δN2+2/3δ

Inte

nsity

(Cou

nts)

2θ (deg.)

(b)

Fig. 11.1. The observed ( ) and simulated (solid line) X-ray powder diffraction pattern

of (a) EuSi2O2N2 and (b) SrSi2O2-δN2+2/3δ (δ ≈ 1) [19]. The difference profile (observed –

calculated) is shown at the bottom. The bars below the profile indicate the positions of all

reflections allowed for EuSi2O2N2.

199

Chapter 11

11.3.3. Optical properties

11.3.3.1. Diffuse reflection spectrum

Fig. 11.2 shows the diffuse reflection spectrum of EuSi2O2N2. A bright yellow color is

in agreement with a high reflection for wavelengths > 500 nm and a high absorption for

wavelengths < 500 nm (Fig. 11.2). A very sharp drop in reflection over a small

wavelength range (i.e. from the onset around 540 nm to 460 nm) is observed indicating a

high color purity of EuSi2O2N2. In addition, EuSi2O2N2 is found to show high tinting

strength which suggests that this material could be used as a special yellow pigment. The

color point of EuSi2O2N2 is at x = 0.452 and y = 0.492, as shown in a CIE (1931)

chromaticity diagram (see the inset in Fig. 11.2). The estimated optical absorption edge is

about 2.43 eV (~ 512 nm) which is consistent with its yellow color. As expected, this

value is larger than that of the red colored nitride compound Eu2Si5N8 (~ 2.06 eV/604 nm

[22]), similar to that of the goldenrod oxide compound Eu2SiO4 (~2.46 eV/508 nm [31]),

and much smaller than that of grey-white SrSi2O2-δN2+2/3δ (δ ≈ 1) (~ 4.52 eV/275 nm [19]).

Clearly, the replacement of Sr by Eu significantly decreases the optical absorption edge,

whose effect is more profound for the nitride lattice as compared to the oxide lattice. The

yellow color of EuSi2O2N2 is ascribed to the 4f 5d transitions of Eu2+ similar to the

Ce3+ ion in the case of red γ-Ce2S3 [32].

150 200 250 300 350 400 450 500 550 600 650 700 750

0

10

20

30

40

50

60

70

80

90

100

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.90.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

Ref

lect

ion

(%)

Wavelength (nm)

Y

X

Fig. 11.2. Diffuse reflection spectrum of EuSi2O2N2. The inset shows the color point in

the CIE chromaticity coordination.

200

Chapter 11

11.3.3.2. Luminescence properties

The excitation and emission spectra of EuSi2O2N2 are shown in Fig. 11.3 with the

broad bands.

200 300 400 500 600 700 800 900

0.0

0.2

0.4

0.6

0.8

1.0In

tens

ity (a

.u.)

Wavelength (nm)

λem= 570 nm λexc= 460 nm

Fig. 11.3. Excitation (left) and emission (right) spectra of EuSi2O2N2 (λem = 570 nm, λexc

= 460 nm).

In general, a broad charge-transfer excitation band of Eu3+ lies at wavelengths < 300 nm,

and the emission of Eu3+ shows sharp line features due to optical transitions between

levels of the 4fn configuration [1]. Therefore, a broad emission band without sharp line-

emission indicates that europium is in the divalent state in EuSi2O2N2. Four obvious

bands can be observed in the excitation spectrum. The one at the shortest wavelength

around 250 nm corresponds to the host lattice excitation (i.e. the valence to conduction

band transitions) close to the case of MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba) [19]. The other

three bands centered at about 340, 391 and 465 nm (Table 3), along with a very weak

shoulder around 300 nm, can be essentially assigned to the 4f 5d transitions of the

Eu2+ ion. Such long-wavelength excitation bands are very similar to the case of Eu2Si5N8

where the excitation band also significantly extends to the lower energy range i.e. ~ 472

nm (Table 3) [22]. A single broad emission band with a maximum at about 568 nm

(FWHM ~ 87 nm) can be observed from the emission spectrum (Fig. 11.3, Table 3). In

201

Chapter 11

addition, by varying the excitation wavelength, the emission band does not show any shift

and shape-change in its spectrum. The symmetrical profile (Fig. 11.3) indicates that the

Eu centers from which the emission is observed (possibly after energy transfer from other

centers) should have rather similar surroundings or are insensitive to the local structures.

The Stokes shift estimated for the 5d ↔ 4f transition of Eu2+ in EuSi2O2N2 (≈ 3700 cm-1)

is lower than for Eu2Si5N8 (≈ 4900 cm-1) (Table 3). As compared to the luminescence

properties of Eu2+-doped MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba), EuSi2O2N2 shows most

resemblance with SrSi2O2-δN2+2/3δ:Eu2+ [19] in agreement with the fact that the crystal

structures of both compounds are strongly related, while they significantly differ from

those of MSi2O2N2:Eu2+ (M = Ca, Ba).

11.3.4. Magnetic properties

From the magnetic susceptibility as a function of temperature, it is clearly seen that

EuSi2O2N2 shows Curie behavior (Fig. 11.4).

0 50 100 150 200 250 300

0.0

0.5

1.0

1.5

2.0

0 50 100 150 200 250 300

0

5

10

15

20

25

30

35

40

χ (e

mu/

mol

)

Temperature (K)

measured fitted

1/χ

(mol

/em

u)

T(K)

Fig. 11.4. Temperature dependence of the molar magnetic susceptibility of EuSi2O2N2. The inset shows the inverse magnetic susceptibility of EuSi2O2N2.

202

Chapter 11

Table 11. 3. Optical and magnetic data for EuSi2O2N2, Eu2Si5N8 and Eu2SiO4.

Compound Color Absorption edge (nm)

Excitation band (nm)

Emission band (nm)

Stokes shift (cm-1)

Magnetic moment (µB)

References

EuSi2O2N2 Yellow 512 340, 391, 465 568 3700 7.89 This work

Eu2Si5N8

Red

604

263, 344, 394, 472, 511

680

4900

7.66 7.67

This work [21]

Eu2SiO4 Goldenrod 508 - 570 - 6.60-7.01

[31] [20]

Chapter 11

The experimental magnetic moment is 7.89(3) µB as derived from the fitted 1/χ versus T

(5 to 300 K) dependence (see inset in Fig. 11.4). The obtained magnetic moment is very

close to the theoretical value of 7.94 µB for free Eu2+ ions [33] which further confirm that

Eu is in the divalent state in EuSi2O2N2, in agreement with its luminescence properties.

No magnetic ordering was detected down to 5 K in EuSi2O2N2. In fact, the Weiss

temperature determined from the fit of the inset of Fig. 11.4 is less than 0.5 K, suggesting

that any magnetic interactions are very weak indeed. The magnetic moment of Eu2Si5N8

(7.66 µB, Table 3) is also in agreement with divalent Eu, while the lower values reported

for Eu2SiO4 (6.60-7.01 µB, Table 3) indicate the presence of some Eu3+ in addition to

Eu2+ [20].

11.4. Conclusions

A new europium-silicon-oxynitride compound with composition EuSi2O2N2 was

obtained, which crystallizes in a monoclinic with a = 13.151(5) Å, b =17.311(5) Å, c =

7.956(2) Å, β = 104.12(4)° and V = 1756.56 Å3. The optical absorption edge is 512 nm

(i.e. 2.43 eV) in agreement with its yellow color. In addition, EuSi2O2N2 possesses a

highly saturated tint. Excitation in the range of 370 – 485 nm yields a band emission with

a maximum at about 568 nm (FWHM ~ 87 nm) ascribed to the Eu2+ 5d 4f transition.

An unusual long-wavelength excitation band is observed at about 465 nm due to the

presence of nitrogen. EuSi2O2N2 perfectly follows the Curie law showing paramagnetic

behaviour with a magnetic moment of 7.89(3) µB down to 5 K due to 7 unpaired spins of

Eu2+. Based on the luminescence and magnetic properties, the europium ions are

confirmed to be present in the divalent state in EuSi2O2N2.

204

Chapter 11

References:

1. G. Blasse and B.C. Grabmaier: Luminescent materials. (Springer-Verlag, Berlin,

1994).

2. P. Mohn: Magnetism in the solid state: an introduction. (Springer, Berlin, 2003).

3. A.H. Morrish: The physical principle of magnetism. (Wiley, Chichester, 1965).

4. M. Catti, G. Gazzoni, G. Ivaldi, G. Zanini, Acta Cryst. B39, 674 (1983).

5. R Marchand, P l'Haridon, Y. Laurent, J. Solid State Chem. 24, 71 (1978).

6. H. Huppertz and W. Schnick, Acta Cryst. C53, 1751 (1997).

7. T. Schlieper, W. Milius and W. Schnick, Z. Anorg. Allg. Chem. 621, 1380 (1995).

8. Y.Q. Li, C.M. Fang, G. de With and H.T. Hintzen, J. Solid State Chem. 177, 4687

(2004).

9. H. Huppertz and W. Schnick, Anorg. Allg. Chem. 623, 212 (1997).

10. C. Wickleder, Z. Anorg. Allg. Chem. 627, 1693 (2001).

11. O. Reckeweg and F.J. DiSalvo, Z. Anorg. Allg. Chem. 629, 177 (2003).

12. U.Berger, W. Schnick, J. Alloys Compd. 206, 179 (1994).

13. E. Ekambaram, J. Alloys Comp. 390, (L1) 2005.

14. D. de Graaf, H.T. Hintzen, S. Hampshire and G. de With, J. Eur. Ceram. Soc. 23,

1093 (2003).

15. H.T. Hintzen, J.W.H. van Krevel, D. de Graaf, R. Metselaar, Y. Menke and S.

Hampshire, J. Mater. Sci. 39, 2237 (2004).

16. J.R. Qiu, K. Miura, N. Sugimoto, K. Hirao, J. Non-Crystalline Solids. 213, 266

(1997).

17. Y. Zhang, A. Quaranta b, G.D. Soraru, Optical Materials. 24, 601 (2004).

18. W.H. Zhu, P.L. Wang, W.Y. Sun, D.S. Yan, J. Mater. Sci. Lett. 13, 560 (1994).

19. Y.Q. Li, G. de With and H.T. Hintzen, Chem. Mater. 17, 3242 (2005).

20. M.W. Shafer, T.R. McGuire, J.C. Suits, Phys. Rev. Letters. 11, 251 (1963).

21. H.A. Höppe, H. Trill, B.D. Mosel, H. Eckert, G. Kotzyba, R. Pottgen. W. Schnick,

J. Phys. Chem. Solids. 63, 53 (2002).

22. A. Boultif, D. Louer, J. Appl. Cryst. 37, 724 (2004).

23. D. Louer and M. Louer, J. Appl. Cryst. 5, 271 (1972).

205

Chapter 11

24. A. Boultif and D. Louer, J. Appl. Cryst. 24, 987 (1991).

25. A. Le Bail, H. Duroy and J.L. Fourquet, Mater. Res. Bull. 23, 447 (1988).

26. A.C. Larson, R.B. Von Dreele, General structure analysis system (GSAS), Los

Alamos National Laboratory Report LAUR 86-748. (2004)

27. B.H. Toby, J. Appl. Cryst. 34, 210 (2001).

28. Y.Q. Li, G. de With and H.T. Hintzen, Submitted to J. Mater. Chem. (Chapter 10)

29. R. Pocha, M. Tampier, R.D. Hoffmann, B.D. Mosel, R. Poettgen, D. Johrendt, Z.

Anorg. Allg. Chem. 629, 1379 (2003).

30. E. Kaldis, P. Streit, P. Wachter, J. Phys. Chem. Solids. 32, 159 (1971).

31. G. Gauthier, S. Jobic, F. Boucher, P. Macaudiere, D. Huguenin, J. Rouxel, R. Brec,

Chem. Mater. 10, 2341 (1998).

32. M.M. Abd-Elmeguid, J.P. Sanchez and H. Micklitz, J. Phys. C: Solid State Phys. 16,

L543 (1983).

206

Summary White-light LEDs are rapidly growing for general illumination because of their

compelling advantages over traditional incandescent and fluorescent lamps, such as lower

energy consumption, longer lifetime and environmentally friendly. It is expected that

white-light LEDs will be the next generation lighting. This is a great challenge for the

conversion phosphors in this lighting revolution. As an explorative research this thesis

focuses on search for and creation of novel rare-earth doped silicon-nitride based

materials with improved properties for white-light LED applications. According to the

composition of the host lattices, this thesis deals with three major types of rare-earth

(Eu2+, Ce3+ and Tb3+) doped silicon-nitride based host lattices: 1. nitrides (Chapters 1 - 5).

2. nitride-carbides (Chapters 6 – 7) and 3. oxynitrides (Chapters 8 – 11). The emphasis

throughout this thesis is on understanding the relationship between the luminescence

properties on the one hand and the structure and chemical composition on the other hand.

In addition, the invention of patentable materials for use in white LED lighting is part of

this thesis work.

In chapter 1 the luminescence properties of Ce3+-activated M2Si5N8 (M = Ca, Sr, Ba)

using Na+ or Li+ as a charge compensator are described. It is shown that the maximum

solubility of Ce3+ with Li+ in M2Si5N8 is about 2.5 mol% for both Ca2Si5N8 and Sr2Si5N8,

and about 1.0 mol% for Ba2Si5N8. The Ce3+-doped M2Si5N8 phosphors exhibit broad

emission bands with maxima at 470, 553 and 451 nm for M = Ca, Sr, Ba, respectively. In

addition, M2Si5N8: Ce3+, Li+ (M = Sr, Ba) obviously shows two Ce3+ emission centers

due to the fact that the Ce3+ ions occupy two different M sites. The influence of using Na+

instead of Li+ ion as charge compensator on emission and excitation properties is small

but Na+ enhances the emission intensity because of larger solubility of Ce3+ in M2Si5N8

(M = Ca, Sr). With increasing the ionic radius going from Ca to Sr and Ba, the emission

intensity of Ce3+ (for excitation 395 – 400 nm) decreases related to a lower Ce3+

solubility in this sequence. An intense absorption and excitation band in the UV-blue

range (370 – 450 nm) points out that these materials, in particular Sr2Si5N8:Ce,Li(Na), are

promising conversion phosphors for white-light LEDs.

207

Summary

In chapter 2 the luminescence properties of red-emitting M2Si5N8:Eu2+ (M = Ca, Sr, Ba)

have been studied. Eu2+ can be completely incorporated into the M2Si5N8 (M = Sr, Ba)

lattice because these compounds are isostructural with Eu2Si5N8. In contrast, the

maximum solubility of Eu2+ is only about 7 mol% in Ca2Si5N8 with a different structure.

Eu2+-doped M2Si5N8 shows a typical broad band emission from yellow-orange to red

(575 – 680 nm) depending on the type of M (e.g., orange to red for M = Ca, Sr; yellow to

red for M = Ba). A high covalency and large crystal field splitting on the 5d band of Eu2+

due to the presence of nitrogen is believed to be responsible for such unusual long-

wavelength emission. The emission band shifts to longer wavelength with increasing the

Eu concentration, which is attributed to the increasing Stokes shift and reabsorption by

Eu2+. The conversion efficiency of M2Si5N8:Eu2+ (M = Sr, Ba) are higher than that of

Ca2Si5N8:Eu2+ excited by 465 nm. The results demonstrate that red-emitting

Sr2Si5N8:Eu2+ phosphor has high potential for white-LED applications.

In chapter 3 the effect of Ca substitution for Sr on the structure and luminescence of

Sr2Si5N8:Eu2+ (5 mol%) is described. The Ca ion preferentially occupies the larger Sr site

in Sr2Si5N8:Eu2+ in order to keep the structure stable, as evidenced by the Rietveld

refinement analysis and in agreement with the lattice energy calculations, while the Eu

ion statistically distributes over two crystallographic Sr sites. Incorporation of Ca hardly

modifies the excitation spectrum of Eu2+, but the position of the emission band shifts to

longer wavelength due to an increase of the Stokes shift. Accordingly, the emission

properties of Sr2Si5N8:Eu2+ can be tailored by partial replacement of Sr by Ca.

In chapters 4 and 5 the synthesis, structure and luminescence properties of undoped

and Eu2+ or Ce3+-doped MYSi4N7 (M = Sr, Ba) are reported. Three new compounds

MYSi4N7 (M = Sr, Ba, Eu) are found to be isostructural with BaYbSi4N7. Eu2+-doped

MYSi4N7 yields a broad green emission band in the range of 503 - 527 nm and 548 – 570

nm for M = Ba, Sr, respectively, depending on the Eu2+ concentration. The Eu2+ emission

band shows a red shift with increasing Eu2+ concentration mainly caused by an increase

of the crystal field splitting and Stokes shift. Ce3+-doped MYSi4N7 exhibits a bright blue

emission band with a maximum at about 417 and 450 nm for M = Ba, Sr, respectively. In

contrast to the Eu-doped materials, the position of the emission band of Ce3+ is

independent of the Ce3+ concentration which is ascribed to a lower solubility of Ce3+. On

208

Summary

a larger lattice site (i.e. for M = Ba), both Eu2+ and Ce3+ exhibit a shorter wavelength

emission due to a smaller crystal field splitting and smaller Stokes shift. SrYSi4N7: Eu2+

can be very well excited by 390 nm radiation, which makes this material attractive as

conversion phosphor for LED lighting applications.

In chapters 6 and 7 two new rare-earth-silicon-nitride-carbides (viz. YTbSi4N6C and

Y2Si4N6C) are achieved and their crystal structures determined. The two compounds are

isostructural and crystallize in the monoclinic crystal system with the space group P21/c.

These materials can be deduced from the MYSi4N7 lattice through a special chemical

replacement of MN by LnC (Ln = Tb and Y), but the resulting crystal structure is

different due to the size difference between the two cation ions. YTbSi4N6C shows an

unusual long-wavelength 4f-5d excitation band of Tb3+ at about 300 nm due to the highly

covalent silicon-nitride-carbide network and large crystal field splitting, and typical Tb3+

green 5D4 7F5 line emissions. More particular, through the direct Ce3+ Tb3+ energy

transfer to the 5D4 level of Tb3+ (normally to 5D3 level) in Ce3+-doped YTbSi4N6C, a

green Tb3+ line emission can be realized by the excitation of Ce3+ ions in the visible range

(i.e. 390 – 480 nm). This demonstrates a new approach to use the line emission of

trivalent rare-earth ions for white-light LED applications (Chapter 6). Ce3+-doped

Y2Si4N6C shows an unusual long-wavelength 4f-5d excitation band of Ce3+ in the range

of 380 - 450 nm due to the highly covalent silicon-nitride-carbide network combined with

large crystal field splitting due to the coordinating N3- ions. For excitation in the UV-blue

range (370 – 450 nm), Y2Si4N6C:Ce3+ gives rise to a green emission with a maximum in

the range of 530 – 560 nm also showing high promise for use as a conversion phosphor in

white-emitting LEDs (Chapter 7).

Chapter 8 deals with the preparation, structural and luminescence properties of Eu-

doped alkaline-earth silicon aluminium oxynitrides, M2Al2-xSixO4-xNx (M = Ca, Sr, Ba).

In contrast to the case of Sialons (deduced from the nitride material Si3N4 by replacement

of (SiN)+ by (AlO)+), these materials are obtained starting from the oxide MAl2O4,

through (SiN)+ substitution for (AlO)+. The maximum solubility of (SiN)+ in MAl2-

xSixO4-xNx with stuffed tridymite structure strongly depends on the type of M ion. The

solubility of (SiN)+ is negligible for M = Ca and Sr, whereas it is up to x = 0.6 for M =

Ba probably related to its largely distorted (AlO4) tetrahedra (i.e. largely different Al-O

209

Summary

distances). As a consequence the luminescence of Eu2+ is hardly changed by introducing

nitrogen for Eu-doped MAl2-xSixO4-xNx (M = Ca, Sr). In contrast, BaAl2-xSixO4-xNx:Eu2+

(x = 0.3) exhibits a long-wavelength excitation band peaking at about 440 nm

corresponding to a green emission at about 500 - 526 nm. This red-shift of both excitation

and emission bands due to the incorporation of nitrogen (for BaAl2O4:Eu2+, λexc ≈ 390 nm

and λem ≈ 500 nm) can be understood from increased covalency and crystal field splitting.

The luminescence properties of BaAl2-xSixO4-xNx:Eu2+ can be tailored by not only the

amount of (SiN)+ but also the Eu concentration and as a consequence BaAl2-xSixO4-

xNx:Eu2+ is a valuable green-emitting phosphor for use in white-light LEDs.

For x = 2 in M2Al2-xSixO4-xNx (M = Ca, Sr, Ba), the composition MSi2O2N2 results

which is studied in chapters 9, 10 and 11. The focus of these chapters is on the synthesis,

structural characterization and luminescence properties of a new class of oxynitride

phosphors, with general composition MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba, Eu; with an ideal

composition MSi2O2N2 for δ = 0) using Eu2+ and Ce3+ as the activators. Two new

oxynitride compounds, BaSi2O2N2 (Chapter 9) and EuSi2O2N2 (Chapter 11), are found

and crystallographically indexed. It is shown that all MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba, Eu)

compounds have a monoclinic unit cell but with different structure. MSi2O2-δN2+2/3δ:Eu2+

materials can be efficiently excited in the UV to blue region (370 - 460 nm), yielding a

blue-green emission at 490-500 nm for BaSi2O2N2: Eu2+, a yellow emission at 560 nm for

CaSi2O2-δN2+2/3δ:Eu2+ (δ ≈ 0) and a green-yellow emission peaking from 530 to 570 nm

for SrSi2O2-δN2+2/3δ:Eu2+ (δ ≈ 1). The results point out that MSi2O2-δN2+2/3δ:Eu2+ is a

highly promising class of phosphor materials for use in white-light LEDs.

Ce3+-doped MSi2O2-δN2+2/3δ shows UV-blue emission with maxima at about 392, 473

and 396 nm for M = Ca, Sr and Ba, respectively, of which CaSi2O2N2:Ce,Na has the

highest luminescence efficiency. Interestingly, BaSi2O2N2:Ce3+ by itself emits white light

when excited with 365 nm radiation (Chapter 10).

In the last chapter (Chapter 11) it is described that EuSi2O2N2 possesses a bright-

yellow color with high purity. EuSi2O2N2 can be excited in the visible range (370 – 485

nm) and shows a broad band emission peaking at about 568 nm. By combination of the

luminescence and magnetic properties, the europium ions are confirmed to be present in

the divalent state in EuSi2O2N2.

210

Summary

In short, with respect to the emission and excitation characteristics and conversion

(quantum) efficiency, M2Si5N8:Eu2+, Sr2Si5N8:Ce3+, MSi2O2-δN2+2/3δ:Eu2+, BaAl2-xSixO4-

xNx:Eu2+ and Y2Si4N6C:Ce3+ are promising materials as potential conversion phosphors

for white LED lighting applications, and therefore, several patents have been applied.

Moreover, such silicon-nitride based materials show high chemical and thermal stability.

Particularly, this study has provided a better understanding of the relationship between

the crystal structure/chemical composition and the luminescence properties of rare-earth

ions (i.e. Eu2+, Ce3+ and Tb3+) in silicon-nitride based materials.

211

Samenvatting

Toepassing van wit-licht LED’s is sterk groeiende voor algemene

verlichtingsdoeleinden vanwege een aantal grote voordelen t.o.v. de traditionele

gloeilampen en fluorescentie lampen, zoals lager energieverbruik, langere levensduur en

milieuvriendelijkheid. De verwachting is dat wit-licht LED’s de volgende generatie

lampen zal zijn. In deze revolutie op verlichtingsgebied is een grote rol weggelegd voor

conversie-fosforen. Als een stuk exploratief onderzoek ligt in dit proefschrift de nadruk

op het bedenken en realiseren van nieuwe zeldzame-aard gedoteerde siliciumnitride-

gebaseerde materialen met verbeterde eigenschappen voor wit-licht LED toepassingen.

Het proefschrift handelt over drie klassen van zeldzame-aard (Eu2+, Ce3+ en Tb3+)

gedoteerde siliciumnitride-gebaseerde materialen, die wat betreft samenstelling van het

gastrooster als volgt ingedeeld kunnen worden: 1. nitrides (hoofdstukken 1-5); 2. nitride-

carbides (hoofdstukken 6-7); en 3. oxynitrides (hoofdstukken 8-11). De nadruk ligt in dit

proefschrift op het begrijpen van de relatie tussen structuur en samenstelling aan de ene

kant en luminescentie eigenschappen aan de andere kant. Bovendien is het uitvinden van

patenteerbare materialen voor gebruik in wit-licht LED’s onderdeel van het beschreven

onderzoek.

In hoofdstuk 1 worden de luminescentie eigenschappen van Ce3+- geactiveerd

M2Si5N8 (M = Ca, Sr, Ba) beschreven, waarbij Na+ en Li+ als ladingscompensator is

gebruikt. De maximale oplosbaarheid van Ce3+ in M2Si5N8 met Li+ als

ladingscompensator is ongeveer 2.5 mol% voor zowel Ca2Si5N8 als Sr2Si5N8, en ca. 1.0

mol% in Ba2Si5N8. De Ce3+-gedoteerde M2Si5N8 fosforen vertonen brede emissie banden

met maxima bij 470, 553 en 451 nm voor M = Ca, Sr, Ba, respectievelijk. Bovendien is

het duidelijk dat in M2Si5N8: Ce3+, Li+ (M = Sr, Ba) twee Ce3+ emissie centra aanwezig

zijn ten gevolge van het feit dat de Ce3+ ionen twee verschillende M roosterplaatsen

bezetten. De invloed van Na+ in plaats van Li+ als ladingscompensator op de emissie- en

excitatie-eigenschappen is gering, maar Na+ vergroot de emissie-intensiteit ten gevolge

van een grotere oplosbaarheid van Ce3+ in M2Si5N8 (M = Ca, Sr). Met toenemende

ionstraal in de volgorde Ca < Sr < Ba, wordt de emissie-intensiteit van Ce3+ kleiner, wat

gerelateerd is aan een in deze volgorde afnemende Ce3+ oplosbaarheid. Een intense

212

Samenvatting

absorptie- en excitatie-band in het UV-blauwe gebied (370 – 450 nm) geeft aan dat deze

materialen, in het bijzonder Sr2Si5N8:Ce,Li(Na), veelbelovende conversie-fosforen zijn

voor wit-licht LED’s.

In hoofdstuk 2 worden de luminescentie eigenschappen van rood-emitterend

M2Si5N8:Eu2+ (M = Ca, Sr, Ba) bestudeerd. Eu2+ kan volledig ingebouwd worden in het

M2Si5N8 (M = Sr, Ba) rooster omdat deze verbindingen isostructureel zijn met Eu2Si5N8.

In tegenstelling hiermee is de maximale oplosbaarheid van Eu2+ in Ca2Si5N8 met een

andere structuur slechts ca. 7 mol%. Eu2+-gedoteerd M2Si5N8 vertoont een kenmerkende

brede band emissie die varieert van geel-oranje naar rood (575 – 680 nm), afhankelijk

van het type M (b.v., oranje tot rood voor M = Ca, Sr; geel tot rood voor M = Ba). Een

dergelijke emissie bij ongewoon hoge golflengte wordt veroorzaakt door de invloed van

een hoge covalentie en een grote kristalveldsplitsing van de 5d band van Eu2+ ten gevolge

van de aanwezigheid van stikstof. Voor grotere Eu concentraties schuift de emissie band

naar langere golflengten, wat wordt toegeschreven aan een toenemende Stokes shift en

reabsorptie door Eu2+. Voor een excitatie-golflengte van 465 nm is het conversie-

rendement van M2Si5N8:Eu2+ (M = Sr, Ba) groter dan dat van Ca2Si5N8:Eu2+. De

resultaten laten zien dat de rood-emitterende Sr2Si5N8:Eu2+ fosfor een hoge potentie heeft

voor toepassing in wit-licht LED’s.

In hoofdstuk 3 wordt het effect van de vervanging van Sr door Ca op de structuur

en luminescentie van Sr2Si5N8:Eu2+ (5 mol%) beschreven. Zoals blijkt uit Rietveld

structuur verfijning en in overeenstemming met rooster-energie berekeningen, bouwt het

Ca ion bij voorkeur in op de grotere Sr roosterplaats in Sr2Si5N8:Eu2+ om de structuur

stabiel te houden, terwijl de Eu ionen statistisch verdeeld zijn over de twee

kristallografische Sr roosterplaatsen. De inbouw van Ca verandert het excitatiespectrum

van Eu2+ nauwelijks, maar de positie van de emissieband schuift naar langere golflengten

ten gevolge van de toename van de Stokes shift. Als gevolg kunnen de emissie

eigenschappen van Sr2Si5N8:Eu2+ op maat ingesteld worden door de gedeeltelijke

vervanging van Sr door Ca.

In hoofdstukken 4 en 5 worden de synthese, structuur en luminescentie

eigenschappen van Eu2+-, Ce3+- en niet-gedoteerd MYSi4N7 (M = Sr, Ba) gerapporteerd.

Drie nieuwe verbindingen MYSi4N7 (M = Sr, Ba, Eu) blijken isostructureel te zijn met

213

Samenvatting

BaYbSi4N7. Eu2+-gedoteerd MYSi4N7 vertoont een brede groene emissieband in het

golflengtegebied van 503-527 nm en 548-570 nm voor M = Ba en Sr, respectievelijk,

afhankelijk van de Eu2+ concentratie. Met toenemende Eu2+ concentratie schuift de Eu2+

emissieband naar het rode gebied, wat voornamelijk veroorzaakt wordt door een toename

van de kristalveldsplitsing en Stokes shift. Ce3+-gedoteerd MYSi4N7 vertoont een heldere

blauwe emissieband waarvan het maximum ligt rondom 417 en 450 nm voor M = Ba en

Sr, respectievelijk. In tegenstelling tot de Eu-gedoteerde materialen, is de positie van de

Ce3+ emissieband onafhankelijk van de Ce3+ concentratie, wat wordt toegeschreven aan

een lagere oplosbaarheid van Ce3+ in het rooster. Voor een grotere roosterplaats (d.w.z.

voor M = Ba) vertoont zowel Eu2+ als ook Ce3+ emissie bij een kortere golflengte, ten

gevolge van een kleinere kristalveldsplitsing en een kleinere Stokes shift. SrYSi4N7:Eu2+

kan zeer goed geëxciteerd worden m.b.v. 390 nm straling, wat dit materiaal aantrekkelijk

maakt als conversie-fosfor voor LED verlichtingsdoeleinden.

In hoofdstukken 6 en 7 worden twee nieuw verkregen zeldzame-aard silicium-

nitride-carbides (nl. YTbSi4N6C and Y2Si4N6C) gepresenteerd en hun kristalstructuren

bepaald. De twee verbindingen zijn isostructureel en kristalliseren in het monocliene

kristalsysteem met ruimtegroep P21/c. De materialen kunnen afgeleid worden van het

MYSi4N7 rooster door chemische vervanging van MN door LnC (Ln = Tb and Y), maar

de resulterende kristalstructuur is anders ten gevolge van het verschil in grootte tussen de

twee kationen. Ten gevolge van het sterk covalente silicium-nitride-carbide netwerk en

de grote kristalveldsplitsing vertoont YTbSi4N6C een Tb3+ 4f-5d excitatieband bij

ongewoon lange golflengten (ca. 300 nm), en de kenmerkende groene 5D4 7F5

lijnemissie van Tb3+. Meer speciaal kan voor Ce3+-gedoteerd YTbSi4N6C een groene Tb3+

lijnemissie verkregen worden door excitatie van Ce3+ ionen in het zichtbare gebied (d.w.z.

390 – 480 nm) als gevolg van energie-overdracht van Ce3+ naar het 5D4 niveau van Tb3+

(normaliter is dit het 5D3 niveau). Dit geeft een nieuwe route aan voor het gebruik van

lijnemissie van zeldzame-aard ionen ten behoeve van wit-licht LED applicaties

(Hoofdstuk 6). Ce3+-gedoteerd Y2Si4N6C vertoont een 4f-5d Ce3+ excitatieband bij

ongewoon lange golflengten in het gebied 380-450 nm, ten gevolge van het sterk

covalente silicium-nitride-carbide netwerk gecombineerd met de grote kristalveldsplitsing

door de coördinerende N3- ionen. Bij excitatie in het UV-blauwe gebied (370 – 450 nm),

214

Samenvatting

emitteert Y2Si4N6C:Ce3+ groene straling met een maximum in het gebied 530–560 nm,

zodat dit materiaal veelbelovend is voor gebruik als conversie-fosfor in wit-licht LED’s

(Hoofdstuk 7).

Hoofdstuk 8 handelt over de synthese, structuur en luminescentie eigenschappen van

Eu-gedoteerde aard-alkali silicium-aluminium-oxynitrides, M2Al2-xSixO4-xNx (M = Ca, Sr,

Ba). Precies omgekeerd als bij Sialons (die afgeleid zijn van de nitride verbinding Si3N4

door de vervanging van (SiN)+ door (AlO)+), worden deze materialen verkregen door

vervanging van (AlO)+ door (SiN)+ in het oxide MAl2O4. De maximale oplosbaarheid

van (SiN)+ in MAl2-xSixO4-xNx met de “stuffed” tridymiet structuur hangt sterk af van het

type M ion. De oplosbaarheid van (SiN)+ is verwaarloosbaar voor M = Ca en Sr, terwijl

die voor M = Ba maximaal x = 0.6 is, wat waarschijnlijk gerelateerd is aan de sterk

verstoorde (AlO4) tetraëders (d.w.z. grote verschillen in Al-O afstanden). Ten gevolge

hiervan is de luminescentie van Eu2+ nauwelijks veranderd voor Eu-gedoteerd MAl2-

xSixO4-xNx (M = Ca, Sr) door toevoeging van stikstof. In tegenstelling hiermee, vertoont

BaAl2-xSixO4-xNx:Eu2+ (x = 0.3) een excitatieband bij lange golflengte (maximum rond

440 nm), corresponderend met een groene emissie bij ongeveer 500-526 nm. Deze rood-

verschuiving van zowel de excitatie- als ook de emissie-band ten gevolge van de inbouw

van stikstof (ter vergelijking voor BaAl2O4:Eu2+: λexc ≈ 390 nm en λem ≈ 500 nm) kan

begrepen worden op basis van een toegenomen covalentie en kristalveldsplitsing. De

luminescentie eigenschappen van BaAl2-xSixO4-xNx:Eu2+ kunnen niet alleen ingesteld

worden door de hoeveelheid (SiN)+, maar ook door de Eu concentratie en dientengevolge

is BaAl2-xSixO4-xNx:Eu2+ een bruikbare groen-emitterende fosfor voor toepassing in wit-

licht LED’s.

Voor x = 2 in M2Al2-xSixO4-xNx (M = Ca, Sr, Ba) resulteert de samenstelling

MSi2O2N2, die wordt bestudeerd in de hoofdstukken 9, 10 en 11. In deze hoofdstukken

wordt de aandacht gericht op de synthese, structurele karakterisering en luminescentie

eigenschappen van een nieuwe klasse van oxynitride fosforen met de algemene

samenstelling MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba, Eu; met ideale samenstelling MSi2O2N2

voor δ = 0) gedoteerd met Eu2+ en Ce3+ als de activatoren. Twee nieuwe oxynitride

verbindingen, nl. BaSi2O2N2 (Hoofdstuk 9) en EuSi2O2N2 (Hoofdstuk 11), zijn ontdekt

en kristallografisch geïndexeerd. Er is aangetoond dat alle MSi2O2-δN2+2/3δ (M = Ca, Sr,

215

Samenvatting

Ba, Eu) verbindingen een monocliene eenheidscel hebben, maar met verschillende

kristalstructuur. MSi2O2-δN2+2/3δ:Eu2+ kan efficiënt aangeslagen worden in het UV-blauwe

gebied (370 - 460 nm), resulterend in een blauw-groene emissie rond 490-500 nm voor

BaSi2O2N2:Eu2+, een gele emissie bij 560 nm voor CaSi2O2-δN2+2/3δ:Eu2+ (δ ≈ 0) en een

groen-gele emissie met een maximum tussen 530 en 570 nm voor SrSi2O2-δN2+2/3δ:Eu2+ (δ

≈ 1). De resultaten geven aan dat MSi2O2-δN2+2/3δ:Eu2+ een zeer veelbelovende klasse van

fosfor materialen is voor gebruik in wit-licht LED’s.

Ce3+-gedoteerd MSi2O2-δN2+2/3δ vertoont UV-blauwe emissie (met maxima bij

ongeveer 392, 473 en 396 nm voor M = Ca, Sr en Ba, respectievelijk), waarvan

CaSi2O2N2:Ce,Na de grootste luminescentie-efficiëntie heeft. Interessant om te

vermelden is dat BaSi2O2N2:Ce3+ een wit-achtig licht emitteert voor excitatie met 365

nm straling (Hoofdstuk 10).

In het laatste hoofdstuk (Hoofdstuk 11) wordt beschreven dat EuSi2O2N2 sterk geel

gekleurd is met een hoge kleurzuiverheid. EuSi2O2N2 kan aangeslagen worden in het

zichtbare gebied (370 – 485 nm) en vertoont een brede band emissie met een maximum

bij ongeveer 568 nm. Door combinatie van luminescentie gegevens en magnetische

eigenschappen, is bevestigd dat europium als tweewaardig ion aanwezig is in EuSi2O2N2.

Samengevat: met betrekking tot emissie- en excitatie-kenmerken als ook conversie-

rendement, zijn M2Si5N8:Eu2+, Sr2Si5N8:Ce3+, MSi2O2-δN2+2/3δ:Eu2+ en Y2Si4N6C:Ce3+

veelbelovende materialen als potentiële conversie-fosforen voor LED

verlichtingstoepassingen, en daarom zijn verscheidene materialen gepatenteerd.

Bovendien vertonen dergelijke silicium-nitride gebaseerde materialen een hoge

chemische en thermische stabiliteit. Meer specifiek, heeft deze studie een verbeterd

begrip opgeleverd van de relatie tussen de chemische samenstelling en kristalstructuur

enerzijds en de luminescentie eigenschappen van zeldzame-aard ionen (Eu2+, Ce3+ and

Tb3+) in silicium-nitride gebaseerde materialen anderzijds.

216

Curriculum Vitae Yuan Qiang Li was born in Shandong, China, on 29th December 1962. He received his

B.Sc. degree in Materials Science and Engineering from Wuhan University of Science

and Technology in Wuhan, China, in June 1986. In June 1989, he obtained his M.Sc.

degree in Materials Science from Nanjing University of Science and Technology in

Nanjing, China. His M.Sc. thesis work was in the synthesis and mechanical properties of

the zirconia-alumina materials. After that, he worked as a lecturer in inorganic materials

at the same university. On May 9th, 2001, he started his Ph.D. research project in the

Laboratory of Solid State and Materials Chemistry (now Laboratory of Materials and

Interface Chemistry) at Eindhoven University of Technology under the supervision of Dr.

H.T. Hintzen and Prof. Dr. G. de With, which resulted in this thesis.

217

Acknowledgements First of all, I would like to thank my first promotor Prof. dr. G. de With for providing me

the opportunity to start and finish my doctoral degree research in the Laboratory of

Materials and Interface Chemistry. Thank you very much for your continuous support,

encouragement and open-mindedness. My special thanks for your valuable comments and

suggestions. I really enjoyed the freedom of doing research work at your laboratory.

Second, I am greatly grateful to my co-promotor dr. Bert Hintzen. I will never forget that

you lead me to the lighting world, especially to a new family of nitride-based phosphor

materials. Many thanks for your patience, knowledge, critical comments. Without your

guidance, reading and corrections, it would have been impossible to finish this thesis.

I would like to thank Prof. dr. R. Metselaar for his valuable comments and discussions on

my manuscripts.

I am grateful to my second promotor Prof. dr. ir. M.C.M. van de Sanden for valuable

comments and suggestions.

I also would like to thank Prof. dr. K.V. Ramanujachary (Rowan University, USA) for his

interest in my project, advices and remarks. In particular, I am thankful to you for the

magnetic measurements and explanation of the basic principles.

Our thanks also go to Prof. dr. R. Marchand (Université de Rennes I, France), Prof. dr. A.

Meijerink (Utrecht University) and Dr. Detlef Starick (LWB, Breitungen GmbH,

Germany) for being a member of my Ph.D. committee. Also the interest of Prof. dr. C.R.

Ronda (Philips Research, Aachen, Germany) is greatly appreciated.

Furthermore, I would like to thank Anneke Delsing for her constantly technical supports

in the laboratory. In addition, thank you very much for explanations and translations of a

number of Dutch documents to me.

218

I am grateful to the former group members Dr. Changming Fang and Dr. Qinshan Zhu for

scientific and experimental helps. Many of your ideas and valuable suggestions have been

incorporated into this thesis. Particular thanks to Changming, you let me know the

Computational Materials Science and Chemistry.

I would like to express my deep thanks to Marco Hendrix. You always keep the X-ray

powder diffractometer (Rigaku) in a perfect condition which contributes very much to

this thesis.

I am also grateful to Huub van der Palen for maintenance of the furnaces and solving

practical problems, and Niek Lousberg for the help with SEM analysis.

Immense thanks to Dr. A.A. Kodentsov, my evening and weekend work-partner, you

always encouraged and supported me!

We acknowledge the financial support of OSRAM Opto Semiconductors GmbH

(Germany) and the TU/e department TDO (Technology for Sustainable Development).

I also would like to thank several people who kindly provided some scientific programs

used for this thesis, Dr. Angela Altomare (Institute of Crystallography CNR, Italy) for

EXPO, Prof. J.D. Gale (Curtin University of Technology, Australia) for GULP (General

Utility Lattice Program), Dr. T. Balic Zunic (University of Copenhagen, Denmark) for

IVTON and Dr. Jürgen Köhler (Max-Planck-Institut für Festkörperforschung, Stuttgart,

Germany) for MAPLE (Madelung part of the lattice energy). Also I am thankful to Dr.

O.K. Andersen (Max-Planck-Institut für Festkörperforschung, Stuttgart, Germany) for

TB-LMTO-ASA and helpful discussions for the compilation under Linux system, and

Prof. dr. George M. Sheldrick (University of Göttingen, Germany) for providing

SHELX97. In addition, I wish to express my sincere gratitude to a great number of my

network colleagues @ http://www.biolover.com, http://www.crystalstar.org and

http://groups.yahoo.com/group/sdpd, many thanks for your valuable help and sharing

your experiences on the programming and X-ray crystallography.

219

I would like to thank my former boss Prof. Jie Xu, for his understanding and sharing his

enormous experiences. These hands-on skills are really invaluable, which cannot be

obtained from any textbooks. Thanks again, I wish you would hear my blessing.

I also owe my special thanks to all of my former and present colleagues, Linda van Loon-

Nunen, Imanda Scholten-Kamstra, Dr. G.F. Bastin, Dr. Jos Laven, Dr. W. Ming and all

the postdoc’s and Ph.D. students in SVM/SMG for kind helps. Thanks are also given to

Edwin van Steen, Roel Copic, Sjoerd Stelwagen and Laurent Grygiel.

I will extend my thanks to my Chinese friends, Zhenhua, Xiaojie, Xuanwen, Qianyao,

Mingwen, Qinjun, Huiqi, Z. Chen, Zhili, Y. Ma, Xiaoniu, D. Wu, M. Yuan.

Finally, I am grateful for my family. My parents, my wife Yuqin and my son Tianqi，谢

谢你们自始至终的理解和支持！

220

List of Publications 1. C.M. Fang, Y.Q. Li, H.T. Hintzen and G. de With, “Structure and electric structure

calculations of MYSi4N7 (M = Sr, Ba)”, J. Mater. Chem., 2003, 13, 148.

2. Y.Q. Li, G. de With and H.T. Hintzen, “Synthesis, structure and luminescence

properties of Eu2+ and Ce3+activated BaYSi4N7”, J. Alloys Comp., 2004, 385, 1.

3. H.T. Hintzen and Y.Q. Li, “Novel nitride phosphors: Rare-earth-doped Silicon-

Aluminum-(Oxy)nitride Materials”, Encyclopedia of Materials Science and

Technology, Elsevier Science Ltd., (2004).

4. Y.Q. Li, C.M. Fang, G. de With and H.T. Hintzen, “Preparation, structure and

photoluminescence properties of Eu2+ and Ce3+-doped SrYSi4N7”, J. Solid State

Chem., 2004, 177, 4687.

5. Y.Q. Li, A.C.A. Delsing, G. de With and H.T. Hintzen, “Luminescence properties of

Eu2+-activated alkaline earth silicon oxynitride MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba): a

promising class of novel LED conversion phosphors”, Chem. Mater., 2005, 17,

3242.

6. Y.Q. Li, G. de With and H.T. Hintzen, “Luminescence properties of Ce3+-activated

M2Si5N8 (M = Ca, Sr, Ba) materials”, J. Luminescence, 2005, in press.

7. Y.Q. Li, G. de With and H.T. Hintzen, “Luminescence properties of Eu2+-doped

MAl2-xSixO4-xNx (M = Ca, Sr, Ba) conversion phosphor for white-LED

applications”, accepted for publication in J. Electrochem. Soc..

8. Y.Q. Li, G. de With and H.T. Hintzen, “Luminescence of a new class UV-blue-

emitting phosphors MSi2O2-δN2+2/3δ:Ce3+ (M = Ca, Sr, Ba)”, accepted for publication

in J. Mater. Chem..

9. Y.Q. Li, K.V. Ramanujachary, S.E. Lofland, G. de With and H.T. Hintzen, “Optical

and magnetic properties of EuSi2O2N2”, submitted to J. Mater. Res.

10. Y.Q. Li, G. de With and H.T. Hintzen, “The effect of replacement of Sr by Ca on the

structural and luminescence properties of red-emitting Sr2Si5N8:Eu2+ phosphor”, to

be submitted.

11. Y.Q. Li, G. de With and H.T. Hintzen, “Structure and luminescence properties of

YTbSi4N6C”, to be submitted.

221

12. Y.Q. Li, K.V. Ramanujachary, G. de With and H.T. Hintzen, “Structure and

luminescence properties of Ce3+-doped Y2Si4N6C”, to be submitted.

13. Y.Q. Li, G. de With and H.T. Hintzen, “Luminescence properties of rare-earth-doped

BaSi7N10”, to be submitted.

14. Y.Q. Li, J.E.J. van Steen, A.C.A. Delsing, G. de With and H.T. Hintzen,

“Luminescence properties of red-emitting M2Si5N8:Eu2+ (M = Ca, Sr, Ba) LED

conversion phosphors”, to be submitted.

15. Y.Q. Li, G. de With and H.T. Hintzen, “Luminescence properties of Tb-doped

Y2Si4N6C”, in preparation.

16. H.T. Hintzen, Y.Q. Li and K.V. Ramanujachary, “Structural, optical and magnetic

properties of YCeSi4N6C”, in preparation.

17. Y.Q. Li, G. de With and H.T. Hintzen, “Luminescence properties of Tb-doped

MSi2O2-δN2+2/3δ (M = Ca, Sr, Ba)”, in preparation.

18. Y.Q. Li, G. de With and H.T. Hintzen, “Luminescence properties of Tb-doped

M2Si5N8 (M = Ca, Sr, Ba)”, in preparation.

Patents

1. H.T. Hintzen, Y.Q. Li and A.C.A Delsing, “Luminescent material, especially for LED

application”, WO 2004/30109 A1, 2004.

2. H.T. Hintzen and Y.Q. Li, “Luminescent material and light emitting diode using the

same”, WO 2004/029177 A1, 2004.

222